Search

The rapid and accurate detection of defects in solar energy plants is of great importance to reduce efficiency losses and extend the lifespan of photovoltaic systems. In this study, the effectiveness and advantages of You Only Look Once (YOLO) algorithms for hotspot detection in solar energy plants have been investigated. YOLO algorithms can be efficiently used in large-scale facilities due to their ability to detect objects in images in a single scan at high speeds. In this context, the performances of YOLOv5, YOLOv6, YOLOv7, and YOLOv8 algorithms were compared, and the best-performing model was determined. According to the results of the experiments, 80% of the 100 images obtained by an unmanned aerial vehicle in the dataset were used for training, and the remaining 20% were used for testing the YOLO algorithms. The results indicated that the YOLOv8 algorithm outperformed other models with 88.7% specificity, 80.5% sensitivity, and 83.8% mean Average Precision (mAP) values. The dataset used in the study consisted of images obtained from real solar panels, ensuring that the results of the study were tested in accordance with real-world scenarios. The findings demonstrate that YOLO algorithms are an effective method for detecting hotspot defects in solar panels. This study highlights the importance of using object detection algorithms to make solar energy plants more efficient. Additionally, it can be considered as a guiding and contributing study to the literature, providing insights for future research. [ABSTRACT FROM AUTHOR]

Cracks in buildings are a source of concern as they may indicate structural problems. Cracks in buildings are one of the major problems with concrete structures, as they affect the appearance of the building, compromise the integrity of the masonry, jeopardize building safety, and reduce the durability of the structure. Cracks in buildings can be a cause for concern and may indicate a potential structural problem that could jeopardize the safety and stability of the building. Understanding the root causes of these cracks is crucial to determining appropriate preventive measures and repair methods. In this study, crack and slope control in buildings was performed with an image processing-based measurement system developed using CNN deep learning algorithms. A dataset of 294 photographs was used for the study. The data were subjected to image processing, preprocessing, and detection with CNN. The accuracy rate is 0.9716 for the VGG-16 architecture and 0.9513 for the Inception-V3 architecture. It was concluded that there is no significant difference between the accuracy rates of the architectures. All vertical and horizontal cracks with a thickness of 2 mm and a continuity of 4 cm were detected. Thus, crack control in buildings can be done quickly and reliably, and building safety will be ensured. [ABSTRACT FROM AUTHOR]

With the rapid advancement of technology, the use of artificial intelligence (AI) is increasing in various sectors such as education, health, security and defense. A critical application of AI is highway management, especially with the rise of autonomous vehicles. The focus of this study is to address the problem of deformations in highway marking lines that pose challenges for autonomous vehicles and affect traffic safety. The research involves using unmanned aerial vehicle (UAV) to create an original image dataset of highway lines. This data set will be processed with image enhancement techniques and deep learning models. The first stage involves cleaning the images from foreign matter. Deep learning models will then identify potential line deformations. These models will be developed and trained for optimum accuracy using various performance metrics. In the study, the mobilenet v3 model, trained with the images in the data set, reached an accuracy rate of 89.58%, the resnet50 v2 model 77.78% and the Convolutional Neural Network model 92.55%. The ultimate goal is to implement a real-time system to accurately detect and report differences in highway lines by combining the UAV with the computer system on the ground. This will ensure timely notification to authorities and help prevent traffic safety problems related to line deformations. This approach demonstrates the practical applications of artificial intelligence in improving road safety and autonomous vehicle navigation. [ABSTRACT FROM AUTHOR]

The properties of the blood vessels in the retina are very important in the diagnosis of retinopathy of prematurity (ROP). In premature infants, the blood vessels of the retina cannot complete their development. Post-natal, certain triggering circumstances cause the creation of regular veins to cease and abnormal blood vessels begin to enlarge forming abnormal tissue. With the increase in the degree of this condition, retinal damage may occur. It is crucial to follow the progress of the disease by following the developments in the vascular networks, especially since babies born prematurely are at greater risk for ROP. The purpose of this study is to develop methods for the detection and segmentation of vascular pathways in ROP images by applying image processing methods to retinal images of preterm neonatal. These methods have been applied to ROP images and the results have been compared numerically. As a result, according to the Peak Signal to Noise Ratio (PSNR) values of the most appropriate image processing method; It has been determined that there is an OTSU filter as the thresholding method and a Gaussian filter as the filtering algorithm. [ABSTRACT FROM AUTHOR]

As Android devices occupy more of people's lives, they have also become a target of malicious software. It is important to detect malicious software and to prevent the losses and damages that may arise from this software. For this purpose, various studies are being carried out about malware detection. Recently, imagebased methods and machine learning studies have come to the fore. In these studies, binary files used in static and dynamic analysis are converted into image files. Global and local features extracted from the images are classified by various machine learning methods. In this study, global features were extracted on the malimg dataset and a feature matrix (2000, 532) long was obtained. The obtained features were classified using machine learning methods (LR, LDA, K-NN, CART, RF, NB, SVM). The results were evaluated using the K-fold crossover validation method, and a highest accuracy rate of 96.72% was obtained with K-NN and 97.44% with RF. This study contributes to the literature by reaching a higher accuracy value compared to other studies on the same data set. [ABSTRACT FROM AUTHOR]

Various proposed approaches for autonomous driving basically involve an image processing and a machine learning process. It is extremely important to use appropriate image processing techniques and a comprehensive data set in these approaches. Moreover, the proposed model must work in real-time. On the other hand, designing and manufacturing an autonomous vehicle model results in serious hardware costs. In addition, the design and manufacturing processes need to be repeated to develop new approaches. In this context, utilizing a real-time simulation environment can be seen as a suitable approach for a less costly prevalidation of such models. In this study, a real-time simulation architecture is developed with Unity framework to test an autonomous driving model. In addition, an autonomous driving model that includes lane tracking and object recognition approaches is proposed, and an autonomous vehicle simulation is created. Finally, the feasibility of the proposed simulation architecture is tested with the convolutional neural networks-based YOLO algorithm and R-CNN algorithm versions. According to the findings, it is observed that Faster R-CNN, Mask R-CNN and YOLO-v4 algorithms produce results with 91%, 93% and 95% accuracy, respectively. It has been determined that these results are close to the accuracy rates obtained on different traffic sign data sets in the literature. Considering the outcomes, it is argued that a vehicle simulation with an autonomous driving model has been successfully tested in the proposed system architecture. [ABSTRACT FROM AUTHOR]

