Search

This paper proposes a method for identifying the bicycle riding environment using only onboard equipment. Initially, a fundamental subsystem is established for identifying the bicycle riding environment, and its functionality is validated. The findings indicate that the subsystem, utilizing an open-source trained model, can detect riding on roadways but not on sidewalks. Consequently, we emphasize the need for transfer learning, specifically using sidewalk viewpoint images, to enable the identification of bicycle riding environments. Subsequently, we conduct bicycle riding environment identification by employing a transfer learning model with manually labeled training data. The results demonstrate that after transfer learning, sidewalk riding detection, which was previously unachievable, becomes feasible. The identification rate was over 80%. Furthermore, we develop four riding environment identification algorithms, including the transfer learning model, and compare their performance across various road environments and riding conditions. Consequently, it is established that the region of interest (ROI) extension identification algorithm exhibits the highest identification performance (93% on average). As a result, this paper contributes valuable insights into the realization of bicycle riding environment identification, particularly in the context of detecting traffic violations within the riding safety support information system.

In this study, we conducted bicycle riding experiments that focus on the method of information presentation by an automated system to cyclists, with the intent to design an information system related to bicycle safety information. In an evaluation of information presentation, we confirmed that auditory presentation was the most effective from the viewpoints of both recognition and comprehension. In addition, information presented using the auditory method was more likely to be recognized than information presented using the visual and tactile method combined. The results of the analysis of variance for recognition showed that the visual presentation method had no significant effect. In addition, the experiments involved behavioral observation using the system and a questionnaire survey on changes in behavior before and after experiencing the system. The results showed that information was correctly conveyed by the system using a combination of auditory, tactile, and visual presentation methods, regardless of the users' degree of awareness of the system. Two weeks after the experiment, a survey was conducted to investigate the changes in the participants' awareness and behavior regarding risky or illegal riding. The results showed an improvement in the awareness of risky or illegal riding in daily life and an increase in right behavior, confirming the effectiveness of this system. Consequently, the results demonstrate the effectiveness of the system and provide knowledge for realizing a bicycle riding safety support information system.

In recent years, the agricultural field has been confronting difficulties such as the aging of farmers, a shortage of workers, and difficulties for new farmers. Harvesting time prediction has the potential to solve these problems, and is expected to effectively utilize human resources, save labor, and reduce labor costs. To achieve harvesting time prediction, various works are being actively conducted. Methods for harvesting time prediction using meteorological information such as temperature and solar radiation, etc., and methods for harvesting time prediction using neural networks based on color information from fruit bunch images are being investigated. However, the prediction accuracy is still insufficient, and the harvesting time prediction for individual tomato fruits has not been studied. In this study, we propose a novel method to predict the harvesting time for individual tomato fruits. The method uses Mask R-CNN to detect tomato bunches and uses two types of ripeness determination to predict the harvesting time of individual tomato fruits. The experimental results showed that the accuracy of the prediction using the ratio of R values was better for the harvesting time prediction of tomatoes that are close to the harvesting time, and the accuracy of the prediction using the average of the differences between R and G in RGB values was better for the harvesting time prediction of tomatoes that are far from the harvesting time. These results show the effectiveness of the proposed method.

Due to an increasing need for face recognition under poor lighting conditions, near infrared (NIR) face recognition based on deep convolutional neural networks (DCNN) has become an active area of research. However, in NIR face images of eyeglasses wearers, reflected light is generated around the eyes due to active NIR light sources, and it is one of the main contributors to performance degradation in NIR face recognition. In addition, there have to date been no attempts to lighten DCNN models for NIR face recognition. To solve these problems, we propose a DCNN-based fast NIR face recognition system which is robust to reflected light. This work has two main contributions: 1) We generated synthetic face images of individuals with and without eyeglasses using our proposed CycleGAN-based Glasses2Non-glasses (G2NG) data augmentation. We then constructed an augmented training database by adding the synthetic images, and the database helps to make the NIR face recognition system robust against reflected light. 2) A lightweight NIR FaceNet (LiNFNet) architecture was developed to reduce the computational complexity of the proposed system by adapting the depthwise separable convolutions and linear bottlenecks to VGGNet 16. The proposed architecture reduces the computation required, while improving the performance of NIR face recognition. Through the experiments reported in this paper, we verified that the proposed G2NG data augmentation improved the face recognition validation rate by 99.09% for NIR face images which have the reflected light from eyeglasses. Also, LiNFNet reduces the number of multiplication operations by 4.4×109 compared with VGGNet 16.

