YOLOv3 Network Research Articles

Research on Radar Target Detection Based on the Electromagnetic Scattering Imaging Algorithm and the YOLO Network

In this paper, the radar imaging technology based on the time-domain (TD) electromagnetic scattering algorithm is used to generate image datasets quickly and apply them to target detection research. Considering that radar images are different from optical images, this paper proposes an improved strategy for the traditional You Only Look Once (YOLO)v3 network to improve target detection accuracy on radar images. The speckle noise in radar images can cover the real information of a target image and increase the difficulty of target detection. The attention mechanisms are added to the traditional YOLOv3 network to strengthen the weight of the target region. By comparing the target detection accuracy under different attention mechanisms, an attention module with higher detection accuracy is obtained. The validity of the proposed detection network is verified on a simulation dataset, a measured real dataset, and a mixed dataset. This paper is about an interdisciplinary study of computational electromagnetics, remote sensing, and artificial intelligence. Experiments verify that the proposed composite network has better detection performance.

Enhanced YOLO v3 for precise detection of apparent damage on bridges amidst complex backgrounds

A bridge disease identification approach based on an enhanced YOLO v3 algorithm is suggested to increase the accuracy of apparent disease detection of concrete bridges under complex backgrounds. First, the YOLO v3 network structure is enhanced to better accommodate the dense distribution and large variation of disease scale characteristics, and the detection layer incorporates the squeeze and excitation (SE) networks attention mechanism module and spatial pyramid pooling module to strengthen the semantic feature extraction ability. Secondly, CIoU with better localization ability is selected as the loss function for training. Finally, the K-means algorithm is used for anchor frame clustering on the bridge surface disease defects dataset. 1363 datasets containing exposed reinforcement, spalling, and water erosion damage of bridges are produced, and network training is done after manual labelling and data improvement in order to test the efficacy of the algorithm described in this paper. According to the trial results, the YOLO v3 model has enhanced more than the original model in terms of precision rate, recall rate, Average Precision (AP), and other indicators. Its overall mean Average Precision (mAP) value has also grown by 5.5%. With the RTX2080Ti graphics card, the detection frame rate increases to 84 Frames Per Second, enabling more precise and real-time bridge illness detection.

Style transfer and recognition of infrared images of transformers under different lighting based on CycleGAN

In order to solve the problem that the transformer in the infrared image of the transformer is difficult to be accurately detected due to strong light irradiation under different lighting conditions, a method for unifying the illumination and equipment identification of the infrared image of the transformer based on CycleGAN and yolov3 is proposed. CycleGAN is used to unify the illumination intensity of the infrared image of the transformer, and an adversarial loss function and a cycle consistency loss function are introduced to ensure high-quality generation of infrared images, and the expanded infrared sample library is richer. Finally, the yolov3 network is trained using the uniformly illuminated image set to achieve accurate identification of transformers. The experimental results show that the infrared image of the transformer that has been unified by illumination can identify the transformer with an accuracy of 95.2%, which is 3.1% higher than before enhancement.

Deep learning methods for fully automated dental age estimation on orthopantomograms.

This study aimed to use all permanent teeth as the target and establish an automated dental age estimation method across all developmental stages of permanent teeth, accomplishing all the essential steps of tooth determination, tooth development staging, and dental age assessment. A three-step framework for automatically estimating dental age was developed for children aged 3 to 15. First, a YOLOv3 network was employed to complete the tasks of tooth localization and numbering on a digital orthopantomogram. Second, a novel network named SOS-Net was established for accurate tooth development staging based on a modified Demirjian method. Finally, the dental age assessment procedure was carried out through a single-group meta-analysis utilizing the statistical data derived from our reference dataset. The performance tests showed that the one-stage YOLOv3 detection network attained an overall mean average precision 50 of 97.50 for tooth determination. The proposed SOS-Net method achieved an average tooth development staging accuracy of 82.97% for a full dentition. The dental age assessment validation test yielded an MAE of 0.72years with a full dentition (excluding the third molars) as its input. The proposed automated framework enhances the dental age estimation process in a fast and standard manner, enabling the reference of any accessible population. The tooth development staging network can facilitate the precise identification of permanent teeth with abnormal growth, improving the effectiveness and comprehensiveness of dental diagnoses using pediatric orthopantomograms.