Currently, image sensors (cameras), are widely used for image analysis (classification, segmentation, etc.) and synthesis (object detection, tracking, distance detection, etc.). The aim of the study is to lay the theoretical foundations for the development of an image meter based measurement device (image meter) that can be used for laser-meter, lidar-meter, radar and similar industrial purposes. For this purpose, distance detection methods based on image processing in the literature have been studied. The main factors that negatively affect the performance of these methods have been identified, and a new method has been developed that is not affected by these factors. The measuring device, which is planned to be developed, is based on theoretical foundations. The operation of these theoretical foundations is based on hardware and software components. In the study, these theoretical foundations were given and designs of hardware and software components were realized. In study, as a result of calculations made for 1-1000m, it was determined that a success rate below 0.2% could be achieved. The study consists of an electro-mechanical component have hardware and software projects. It is obvious that errors caused by hardware and software components will reduce the success rate. These errors will consist of standard and random errors. In the study, these errors are foreseen and added to the distance measurement equation. Detection of predicted errors will be determined in the future study with the development of the hardware prototype and software components. [ABSTRACT FROM AUTHOR]

Grasshoppers damage crops and causes millions of tons of food to be destroyed every year. The development of effective and accurate locust identification systems is critical in controlling locust species and preventing food loss. In this study, 11 different plant pest grasshopper species seen in various parts of our country and the world were classified using various convolutional neural network models. The dataset used in the study was obtained by observing the Eastern and Southeastern Anatolia regions of our country. The novelity of this study is that a dataset named GHCD11 has been created for 11 different plant pest grasshopper species in our country. In addition, VGG16, VGG19, ResNet50, DenseNet121, EfficientNet, MobileNet, which are in the Keras library and are widely used in image classification, were used for the automatic classification of 11 different grasshopper species in the study. As a result of experimental studies on the GHCD11 dataset with learning transfer, remarkable classification accuracies in the range of 95% to 99% were obtained. The study is important because it not only presents a novel dataset, but also demonstrates that automatic identification and detection of plant pest grasshopper species can be done with high accuracy using convolutional neural network architectures. [ABSTRACT FROM AUTHOR]

Image processing is used in many sectors in parallel with the developing software and hardware technology. Differences are observed in the image techniques used according to the needs of the sectors. In this context, very precise measurements and powerful algorithms are required to be used in the field of health. In this article, an image mosaic was made on retinal images using datasets such as D-EYE and FIRE. Feature detection, feature matching, image matching, image wrapping and image mosaicing processes are applied on the images. Harris corner detector was used to find the corners in the image. The positions of the points in the 2 corresponding images were found with the corner detector. Scale-Invariant Feature Transform (SIFT) was used as the feature recognition algorithm between the coordinates. Feature and image matching processes were performed between the determined features. Weak spots are eliminated with the help of Random Sample Consensus (RANSAC). The homography is defined for the projective transformation with the remaining points. In the last step, a geometric transformation was made with the homography matrix. The image mosaic application in our study; It can be applied to a wide area, especially retinal images, unmanned aerial vehicle images, bacteria or tomography images. It has been tested with different data sets and a successful result has been obtained. [ABSTRACT FROM AUTHOR]

In this study, the Finsler curve evolution model, which was established with Finsler geometry and turned into a curved evolution model with the calculation of variations, will be examined. This model is not constructed by considering only locations in the image space as in the Riemann curve evolution model. It was established in an anisotropic space by considering both locations and directions. The model offers a flexible operation in the image by eliminating the necessity of isotropic spaces. Therefore, it has become a model that researchers working on image processing mostly work on. It is aimed to contribute to future studies in the field of image processing by using the mathematical analysis of the Finsler curve evolution model discussed in this study. [ABSTRACT FROM AUTHOR]

Measurements of serial production workpieces in the industry are performed by camera-controlled systems thanks to the advantage of speed. The measurement success of camera systems largely depends on the measurement algorithm and ambient conditions. Area-based approaches are not preferred in industrial machine vision applications due to the undesired environmental conditions. In this paper, an area-based diameter measurement algorithm that can be used in industrial machine vision applications is proposed. The success of the proposed method is demonstrated based on the sub-computation metric. In the proposed method, firstly, the noise on the obtained image is cleaned according to the connected component analysis. Then, the inner and outer diameters of the largest component are determined according to the area calculation. In the designed experimental setup, a back lighting illumination has been preferred. According to 3 different positioning types in the field of view of the camera, a total of 40 stamps of 4 types were measured 20 times with 3 different lenses. According to the test results, it has been observed that the position of the part on the field of view greatly affects the repeatability measurements. Also, sub-computation metric (C) is measured 2 in random positioning. This value increases up to 5 in the limited positioning that meets the industrial conditions. Tests have shown that the proposed method can measure the diameters of workpieces with precise tolerances in an industrial setting. [ABSTRACT FROM AUTHOR]

