Search

After apple fruit maturation, the optimal harvest period is short, and the picking robot is expected to improve harvesting efficiency. While it is common for apples to be overlapped and occluded by branches and leaves, which pose challenges to the robot's apple harvesting. Therefore, precise and swift identification and localization of the target fruit is crucial. To this end, this paper proposes a lightweight apple detection method, YOLOv5s-ShuffleNetV2-DWconv-Add, or "YOLOv5s-SDA" for short. The red and green apple datasets in natural environment were collected by a mobile phone, which were divided into four categories: red and green apples that can be directly grasped and cannot be directly grasped, in order to avoid damage to the robotic arm. Different deep learning object detection models were compared, with the YOLOv5s algorithm providing superior recognition performance. To improve harvest efficiency and portability of hardware devices, modifications are made to the YOLOv5s algorithm, replacing the Focus, C3, and Conv structures within the backbone with 3 × 3 Conv structures and ShuffleNetV2, removing SPP and C3 structures; substituting the C3 in the Neck portion with DWConv modules; and replacing two Concat layers in the PANet structure with smaller computational Add layers. Results demonstrate that the model achieved a mAP of 94.6% on the test set, doubled the detection speed, and compressed the model weight to 11.8% of its original value, while maintaining model accuracy. This new method exhibits promising performance in fruit target recognition in natural scenes, providing an effective means of visual acquisition for fruit picking robots. [ABSTRACT FROM AUTHOR]

In response to the current challenges of numerous background influencing factors and low detection accuracy in the open railway foreign object detection, a real-time foreign object detection method based on deep learning for open railways in complex environments is proposed. Firstly, the images of foreign objects invading the clearance collected by locomotives during long-term operation are used to create a railway foreign object dataset that fits the current situation. Then, to improve the performance of the target detection algorithm, certain improvements are made to the YOLOv7-tiny network structure. The improved algorithm enhances feature extraction capability and strengthens detection performance. By introducing a Simple, parameter-free Attention Module for convolutional neural network (SimAM) attention mechanism, the representation ability of ConvNets is improved without adding extra parameters. Additionally, drawing on the network structure of the weighted Bi-directional Feature Pyramid Network (BiFPN), the backbone network achieves cross-level feature fusion by adding edges and neck fusion. Subsequently, the feature fusion layer is improved by introducing the GhostNetV2 module, which enhances the fusion capability of different scale features and greatly reduces computational load. Furthermore, the original loss function is replaced with the Normalized Wasserstein Distance (NWD) loss function to enhance the recognition capability of small distant targets. Finally, the proposed algorithm is trained and validated, and compared with other mainstream detection algorithms based on the established railway foreign object dataset. Experimental results show that the proposed algorithm achieves applicability and real-time performance on embedded devices, with high accuracy, improved model performance, and provides precise data support for railway safety assurance. [ABSTRACT FROM AUTHOR]

Synthetic Aperture Radar (SAR) finds widespread applications in environmental monitoring, disaster management, ship surveillance, and military intelligence. However, existing target detection methods are ineffective in SAR scenes due to the intricate background environments, small target displays, and irregular appearances. To address these challenges, this thesis introduces a target detection model named YOLOv8-FDF, tailored for SAR scenes based on the YOLOv8 architecture. The model effectively incorporates the FADC module to distinguish targets from complex backgrounds and integrates a deformable feature adaptive mechanism to focus on irregular targets. Additionally, this thesis devised a specialized detection head designed to identify small targets in SAR-wide scenes, thereby improving the effectiveness of detecting such targets. The proposed YOLOv8-FDF model is evaluated on the HRSID dataset. Experiment results show a 3.6% improvement in Map75 on both the training and test sets. Furthermore, under the COCO standard, the model achieves improvements of 4.1%, 2.9%, and 5.5% on AP, AP50, and AP75, along with 6.8%, 1.2%, and 1.2% improvements on small, medium, and large-sized ship detection. An accuracy enhancement of 6.8%, 1.0%, and 14.9% is achieved. These experimental findings validate the efficacy of the proposed YOLOv8-FDF model in SAR scenarios.