High-low level task combination for object detection in foggy weather conditions

For the object detection task in foggy weather conditions, image dehazing network is often used as preprocessing method to get a clear input. However, there is not strictly a strong positive correlation between image dehazing task and object detection task. Moreover, the preprocessing module can increase the inference time of the whole model to a certain extent. To alleviate these problems, we propose a novel High-Low level task combination network (HLNet) based on multitask learning, which can learn both high-level and low-level tasks. Specially, instead of restoring the features to clear pixel-wise feature space like common image dehazing method, we opt to perform a restoration in feature level to mitigate the influence of the Batch Normalization (BN) layer of encoder on dehazing task. HLNet jointly learn dehazing task and detection task in an end-to-end fashion, which ensures that the weather-specific information in latent feature space is suppressed. Moreover, we applied the HLNet framework on three different object detection networks, including RetinaNnet, YOLOv3 and YOLOv5s network, and achieved improvements of 1.7 percent, 2.3 percent, and 1.2 percent in mAP respectively. The experimental results demonstrate the effectiveness and generalization ability of our proposed HLNet framework in real foggy scenarios.

Object Recognition and Grasping for Collaborative Robots Based on Vision.

This study introduces a parallel YOLO-GG deep learning network for collaborative robot target recognition and grasping to enhance the efficiency and precision of visual classification and grasping for collaborative robots. First, the paper outlines the target classification and detection task, the grasping system of the robotic arm, and the dataset preprocessing method. The real-time recognition and grasping network can identify a diverse spectrum of unidentified objects and determine the target type and appropriate capture box. Secondly, we propose a parallel YOLO-GG deep vision network based on YOLO and GG-CNN. Thirdly, the YOLOv3 network, pre-trained with the COCO dataset, identifies the object category and position, while the GG-CNN network, trained using the Cornell Grasping dataset, predicts the grasping pose and scale. This study presents the processes for generating a target's grasping frame and recognition type using GG-CNN and YOLO networks, respectively. This completes the investigation of parallel networks for target recognition and grasping in collaborative robots. Finally, the experimental results are evaluated on the self-constructed NEU-COCO dataset for target recognition and positional grasping. The speed of detection has improved by 14.1%, with an accuracy of 94%. This accuracy is 4.0% greater than that of YOLOv3. Experimental proof was obtained through a robot grasping actual objects.

Vehicle target detection method based on improved YOLO V3 network model

For the problem of insufficient small target detection ability of the existing network model, a vehicle target detection method based on the improved YOLO V3 network model is proposed in the article. The improvement of the algorithm model can effectively improve the detection ability of small target vehicles in aerial photography. The optimization and adjustment of the anchor box and the improvement of the network residual module have improved the small target detection effect of the algorithm. Furthermore, the introduction of the rectangular prediction frame with orientation angles into the model of this article can improve the vehicle positioning efficiency of the algorithm, greatly reduce the problem of wrong detection and missed detection of vehicles in the model, and provide ideas for solving related problems. Experiments show that the accuracy rate of the improved algorithm model is 89.3%. Compared to the YOLO V3 algorithm, it is improved by 15.9%. The recall rate is improved by 16%, and the F1 value is also improved by 15.9%, which greatly increased the detection efficiency of aerial vehicles.

A Real-Time Shipping Container Accident Inference System Monitoring the Alignment State of Shipping Containers in Edge Environments