Driver drowsiness detection is an important issue to prevent traffic accidents. 40% of severe traffic accidents are due to drowsiness. Various methods are used for driver drowsiness detection. One of the driver drowsiness detection method is driver drowsiness detection based on the analysis of signals such as EEG and ECG. Another driver drowsiness detection method is driver drowsiness detection based on vehicle-driver interaction. The last driver drowsiness detection method used in the study is driver drowsiness detection from images. This method is more advantageous than the other two methods in terms of cost and usability because no driver intervention required. Classical image processing techniques and deep learning algorithms are used for driver drowsiness detection from images. Recent driver drowsiness detection studies are based on deep learning models. In addition, the model to be developed will need to be able to work on mobile devices in order to ensure widespread use. In the study, Convolutional Neural Networks were used for driver drowsiness detection on mobile devices. In order to increase the success rate of the model, the pre-trained model was reused with the transfer learning technique. The model developed for the training consists of 14 layers and 1,236,217 parameters. The dataset consists of two categories, open-eye and closed-eye images. The developed model achieved 95.65% accuracy, 95.86% precision, 94.32% recall and 95.17% f1 score which achieved better results than previous studies. A dataset of 2425 images was used to train the model. [ABSTRACT FROM AUTHOR]

Nowadays, trade based on logistics and sea transportation has gained importance. Considering the traffic related to this, the classification and discrimination of ship types are important in terms of transportation and storage costs, and safety. The classification of ships performing different tasks on the sea is handled in this study, and a ship image dataset has been created in order to make a high accuracy ship classification thanks to deep learning methods. It is preferred in our deep learning study in comparison to the classical machine learning method, as the features are semantically richer as the higher level, while expressing content from the dataset. It was created by the acquisition of various ship images thanks to the web scraping method. YOLOv5 and Xception deep learning models were trained to obtain the most appropriate classification performance. As a result of the experiments, an accuracy rate of approximately between 96% and 99% was achieved with both models. Scientific findings and discussion are included in our study as well. [ABSTRACT FROM AUTHOR]

Weather conditions appear as an unchangeable structure. However, determining and determining weather conditions can help individuals plan their physical activities. In the study, it has been tried to perform different sky images and weather detection processes with image classification methods, which is one of the popular work subjects in the computer field in recent years. In the study, a data set consisting of images with different weather conditions and resolutions was used. The number of images in the data set has been increased by using various data augmentation methods. The feature maps of the images were obtained by applying image processing techniques to the images. In the next part of the study, the classification process was carried out on the images with an accuracy rate of 96.4% using the DenseNet image classification model. Thanks to the classification performed, weather detection with high accuracy can be achieved. [ABSTRACT FROM AUTHOR]

Pistachio is a product with high commercial value and Turkey has a serious production and export potential in this field. The harvesting and sorting of these products is as important process as their cultivation. Classification of pistachios according to their outer shell color after harvest is one of the important processes that must be done before presenting the pistachio to market. In this study, deep learning techniques were used to sort the pistachios according to their outer color. The data set required for the extract process was obtained using an industrial experimental setup. The AlexNet structure was trained and tested with these data sets. The network structure was trained and tested for two different percentages (70%, 15%, 15% and 60%, 20%, 20%) and three different data numbers to investigate the effect of training, validation and test data in different percentages and numbers on test accuracy. The highest test accuracy was calculated as 98.44% by using of 60%, 20%, 20% percentages and 580x2, 193x2 and 193x2 data for training, validation and testing, respectively. Classification process has been performed with deep learning techniques with high accuracy. Network structures trained with the developed industrial data set are at a level to obtain high accuracy in industrial sorting machines. [ABSTRACT FROM AUTHOR]

Vehicles take place in a large part of our daily life and their number is increasing day by day. This situation has made the tracking of vehicles in motion and the detection of their license plates a necessity. Performing these processes automatically will provide a great convenience in terms of labor and economy. In this study, a system that recognizes a moving target and tracks vehicles using image processing algorithms using the data obtained with a quadrotor DJI Tello Drone camera is proposed. The work consists of three stages: The first one is transferring the images from the drone to the software and the second one is recording the images received at the end of the analysis to the database; In the third stage, semi-autonomous tracking of a specified object within the data obtained is possible. Tracking height and distance can be adjusted manually. The system detects a vehicle in real-time and records five different data such as registration date, license plate text, license plate image, vehicle image, and tracking image. In the study, possible environmental (intense light reflection on the plate, the intensity of transition from dark to light, sticker or dirt on the plate, etc.) errors can be eliminated by comparing the plate text with the plate image before the plate information is recorded in the database. In case of an error, the license plate information is saved in the database after the user corrects it manually. The proposed system can be used in many fields, including film, aerial photography, and the defense industry. According to the results obtained, 79.54% success is achieved in license plate recognition and 85% success in the object tracking systems. [ABSTRACT FROM AUTHOR]

It is aimed to carry out an experimental study by creating an algorithm function that enables the improvement of gel electrophoresis band images with the help of the Python programming language-based, mini gel electrophoresis system image processing interface. With this system, the analysis of gel images under UV (Ultraviolet) light was made with the designed interface software. The images were transferred to the interface software via the camera and filters were applied to highlight the lanes and bands. The two-term power function, which estimates BP (Base Pair) numbers with the most accurate result, was used. The software can be used on Raspberry Pi 3B+ by being integrated into the mini gel electrophoresis system via Python programming language with OpenCV library. In this study, gel images obtained from the mini gel electrophoresis system controlled by the embedded system were used to estimate the BP numbers of the bands in the images. By importing the gel images from the camera or from the file in the system, the interface software algorithm enabled the estimation of BP numbers in the manual measurements and reducing the average error rates to the range of 0.55% - 0.86%. In the interface software, which enables the calculation of lanes and BP numbers in gel images with the lowest error rate, two-term power function has been applied to the image processing algorithm instead of the two-term exponential function, based on the principle that exponential functions can also be defined as power functions. It has been revealed that the BP values calculated with this two-term power function work in accordance with its purpose with the value of R²=0.9999533, which shows the closeness of the actual BP values. [ABSTRACT FROM AUTHOR]