Gesture recognition nowadays has grabbed the attention of researchers as they represent human behaviour in multiple practical ways. Amongst a variety of gestures available, hand gestures play an essential role in the field of human‐computer interaction when recognised efficiently in complex and dynamic environments. In this paper, we propose a dynamic hand gesture recognition system to recognise hand gestures appearing in different indoor and outdoor environments. Hand detection and tracking uses a two‐level system resulting in the formation of gesture trajectory in challenging conditions in which existing detection and tracking algorithms could not do so. A set of 45 features is provided as input to the various classification techniques. The redundancy problem has been reduced by selecting a set of optimum features using the analysis of variance method, which ranks the list of features. An incremental feature selection technique calculates recognition accuracy by selecting features according to rankings. This system provides an accuracy of 96.32% when used with machine learning and 97.5% when used with deep learning techniques. Recognition accuracy is calculated for various environments, including an extra hand, multiple persons in the video frame, and outdoor environment. All machine‐learning classifiers are combined using classifier combination to calculate the accuracy according to the majority‐voting rule. Based on the experimental results, it has been observed that deep learning provides better results compared to machine learning. [ABSTRACT FROM AUTHOR]

Target identification is one of the most popular radar uses in real life. Target identification is a classifier that analyzes whether a signal contains an echo from a target (target-present) or is merely noise (target-absent). Deep learning techniques are a popular topic in classification, and they have evinced to be effective in a range of applications. In this paper, a 64 layers Circular Disk type RADAR Target Detection (CDRTD) model is proposed based on Transfer Learning using the SqueezeNet architecture of Convolutional Neural Network (CNN) that functions directly with processed radar target return eco signal and minimize the requirement of conventional laborious radar signal processing. Further, the proposed 64 layers SqueezeNet-based CNN CDRTD model was then implemented to identify circular disk type targets in complex environment. Finally, the target return eco data was tested to identify the circular disk type radar target in complex environments. We further analyzed target detection probability, false alarm rate, precision, recall, F1 in a complex environment and compared it with the ideal case. We found that our proposed CDRTD model can classify 83.3% of the test samples correctly with an overall accuracy of 94.59% in a noisy and cluttered environment whereas 100% of the test samples are classified correctly with an overall accuracy of 100% in an ideal environment. [ABSTRACT FROM AUTHOR]

• A novel method using deep learning-based amplification and tracking is proposed to measure the micro-vibration. • This method can be used under complex light fields with different illumination and backgrounds. • The target object is first magnified and tracked, and then converted into actual displacement values. A new deep learning-based vision measurement method is proposed to accurately measure the micro-vibration displacements of objects in different illuminations and backgrounds. The measurement method pre-process the video, then the deep learning correlation methods are used to zoom in the target object and track the vibration trajectory, and the pixel displacement is converted to actual displacement by pixel equivalents. By comparing the three sets of experiments, the proposed method has exceptional accuracy. When measuring vibration displacement of 0.1 mm, the Root Mean Square Error (RMSE) and Normalized Root Mean Square Error (NRMSE) are 0.0234 mm and 11.8601 %. By comparing with the Rectangle Detection Algorithm and the Template Matching Algorithm, the proposed algorithm outperforms these two traditional methods, especially for the complex environments. It can be concluded that this method, as a new visual measurement method, can be adapted to a variety of complex environments and can accurately measure micro-amplitude vibrations. [ABSTRACT FROM AUTHOR]

Aiming at the problem that the network models of traditional intelligent single-pixel imaging methods have poor suppression ability in complex environments (e.g., smoke, fog, haze, etc.), a single-pixel imaging method based on random discarding mechanism (DSPINet) is proposed. The method integrates compressed sensing and deep learning. It compresses and samples the image directly, discards part of the information, and reconstructs the image with a small number of the one-dimensional detection signal. In view of the fact that most of the data of traditional single-pixel imaging methods are obtained in a controlled environment and do not have good generalization ability, this paper introduces the random discarding mechanism into the network model to simulate the light scattering loss and system noise in the complex environment. The test results show that when the simulation test's sampling rate is 10.00 %, the proposed DSPINet model is compared with the DLGI (improved), DLBGI, Bsr2-Net, and TVAL3, the structural similarity (SSIM) of reconstructed images is improved by 2.20 %, 4.49 %, 5.68 %, and 63.16 %, respectively. In the actual test of introducing the scattering medium, when the sampling rate is only 3.00–10.00 %, the reconstructed images generated by the DSPINet model have a good ability to suppress the interference and show a high reconstruction accuracy in the complex environment. [ABSTRACT FROM AUTHOR]