Along with the recent development of artificial intelligence technology, convergence services that apply technology are undergoing active development in various industrial fields. In particular, artificial intelligence-based object recognition technologies are being widely applied to the development of intelligent analysis services based on image data and streaming video data. As such, in the port yard, these object recognition technologies are being used to develop port safety services in smart ports. Accidents are a frequent occurrence in port yards due to misaligned loading of ship containers. In order to prevent such accidents, various studies using artificial intelligence technology are underway. In this paper, we propose a real-time shipping container accident inference edge system that can analyze the ship container’s loading status from a safety point of view to prevent accidents in advance. The proposed system includes the collection of video data of the ship container, inferring the safety level of the alignment status of the ship container, and transmitting the inference results for the safety level. In this paper, the proposed inference model is implemented with YOLOv3, YOLOv4 and YOLOv7 networks and can be used in video monitoring to realize the accurate classification and positioning of three different safety levels (safe, caution, and danger) in real time. In the performance evaluation, the detection accuracy of the inference model implemented with the YOLOv4 network was greater than 0.95. Its performance was also significantly better than that of the inference model implemented with the YOLOv3 and YOLOv7 networks. Although it was slightly inferior to the YOLOv4 network in terms of the accuracy, the inference model implemented with the YOLOv3 network had a faster inference speed than the model implemented with the YOLOv4 and YOLOv7 networks. Because of the port safety scenario, in which the inference accuracy is more important than the inference speed, we applied the YOLOv4 algorithm to the inference model of the system.

RSA based improved YOLOv3 network for segmentation and detection of weed species

RSA based improved YOLOv3 network for segmentation and detection of weed species

Based on YOLO v3 Target Recognition Algorithm as Vehicle Tracking Algorithm Analysis

Traditional traffic information acquisition and acquisition are mainly implemented by sensors, and these traditional acquisition and acquisition systems have some great drawbacks. However, with the popularity of traffic monitoring, computer vision technology gradually has a platform foundation that can be applied to identify and track traffic conditions. In this paper, through the research of traditional and Deep learning-based multi-target recognition algorithms and two common multi-target tracking algorithms, a solution of YOLO v3 network combined with deep-sort algorithm is proposed. In this paper, a video of traffic information of urban roads is directly collected for areas with relatively large traffic flow. Interval frames are extracted from the video data set to make relevant data sets for training and verification of YOLO v3 neural networks. Combined with the test results, an open source vehicle depth model dataset is used to train the vehicle depth feature weight file, and Deep-SORT algorithm is used to achieve the target tracking, which can realize the real-time and more accurate multi-target recognition and tracking of moving vehicles.

The Design and Implementation of the Vision-based Water Surface Garbage Cleaning Vehicle

This paper proposes a water surface garbage cleaning vehicle. Its main functions include the detection, tracking and collection of garbage. The object detection algorithm based on the YOLOv3 network is applied to embedded systems. The sliding mode control is designed to guarantee the stability of the tracking process of the unmanned vehicle. Experiments and comparative simulation results are given to show the effectiveness of the proposed solution.

Determination of Vickers Hardness in D2 Steel and TiNbN Coating Using Convolutional Neural Networks

The study of material hardness is crucial for determining its quality, potential failures, and appropriate applications, as well as minimizing losses incurred during the production process. To achieve this, certain criteria must be met to ensure high quality. This process is typically performed manually or using techniques based on analyzing indentation image patterns produced through the Vickers hardness technique. However, these techniques require that the indentation pattern is not aligned with the image edges. Therefore, this paper presents a technique based on convolutional neural networks (CNNs), specifically, a YOLO v3 network connected to a Dense Darknet-53 network. This technique enables the detection of indentation corner positions, measurement of diagonals, and calculation of the Vickers hardness value of D2 steel treated thermally and coated with Titanium Niobium Nitride (TiNbN), regardless of their position within the image. By implementing this architecture, an accuracy of 92% was achieved in accurately detecting the corner positions, with an average execution time of 6 seconds. The developed technique utilizes the network to detect the regions containing the corners and subsequently accurately determines the pixel coordinates of these corners, achieving an approximate relative percentage error between 0.17% to 5.98% in the hardness results.