The traditional karyotype studies are widely used in plant systematics to evaluate the positions of species. However, studies sometimes can not solve systematic problems. In recent years, computer-based systems have gained importance in contributing to the solution of the taxonomical problems. The aim of this study was to identify three Lallemantia species by using the LBP (local binary pattern) texture operator based on chromosome images in mitosis. LBP is the one of the most powerful and easily applicable tool for identifying local image patterns. In this study, microphotographs of 641 cells in the metaphase stage of mitosis were used. The LBP involves preprocessing, feature extraction, feature selection, and classification. Decision tree (DT), linear discriminant (LD), support vector machine (SVM), K-nearest neighbor (KNN), bagged tree (BT) and ensemble subspace nearest neighbor (SNN) were used in the classification stage. This study found that the best acting classifier was SNN because achievement rate was one hundred percent. Also, a dendrogram was formed to measure the similarity among the three species. As a result, LBP can be accepted as a tool for classifying plants by using chromosome images. [ABSTRACT FROM AUTHOR]

The diagnosis and detection of kidney stones in medical images are important in terms of determining disease progression and treatment methods. Noisy and lowresolution Abdominal Computed Tomography (CT) images are factors that complicate the diagnosis of kidney stones. The goal of this study is to determine the borders and number of stones found in the kidney on abdominal CT images. Firstly, the Gamma Correction method is used to sharpen CT images and eliminate noise. Secondly, the kidneys are selected manually using the Region of Interest (ROI) method. The detection of kidney stones is carried out by applying a threshold to the selected kidney regions. By applying morphological erosion and dilation processes to the stones within the limits of the kidney, the stones detected adjacent in the image are separated, and the stones are counted. The confusion matrix is created to calculate the performance of the study. In the performance analysis of the proposed method, the Accuracy, Sensitivity, and Specificity values using the Confusion Matrix are calculated at 97%, 96%, and 100%, respectively. The results show that the proposed method is successful in detecting and counting kidney stones. [ABSTRACT FROM AUTHOR]

The aging of dielectric materials is unavoidable during their service life. In addition to the electrical properties of dielectric materials, the surface properties that affect their stable operation, performance, and service life are among the important factors. Considering that the electrical aging phenomena initiate and grown in the presence of electric field and humidity, it is obvious that wettability is among the surface properties that are of great importance especially in outdoor applications. Wettability allows characterization of the solid surfaces and determination of solidliquid interactions. In this study, snapshots of saltwater droplets with a conductivity of 167.9 mS/cm and a volume of 20 μL dropped onto the dielectric material surface were taken using a digital microscope for 5 different time points, 1st, 10th, 20th, 30th, 40th, and 50th minutes. The obtained images were enhanced using image processing techniques and the droplet image was segmented. Discrete Cosine Transform (DCT) was applied to the saline droplet images and the DCT coefficients were extracted. The standard deviation of the extracted coefficients was plotted as a time-dependent scatter graph. For the time-dependent points obtained, linear and quadratic polynomial curves were fitted and the mathematical equations of the fitted curves were obtained. A DCT-based approach is presented to evaluate and interpret the wettability and evaporation rate for the dielectric materials' surfaces. As a result, the relationship between evaporation rate, time, wettability, and contact angle is observed using image processing techniques. [ABSTRACT FROM AUTHOR]

In this study, an image processing-based system was developed that performs vote detection and counting. The system consists of two parts, hardware and software. The camera used for image capturing and Raspberry Pi3 used for image processing constitute the hardware part. In the software part of the system, the features in the images are matched using Oriented FAST and Rotated BRIEF (ORB) method and Brute-Force Matching method which are the image processing methods. The software's codes are written in the Python programming language and the OpenCV library is used. In the study, YES and NO ballots written in Calibri font, 72 points in size, were used for vote detection and counting. Thanks to the high-resolution camera in the system, the image of the ballot is taken and transferred to the image processing software. The operating logic of the software works according to the principle of determining the corners and milestones, which are distinctive features in the recorded image, and matching them with the images taken. In order to determine who the game was given to, the position of the game on the horizontal plane used is looked at. The system increases the number of yes and no votes according to the matching image. In the study, it is performed that vote determination and counting with a 100% success rate. [ABSTRACT FROM AUTHOR]

XLPE dielectric materials have an important place in electricity transmission and distribution. The aging of XLPE dielectric materials during the service period is inevitable. The surface properties of XLPE dielectric materials play an active role in the determination and evaluation of their service life. Properties such as hydrophobicity and hydrophilicity are the methods used in the analysis of the surface properties of materials. Within the scope of this project, contact angle measurements were carried out under the electric field for evaluation and determination on the wettability behavior of the XLPE material, which plays an important role in electrical insulation, which is widely used in electrical energy transmission and distribution. In the contact angle measurement platform manufactured using the 3D printer, XLPE insulator was placed between the aluminum plane electrodes and exposed to an electric field. A high voltage source of 0-10kV was used as a high voltage generator during the experiments. Snapshots of the 1.8M saltwater droplet dropped on the XLPE insulator surface under the electric field were taken. In the image processing part of the study, images taken with a digital microscope were processed using the image processing techniques with the help of a graphical user interface (GUI) prepared on the MATLAB platform. Contact angle values were obtained as output. Contact angles obtained from XLPE surfaces under the electric field will allow an evaluation of the service life of XLPE dielectric materials. An approach based on the contact angle for evaluation and determination of the wettability of XLPE dielectric materials is presented. As a result, the correlation between the electric field, wettability, and contact angle was observed. [ABSTRACT FROM AUTHOR]