Personal mobility devises become more and more popular last years. Gyroscooters, two wheeled self-balancing vehicles, wheelchair, bikes, and scooters help people to solve the first and last mile problems in big cities. To help people with navigation and to increase their safety the intelligent rider assistant systems can be utilized that are used the rider personal smartphone to form the context and provide the rider with the recommendations. We understand the context as any information that characterize current situation. So, the context represents the model of current situation. We assume that rider mounts personal smartphone that allows it to track the rider face using the front-facing camera. Modern smartphones allow to track current situation using such sensors as: GPS / GLONASS, accelerometer, gyroscope, magnetometer, microphone, and video cameras. The proposed rider assistant system uses these sensors to capture the context information about the rider and the vehicle and generates context-oriented recommendations. The proposed system is aimed at dangerous situation detection for the rider, we are considering two dangerous situations: drowsiness and distraction. Using the computer vision methods, we determine parameters of the rider face (eyes, nose, mouth, head pith and rotation angles) and based on analysis of this parameters detect the dangerous situations. The paper presents a comprehensive related work analysis in the topic of intelligent driver assistant systems and recommendation generation, an approach to dangerous situation detection and recommendation generation is proposed, and evaluation of the distraction dangerous state determination for personal mobility device riders.

We developed a method for the precise estimation of the 3D trajectory of a baseball by modeling the movement of the baseball and estimating the capture delay, using multiple unsynchronized cameras. To develop the proposed algorithm, we mimicked the real-world process of capturing a baseball in simulation space, and analyzed the capture process using a multiple unsynchronized camera system. We represented the movement of the baseball using a piece-wise spline model, and predicted the position of the baseball in the subframes in a manner which is robust to position error and change in direction of movement of the baseball. This method accurately predicts the baseball position over time by modeling the movement of the baseball in a real baseball game environment, and improves the accuracy of the reconstructed 3D baseball trajectories. We defined an objective function to estimate the capture delay, and estimate the optimal capture delay parameter using non-linear optimization method. In addition, we evaluated the performance of the proposed method in simulation space and in a real-world situation. The experimental results show that the proposed method can estimate a 3D baseball trajectory precisely using a multiple unsynchronized camera system and is robust to variations in capture delay, both in the simulation space and in real-world situations.

In recent years, environmental information monitoring in the agricultural field has become an important issue. There is an increasing demand for meteorological information in local areas such as a rice field, a greenhouse, etc., owned by an agricultural worker. Conventional research has been actively conducted on weather stations in local areas. However, weather stations that are inexpensive, highly accurate, and have achieved stable measurements indoors and outdoors for long periods of time (over a year) are not reported. In addition, there is a lack of research that simultaneously acquires weather information, stores weather information, and provides weather information to farmers. These three functions are important in the agricultural field. In this paper, we discuss the development of a meteorological observation device, the construction of a cloud server for storing meteorological information, and the provision of information to users. First, we develop the novel meteorological observation device (KOSEN-Weather Station), which applies a simple Aßmann’s aspiration psychrometer for highly accurate temperature and humidity measurements. To evaluate the reliability of KOSEN-WS, we compare the weather information measured by KOSEN-WS with that of WXT520. As a result, it is shown that KOSEN-WS is viable. Then, KOSEN-WS is installed in the field, and the stability and durability of KOSEN-WS are examined. As a result, the KOSEN-WS has been operating stably over 19 months and provides weather information to users. Then, it is shown that the KOSEN-WS is able to operate continuously under the environment of −16.5 °C to 44.9 °C. Next, for the storage of meteorological information, we construct the cloud server. Then, a webpage is created to provide easy-to-understand weather information to farmers. Furthermore, to prevent damage to crops, if the current temperature is lower than the set temperature, or if the current temperature is higher than the set temperature, an alert is sent to the farmers. As a result, the system is highly evaluated by agricultural workers and JA staff. From the above results, the effectiveness of this system is shown.

To ensure the safety of a handle-type electric wheelchair (hereinafter, electric wheelchair), this paper describes the applicability of using a Kinect sensor. Ensuring the mobility of elderly people is a particularly important issue to be resolved. An electric wheelchair is useful as a means of transportation for elderly people. Considering that the users of electric wheelchairs are elderly people, it is important to ensure the safety of electric wheelchairs at night. To ensure the safety of an electric wheelchair at night, we constructed a hazardous object detection system using commercially available and inexpensive Kinect sensors and examined the applicability of the system. We examined warning timing with consideration to the cognition, judgment, and operation time of elderly people. Based on this, a hazardous object detection area was determined. Furthermore, the detection of static and dynamic hazardous objects was carried out at night and the results showed that the system was able to detect with high accuracy. We also conducted experiments related to dynamic hazardous object detection during daytime. From the above, it showed that the system could be applicable to ensuring the safety of the handle-type electric wheelchair.