Locating Multiple Standard Planes in First-Trimester Ultrasound Videos via the Detection and Scoring of Key Anatomical Structures

This study was aimed at developing a first-trimester standard plane detection (FTSPD) system that can automatically locate nine standard planes in ultrasound videos and investigating its utility in clinical practice. The FTSPD system, based on the YOLOv3 network, was developed to detect structures and evaluate the quality of plane images by using a pre-defined scoring system. A total of 220 videos from two different ultrasound scanners were collected to compare detection performance between our FTSPD system and sonographers with different levels of experience. The quality of the detected standard planes was quantitatively rated by an expert according to a scoring protocol. Kolmogorov-Smirnov analysis was used to compare the distributions of scores across all nine standard planes. The expert-rated scores indicated that the quality of the standard planes detected by the FTSPD system was on par with that of the planes detected by senior sonographers. There were no significant differences in the distributions of the scores across all nine standard planes. The FTSPD system performed significantly better than junior sonographers in five standard plane types. The results of this study suggest that our FTSPD system has significant potential for detecting standard planes in first-trimester ultrasound screening, which may help to improve the accuracy of fetal ultrasound screening and facilitate early diagnosis of abnormalities. The quality of the standard planes selected by junior sonographers can be significantly improved with the assistance of our FTSPD system.

A Machine Learning based Approach for Detection of Osteoarthritis using Thermal Images

Osteoarthritis (OA) of the knee is a common disorder that contributes to physical decline and activity limitation. Early OA diagnosis and treatment can stop the disease's progression. The assessment of a physician's visual examination is impartial, subject to different interpretations, and highly dependent on their level of experience. Therefore, a system that employs machine learning techniques to automatically determine the degree of knee OA has been proposed in this study. At first, a specifically created one stage YOLOv2 network is employed to estimate the size of the kneecap according to the distribution of knee joints in low contrast thermal images. To be more specific, the knee Kellgren-Lawrence (KL) grading assignment is ordinal; therefore, a harsher penalty is provided for misrepresentation with a larger gap between the anticipated and actual KL grade. A machine learning architecture is then constructed and extensive tests are performed to demonstrate how texture properties affect diagnostic performance. Thermal images are used to determine if they might be used to distinguish between radiographs of diseased and healthy knees. The outcomes of machine learning features and manually extracted features are compared. Finally, a stacked model that combines second-level machine learning with predictions of patellar texture and clinical characteristics is provided.

Utilizing Image Processing and the YOLOv3 Network for Real-Time Traffic Light Control

In this study, different strategies used to count vehicles and people in different image areas at a street intersection were analyzed to obtain counts at appropriate times suitable for real-time control of a traffic light. To achieve this, video recordings of cameras placed at the intersection were used to test and verify image processing algorithms and deep learning using the YOLOv3 network implemented on a 4 GB RAM Jetson Nano card. We counted the vehicles and people that stopped and crossed the polygons to delimit the different areas of interest, with a maximum error of ±2 in the validation tests for all cases. In addition, as a strategy, we combined the images from both cameras into a single one, thereby allowing us to make a single detection and subsequently determine if they are inside or outside the polygons used in separating the areas of interest with the respective counts. Furthermore, this enabled us to obtain information on vehicles and people stopped and crossing in a time of 0.73 s on average. Hence, it was established that the inclusion of the control algorithm is appropriate for real-time control of traffic lights.

Cotton Stubble Detection Based on Improved YOLOv3

The stubble after cotton harvesting was used as the detection object to achieve the visual navigation operation for residual film recovery after autumn. An improved (You Only Look Once v3) YOLOv3-based target detection algorithm was proposed to detect cotton stubble. First, field images of residual film recycling were collected. Considering the inconsistency between stubble size and shape, a segmented labeling data set of stubble is proposed. Secondly, the Darknet-53 backbone of the original YOLOv3 network is improved to accommodate tiny targets. Next, the prediction anchor box of the improved detection backbone is clustered using K-means++, and the size of the prediction anchor box suitable for improved YOLOv3 is determined. Finally, for the false detection points after detection, a mean value denoising method is used to remove the false detection points. Feature points are extracted from the denoised stubble, and the candidate points are fitted by the least square method to obtain the navigation line. The optimal model with a mean average precision (mAP) of 0.925 is selected for testing at the test stage. The test results show that the algorithm in this article can detect the stubble of residual film recovery images at different locations, different time periods, and different camera depression angles without misdetection. The detection speed of a single image is 98.6 ms. Based on an improvement over YOLOv3, the improved model has a significantly higher detection rate in different scenarios than YOLOv3. This can provide practical technical support for the visual navigation of residual film recovery.