Bees are one of the oldest living species in the world, having a major impact on the development of living species. The continuity of plants at the bottom of the food chain is directly related to the pollination of bees. Bees are a global insurance because of this characteristic. For this reason, it is very important to check the health status of bees. Depending on the technology developed nowadays, it is possible to control the health status of bees remotely with real-time image processing applications. In the study, feature extraction methods, which are the strengths of deep learning, were executed from two different arms and aggressive changes in images were detected. In the classification process; Instead of Softmax classifier based on probability calculation, multi-layer feedback artificial neural network (MLFB-ANN) model has been used. The success of the designed system has also been compared with the Softmax classifier. As a result of experimental studies, 93.07% success rate can be achieved with Softmax classifier for six different bee diseases on the same data set, while 95.04% success rate has been obtained with the developed system. In this study, a hybrid method based on deep learning methods was proposed for the classification of bee diseases and successful results were obtained. [ABSTRACT FROM AUTHOR]

Plants, which are a part of nature, are used for different reasons, such as alternative medicine as well as adding beauty to our environment. In such cases requiring expertise, the misinformation spread among the public and the use of poisonous plants considering that they are medicinal causes problems that can go up to death. In this study, it is aimed to solve these problems through a system that determines the species of leaves in leaf images using artificial intelligence techniques. Convolutional Neural Network (CNN), one of the most widely used artificial intelligence techniques, has been used recently. Its multi-layer structure, having many parameters and being able to learn features without requiring too much pre-processing is one of the reasons why it is used in many studies. In this study, the training of the convolutional neural network was carried out separately with 5 different data sets consisting of leaf images with a fixed background, and the effect of these training parameters on training was investigated. A combined data set consisting of 270 species obtained by combining these data sets was created. A general purpose leaf classification model is obtained with convolutional neural network. The results obtained by the classification process were compared with the studies in the literature. [ABSTRACT FROM AUTHOR]

3D fabrication technology has gained a lot of attention recently, with the rapid fabrication of 3D parts for use in various applications. This article presents an electrothermal micro-actuator's characterization study fabricated using digital light processing (DLP), one of the 3D manufacturing methods. This actuator has been designed and fabricated to move in two directions according to the specified criteria. Three different experiments were conducted during the fabrication process. During the fabrication process of the first two experiments, deterioration or breakage occurred in the micro-actuator structure. These problems did not occur in the fabrication of the 3rd experiment. A characterization test setup was created to observe the displacement of the actuator under different operating voltage values. This setup includes an optical microscope and digital camera to observe displacements at the probe station. An image processing algorithm was used to detect the observed displacements. In experimental studies, the operating voltage was increased from 0V to 1V until deterioration or breakage in the micro-actuator structure was observed. When 6V voltage is applied to the micro-actuator, a maximum displacement of 3.84 µm occurred in each direction. When more than 6V voltage is applied to the micro-actuator, deformations or breaks occurred in the actuator arms. As a result, the micro-actuator's design is bi-directional, so its maximum displacement is 7.68 µm. [ABSTRACT FROM AUTHOR]

Underwater vehicles are used for regular control of underwater life and pollution, which are legalized and made mandatory in the European Union. In addition to environmental disasters caused by chemical leaks to oceans in different parts of the World, increasing oil and natural gas exploration studies in the seas, especially in our region, have increased the importance of underwater vehicles that are used in drilling platforms. Underwater vehicles have an important role in offshore fishing and sunken ship investigations. In this study, the peripheral units, sensors, operators used in underwater vehicles and their effects on vehicle control and cruise planning and the stages of autonomous movement in unmanned underwater vehicles are given. In this context, the general goals, tasks and objectives that we encounter as criteria during the design stage are explained. Then, the basic concepts that should be considered in designs for mechanical, electronic and software development required to achieve these are mentioned. As an example, it has been discussed how to determine some dynamics parameters of the underwater vehicle Lucky Fin developed in the Electronics and Communication Department of Kocaeli University Faculty of Engineering. Unmanned underwater vehicles use many subjects such as robot kinematics, hydrodynamics, modeling, control, image processing, electronic card design, sensor technologies, embedded systems, drive systems, power management, lighting, underwater communication, image processing algorithms, navigation planning which require an interdisciplinary coordination. The article aims to contribute in providing the reader an overview of the generic concepts and design principles encountered in vehicle design. [ABSTRACT FROM AUTHOR]

Continuous Wave Diffuse Optical Tomography (CWDOT) system is one of the emerging medical imaging modalities. This study aims to apply different three dimensional (3D) image processing methods to breast phantom images and determine the most appropriate image processing methods in CWDOT. A breast phantom was made with a mixture of intralipid, water, and Indocyanine green (ICG). Inclusions were used to represent the tumor inside the phantoms. First, image processing methods of spatial filters such as Average, Gaussian, Laplacian, Laplacian of Gaussian (LoG) filtering methods were implemented. Then, nearest, bilinear, bicubic, and cubic spline interpolation methods were applied to the images. After the image processing using different filters and interpolation techniques, the images were evaluated numerically using the Peak signal-to-noise ratio (PSNR), Mean square error (MSE), and Structural similarity index (SSIM) methods. In this study, the most appropriate image processing algorithm was defined based on the location, shapes, and sizes of the breast phantom inclusions. Our results show that Gaussian filtering and Bicubic Interpolation are the most appropriate filtering and interpolation techniques for the images obtained from CWDOT. [ABSTRACT FROM AUTHOR]

Deep learning, one of the most popular topics of recent years, is frequently used in the biomedical field. Disease and fracture detection, data estimation, tissue and organ segmentation, and many other applications are successfully carried out with deep learning algorithms. However, while these applications are valid for human medicine, veterinary medicine is left behind. So, the deficiency in the literature about animal's X-Ray images processing has been the biggest motivation for this study. In this work, a comprehensive dataset containing dog X-Rays taken from Ankara Metropolitan Municipality Stray Animals Temporary Nursing Home, was processed with deep learning algorithms. It is aimed to determine the type of long bone. Like many other studies, Convolutional Neural Networks'(CNN) architectures are utilized in this study, too. Transfer learning was carried out with AlexNet, GoogLeNet and VGG-19 deep learning models, and the classification performance was tested with Support Vector Machines (SVM). [ABSTRACT FROM AUTHOR]

Converting characters or data to text through image formats means loss of time and labor. Today, It is desired to reduce transaction costs and increase efficiency rates in document transactions. For reading have been different writing styles, font sizes and writing formats on the documents. Computer-generated prepared character conversion and known writing style with success back into editable text format, the success of the conversion value higher then handwritten characters. The biggest step for reading in character, separating of characters from background. Due to the lack training data of the Tesseract library, the success rate, lows in some writing formats. In this study, reading alphabetical character and numbers was performed with the help of Tesseract library used for OCR technology and was made on different writing styles by applying different writing styles reading success comparison. Samples using normal, bold and italic writing formats were used on Times New Roman, Calibri and Arial font styles. Also on the image before using Tesseract library, Error Diffusion algorithms were compared with read rates by making improvements. Thus, it is aimed to find a pre-processing algorithm that increases the success of the OCR recognition method. According to the obtained values; As a pretreatment on the document, it was observed that the Tesseract library made a more accurate reading after using the Floyd Steinberg error distribution algorithm. [ABSTRACT FROM AUTHOR]

Although the architectural design process is carried out with the collaboration of experts who are experienced in many different areas from the main preferences to the detailing stage, the major decisions such as plan organization, mass design etc. are taken by the architect. Computer Aided Design (CAD) programs are generally effective after the major decisions of the design are taken. For this reason, it is common for the main decisions, taken during the design process, to be changed during the analysis of the structural system. In order to prevent this, in the early stages of architectural design, earthquake system awareness and structural system design should be included as an design input; as, the failure of the structural system which did not considered well in the architectural design phase leads to unexpected revisions in the implementation project phase and thus leads to serious losses in both time and cost. The aim of this study is to create an Irregularity Control Assistant (IC Assitant) that can provide architects general information about the appropriateness of structural system decisions to earthquake regulations in the early stages of design process by using the deep learning and image processing methods. In this way, correct decisions will be made in the early stages of the design and unexpected revisions that may occur during the implementation project phase will be prevented. [ABSTRACT FROM AUTHOR]

Face represents an individual's claimed sense of positive image in the process of social interaction. Consumers with a high level of face consciousness may consume conspicuous using branded products. The main reason underlying this behavioral effect is defined as brand affect. Because brand affect is considered as a concept that helps explain the relationship between consumer and the brand. On the other hand, the force of brand affect may differ in terms of gender. From this point of view, the aim of the study is to reveal the moderating role of brand affect and gender in the relationship between face consciousness and conspicuous consumption. In the study, 454 usable completed questionnaires were received by using self-completion survey method. PROCESS Macro software, which was used by installing SPSS program, helped to perform moderating analysis. According to the results, it is revealed that the effect of face consciousness on the conspicuous consumption is more effective for consumers with high brand affect. This effect is more powerful for men with high brand affect. In the conclusion section, after theoretical discussions, marketing managers were advised potential marketing strategies that they could apply in the branded product market. [ABSTRACT FROM AUTHOR]

In this study, physical and sensory changes of C. gibelio were determined during chilled storage and quality changes were assigned by using image processing and instrumental method. Sensory evaluations of specimens were performed by trained panelists in terms of their colour, odour, texture, general acceptability and sensory score and the highest correlation (r=-0.95) between storage time were found in colour parameter. In addition to this specimens were photographed by using high definition camera and analysed in image processing programme to obtain red (R), green (G) and blue (B) parameters. Specimens were analysed simultaneously by colour meter and L, a and b values were obtained according to CIELAB diagram. In accordance with the results of this study, colour components (L, a, b and R, G, B) were not suitable alone to determine the quality changes of C. gibelio however, it is concluded that total colour changes (ΔE and ΔRGB) are more accurate to assign the storage time in terms of correlation (r=0.53 and r=0.38) and regression (R²=0.35 and R²=0.43) coefficients. [ABSTRACT FROM AUTHOR]

Image processing applications in real-time systems have become a popular topic in recent years. Deep learning methods, one of the sub-branches of artificial intelligence, and image processing algorithms used in the field of object detection from images can be used together. In this way, applications are developed in many areas such as autonomous cars, autonomous unmanned aerial vehicles, assist robot technologies, assistant technologies for disabled and elderly individuals. This study aims to detect the tactile paving surfaces with deep learning methods in order to design an assistive technology system that can be used by visually impaired individuals, autonomous vehicles and robots. Contrary to traditional image processing algorithms, deep learning methods and image processing algorithms are used together in this study. The YOLO-V3 model, which is one of the best methods of object detection, is combined with the DenseNet model to create the YOLOV3-Dense model. YOLO-V2, YOLO-V3 and YOLOV3Dense models were trained on the Marmara Tactile Paving Surface (MDPY) dataset, which was created by the researchers and included 4580 images and their performances were compared with each other on the test dataset. It was observed that YOLOV3-Dense model is better than other models in detecting tactile paving surface with 89% F1-score, 92% mean average Precision(mAP) and 81% IoU values. [ABSTRACT FROM AUTHOR]

The correct perception of objects is directly related to the suitability of the ambient illumination. This level of conformity is determined by a measuring instrument which is known as luxmeter which measures on point based. However, these measurements made on a point basis have some disadvantages such as cost, long processing times. In order to eliminate these disadvantages in recent years, there have been studies in which CCD cameras, mobile phones and video cameras have been used as luxmeters. In these studies, said devices are placed directly opposite the light source and used as a light collector such as a luxmeter. They then calculated the illumination intensity by digital image processing techniques using aquired source images. In this study, the camera is not used directly as a luxmeter. However, an overall image representing the entire environment is taken. A new method based on image processing is presented for illumination intensity calculation. The image acquired is separated into its frequency components in the frequency plane by the homomorphic filtering, and the illumination component representing the source's illumination is detected by the low pass filter. The average pixel value of the illumination component is calculated and defined as illumination intensity. In order to test the usability of the proposed method, the measured illumination intensity with a luxmeter and the computed results of the developed method are sta tistically analyzed. The results of this analysis showed that there was a linear relationship between the measurement and calculation values with satisfactory R2 values. [ABSTRACT FROM AUTHOR]