Remote Sensing Image Object Detection Based on Improved YOLOv3 in Deep Learning Environment

A deep learning-based method, improved YOLOv3 algorithm is proposed in the deep learning environment to tackle challenges such as big scale, uneven distribution, largescale variation, and complicated background of small- and medium-sized remote sensing photos. This manufacture uses Densenet as the backbone, replacing Darknet-53 to realize feature reuse and make the feature extraction more effective; introduces the spatial pyramid pooling module into the feature pyramid part for increasing the receptive field and isolating the most prominent contextual features; adds SE attention module in the process of feature extraction and obtains richer features by learning more location information and channel information from the images. Under DOTA dataset, the final results are that the mean Average Precision value is 86.78%, which is 4.16% higher than the baseline YOLOv3 network. The model put forward makes it easier to extract information from the feature map and achieve higher detection accuracy without influencing the real-time performance of detection.

Detail extraction of mechanical component image based on depth learning

Aiming at the problem of feature extraction, recognition and classification of machining features of various mechanical part models, an image detail extraction method based on deep learning architecture is proposed. We use the median filtering method to filter the impulse noise in the image initially. After horizontal tilt correction, the contour image is divided into several fan sub regions. Then, the structural parameters of YOLOV3 network are optimized to make it more suitable for small target classification detection of industrial parts on the basis of fully retaining the feature extraction ability of depth convolution network. The simulation results show that for original YOLOv3 algorithm, the improved network model has good robustness, and the problems of different scales, low feature utilization and occlusion in part detection which are obviously improved. Compared with the traditional methods of feature extraction and classification of mechanical part images, the deep learning strategy also shows greater advantages.

Reflective Safety Clothes Wearing Detection in Hydraulic Engineering Using YOLOv3-CCD

The construction site of the hydraulic engineering has a high danger factor and the correct wearing of reflective safety clothes ensures the safety of workers. Therefore, the detection and testing of reflective safety clothes wearing is an important task in the construction site of a hydraulic engineering. However, the traditional manual supervision strategy has the problems of low efficiency, narrow scope, and poor real-time performance in complex working conditions. Based on YOLOv3 that is a classical target detection model, this paper proposes a reflective safety clothes detection algorithm (YOLOv3-CCD) based on an attention mechanism and an improved loss function. By adding the CA (Coordinate Attention) mechanism module to the backbone network, the characterization ability of the target feature is enhanced, so as to solve the Long-Term dependencies problem in the detection process; The loss function is changed from IOU-Loss (Intersection Over Union Loss) to CIOU-Loss (Complete-IOU Loss), so that the network takes the aspect ratio into consideration when selecting the prediction box, which improves the accuracy of target positioning; In the post-processing of the algorithm, we improved NMS (Non-Maximum Suppression) to solve the problem of dense target detection being missed. Experimental results show that compared with the original YOLOv3 network model, the algorithm has stronger robustness and the overall detection accuracy is 1.8% higher than that of the original network. Moreover, the detection speed is 32 frames per second, which is faster than the original network.

Dynamic Target Tracking Control of Autonomous Underwater Vehicle Based on Trajectory Prediction.

Underwater dynamic target tracking technology has a wide application prospect in marine resource exploration, underwater engineering operations, naval battlefield monitoring, and underwater precision guidance. Aiming at the underwater dynamic target tracking problem, an autonomous underwater vehicle tracking control method based on trajectory prediction is studied. First, a deep learning-based target detection algorithm is developed. For the image collected by the multibeam forward-looking sonar image, this algorithm uses the YOLO v3 network to determine the target in a sonar image and obtain the position of the target. Then, a time profit Elman neural network (TPENN) is constructed to predict the trajectory information of the dynamic target. Compared with an ordinary Elman neural network, its accuracy of dynamic target prediction is increased. Finally, underwater tracking of the dynamic target is realized using the model predictive controller (MPC), and the tracking result is stable and reliable. Through simulations and experiment, the proposed underwater dynamic target tracking control method is demonstrated to be effective and feasible.