Nowadays, the detection of buried plastic anti-personnel mines is an important theoretical and applied research area which is technologically very difficult and is currently being studied. In this regard, with the development of different methods in the literature, thermal imaging method is a preferred method with its relatively easy application and cost efficiency. In this study, in order to determine the circular geometry of the plastic anti-personnel mine in the thermal image from the high-noise images, in addition to the image processing techniques, a new mine detection algorithm based on the circular Hough transformation is performed and the experimental results are presented to the literature. The results show that the proposed method is promising. [ABSTRACT FROM AUTHOR]

It is important for the sensitivity of ecological balance that image processing methods and techniques give better results day by day. This article deals with the classification of flower images which are the basic elements of ecological balance. Recently the use of deep learning methods on flower images has increased. In this study, the dataset available on the internet was used for the classification of flower images. The dataset consists of 4326 images. In the obtained dataset, convolutional neural network (CNN), one of the deep learning models, was used for feature extraction. The CNN architectures AlexNet, VGG-16 and VGG-19 are used in this study. The common feature of the three models is that they have a fully connected layer that give 1000 features. The features obtained from flower images were classified with support vector machines (SVM) and the results were compared. As a result of the comparison, the best classification performance was achieved with VGG-16 architecture. Obtained classification accuracy rate is 86.56%. In the next phase, 1000 features obtained from the last fully connected layer of CNN architectures were combined and 3000 feature set were created. Then, the most efficient 300 features were selected from the 3000 set of features using the Maximum Information Coefficient (MIC), Ridge regression, and Recursive Feature Elimination (RFE) feature selection methods. The most efficient features extracted by feature selection methods were reclassified by SVM method. The classification success rate increased by approximately 4.54% to 91.10%. In this study, it has been observed that using CNN architectures and feature selection methods are effective in the classification of flower images. [ABSTRACT FROM AUTHOR]

To date, different approaches have been used to be correctly identified of plant species. Leaves are the most important approaches as part of the plants which provide many features with advantages such as shape, color and vein texture. In this study, a new approach based on the geometrical properties of the leaf has been proposed. This method called Edge Step (ES), consists of features such as angle, center-edge length and edge distance by using edge points in the shape boundary curve. In addition, Shearlet Transform method, which has features such as good sensitivity to tissue identification, rapid calculation and directional independence, is used. In addition to these methods, Color features and Gray-Level Co-Occurrence Matrix (GLCM) method to extract color and texture properties from leaf images have been applied. Attributes derived from all these methods were tested with the Extreme Learning Machine (ELM) classifier method as separately and combination. The proposed study has been tested by using four different plant leaf datasets such as Flavia, Swedish, ICL and Foliage. Using these datasets, studies based on texture, shape and color characteristics have been compared with the performance of the proposed approach. As a result, the proposed method is identified to be more successful than the other methods. [ABSTRACT FROM AUTHOR]

Supervised pattern recognition is a technique for classification that the prior knowledge is used regarding member of sample category. Classification model is improved by using samples separated into category as training set. Supervised pattern recognition is getting more important for chemistry, biology, pharmacology and food science. There are many supervised pattern recognition methods. The main part is to select the most appropriate method. There are implementations to different inputs for various purposes such as food quality assessment and data interpretation. Wine, oil, honey, dairy products, meat, fruits, beverages, cereals and fish could be given as examples analyzed by supervised pattern recognition techniques. Also by using this techniques, texture and aroma analyses, food verification, food quality assessment, multiple element analysis, classification based on geographical and botanical origins can be performed. In this review, supervised pattern recognition is defined, its application techniques are summarized, and information is provided by exemplifying studies on pattern recognition techniques used in food analysis. [ABSTRACT FROM AUTHOR]

Nodule characteristics used in the evaluation of lung nodules are generally subjective assessments of theexpert opinions. Among these characteristics, the most known and used one for prediction is the degree of malignancy. In the classification studies in the literature, deep features are used besides the traditional features extracted from the nodule appearance and morphological structure. In this study, traditional features, deep features, and the combinations both used in predicting the nodule characteristics. Four classification algorithms with different structures are evaluated for predicting nodule characteristics. Reference data sets of nodule characteristics were generated by means of majority voting from subjective assessments of radiologists. These generated data sets generally have large unbalanced class distributions. Data balancing procedure has been applied to examine the effect of this condition on classification results. With the combinations of these methods, effects of different classification models on the classification accuracy, sensitivity and specificity are examined. The results of the experiments shown that the classification strategy needs to be specifically determined starting from the used features to the classification algorithm according to the performance criterion to be achieved. [ABSTRACT FROM AUTHOR]

In this study, it is aimed to detect and count wild animals based on image processing in wildlife. From the images obtained from a fixed camera in the protected areas, an image processing based system has been developed for detecting and counting wild animals which are added value with hunting tourism to the country's economy. Through this developed system, it is aimed to both determine and count the wild animal species with high performance. For this, firstly, using gaussian mixed models (GMM) technique, the background images were extracted from the image scenes coming from real-time photocapture videotapes. In a real-time complex image scene that is instantaneous in nature where there is a lot of mobility, developed field test, attribute test and color test criteria are used to determine the targeted wild animal. In the experimental studies, 4 species of wild animals including deer, fox, wolf and wild horses were detected with 100% accuracy. The software's processing time per video frame is 0.242 seconds. With the developed methods, it has been seen that species determination for wild animal inventory can be done with less cost camera systems and computer software without human power with 100% success rate. The difference from the wild animal classification studies in the literature is the catching of the 100% recognition rate targeted by wild animal identification process with simpler mathematical operations and color factor developed without using ready-made algorithms on object recognition. The wild animal recognition algorithms we use in our work are obviously to be the infrastructure for dynamic object detection studies in computer vision applications and all other object recognition will increase the performance ratio in the studies. [ABSTRACT FROM AUTHOR]

Recently, rapid developments in image processing have gained different perspective in deep learning models. Deep learning models continue to contribute to the areas of human health. Pneumonia is one of the diseases that people may encounter in any period of their lives. Pneumonia accounts for about 18% of infectious diseases. In some cases, this disease can cause death. In this study, lung x-ray images were used for the diagnosis of pneumonia. The convolutional neural network (CNN) from deep learning models was used for feature extraction in the resulting image set. The results of CNN with different classifiers were compared. As a result of the comparison, a success rate of approximately 95.8% was obtained with support vector machines (SVM). In the early diagnosis of deadly diseases such as pneumonia, deep learning models were found to be faster and more accurate. This study has shown that feature extraction with CNN provides better results in terms of performance than current methods in biomedical field. [ABSTRACT FROM AUTHOR]

Recent developments in image processing have contributed to the advancement of rapidly developing technological systems. In particular, work on image processing in the health field has further increased its popularity. Whether it is medical images or images on the other side, although the success of existing methods is ensured, the deeper learning model is a model that contributes more in terms of time and performance compared to existing methods. Highperformance results can be obtained on multi-layered images with deep learning model while the existing methods are operated on single layer images. The most important feature of deep learning is that it can process the operations on the image in a single pass and discover the parameters that need to be entered manually. Moreover, as technology companies try to deepen their learning and increase their competitive power among themselves, the methods they have built on deep learning in the scientific sense have begun to be preferred over existing methods. It is envisaged that the data sets in the biomedical field, which is one of the limited access areas of the dataset, to be obtained quickly in the recent times will contribute more to the image processing studies in this area together with the deep learning model. [ABSTRACT FROM AUTHOR]

Two dimensional digital filters which perform spatial filtering or frequency domain filtering are used for image enhancement operations. For high resolution image enhancement operations, two-dimensional digital filters which perform filtering in the frequency domain are preferred due to the computational advantage. However, in frequency domain based two dimensional digital filter design methods, there are various issues such as excess of computational complexity, suboptimal bad filter performance, and lack of elasticity. Moreover, in many design methods proposed in the literature, it is seen that there is no application on the image. The proposed method in this study presents a solution with improved filter performance and elasticity in less computational complexity. The proposed method, in which the effect on image enhancement is also studied, is based on the derivation of a two-dimensional Kaiser-Hamming window family obtained using the Huang's method. The proposed method provides the design of many two-dimensional digital filters with various filter characteristics at a fixed length. Simulation results obtained from various filtering examples for two images show that two dimensional digital filters designed by the proposed method can provide different smoothing and sharpening on a wide scale according to the desired level for image enhancement applications such as image smoothing and image sharpening, and also the length and independent shape parameters of the proposed window have opposite effects on the image. [ABSTRACT FROM AUTHOR]

Nodule characteristics used in the evaluation of lung nodules are generally subjective assessments of theexpert opinions. Among these characteristics, the most known and used one for prediction is the degree of malignancy. In the classification studies in the literature, deep features are used besides the traditional features extracted from the nodule appearance and morphological structure. In this study, traditional features, deep features, and the combinations both used in predicting the nodule characteristics. Four classification algorithms with different structures are evaluated for predicting nodule characteristics. Reference data sets of nodule characteristics were generated by means of majority voting from subjective assessments of radiologists. These generated data sets generally have large unbalanced class distributions. Data balancing procedure has been applied to examine the effect of this condition on classification results. With the combinations of these methods, effects of different classification models on the classification accuracy, sensitivity and specificity are examined. The results of the experiments shown that the classification strategy needs to be specifically determined starting from the used features to the classification algorithm according to the performance criterion to be achieved. [ABSTRACT FROM AUTHOR]

Two dimensional digital filters which perform spatial filtering or frequency domain filtering are used for image enhancement operations. For high resolution image enhancement operations, two-dimensional digital filters which perform filtering in the frequency domain are preferred due to the computational advantage. However, in frequency domain based two dimensional digital filter design methods, there are various issues such as excess of computational complexity, suboptimal bad filter performance, and lack of elasticity. Moreover, in many design methods proposed in the literature, it is seen that there is no application on the image. The proposed method in this study presents a solution with improved filter performance and elasticity in less computational complexity. The proposed method, in which the effect on image enhancement is also studied, is based on the derivation of a twodimensional Kaiser-Hamming window family obtained using the Huang's method. The proposed method provides the design of many two-dimensional digital filters with various filter characteristics at a fixed length. Simulation results obtained from various filtering examples for two images show that two dimensional digital filters designed by the proposed method can provide different smoothing and sharpening on a wide scale according to the desired level for image enhancement applications such as image smoothing and image sharpening, and also the length and independent shape parameters of the proposed window have opposite effects on the image. [ABSTRACT FROM AUTHOR]

Noise reduction is one of the important issues in image processing. An image affected by noise means that information stored in the image is distorted. Therefore, it is necessary to remove such noises especially before the later stages of image processing. Salt and pepper noise which is one of these noises, is mostly caused by error in image acquisition or by the recording errors. In this study, a new method is proposed to eliminate the noises on the picture by detecting the pixels that may contain salt and pepper noise by using k-means algorithm which is one of the clustering algorithm. Performance of the of the proposed method with the conventional methods which are median filter and mean filter are compared at different noise ratios and it is observed that the proposed method increases the performance rate especially as the noise level increases. [ABSTRACT FROM AUTHOR]