Intelligent System for Estimation of the Spatial Position of Apples Based on YOLOv3 and Real Sense Depth Camera D415

Andriyanov, Nikita; Khasanshin, Ilshat; Utkin, Daniil; Gataullin, Timur; Ignar, S.; Шумаев, В В; Soloviev, Vladimir

doi:10.3390/sym14010148

Cited by 50 publications

(18 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The speed of the image processing and the quality of the object classification are greatly improved when images are subjected to preprocessing steps. The image-to-image conversion is implemented using image filters [31][32][33][34], thresholding operations [29,30,[35][36][37], morphological operations [31,38], and artificial neural networks [39,40].…”

Section: Theoretical Backgroundmentioning

confidence: 99%

Deep Learning Method for Recognition and Classification of Images from Video Recorders in Difficult Weather Conditions

et al. 2022

View full text Add to dashboard Cite

The sustainable functioning of the transport system requires solving the problems of identifying and classifying road users in order to predict the likelihood of accidents and prevent abnormal or emergency situations. The emergence of unmanned vehicles on urban highways significantly increases the risks of such events. To improve road safety, intelligent transport systems, embedded computer vision systems, video surveillance systems, and photo radar systems are used. The main problem is the recognition and classification of objects and critical events in difficult weather conditions. For example, water drops, snow, dust, and dirt on camera lenses make images less accurate in object identification, license plate recognition, vehicle trajectory detection, etc. Part of the image is overlapped, distorted, or blurred. The article proposes a way to improve the accuracy of object identification by using the Canny operator to exclude the damaged areas of the image from consideration by capturing the clear parts of objects and ignoring the blurry ones. Only those parts of the image where this operator has detected the boundaries of the objects are subjected to further processing. To classify images by the remaining whole parts, we propose using a combined approach that includes the histogram-oriented gradient (HOG) method, a bag-of-visual-words (BoVW), and a back propagation neural network (BPNN). For the binary classification of the images of the damaged objects, this method showed a significant advantage over the classical method of convolutional neural networks (CNNs) (79 and 65% accuracies, respectively). The article also presents the results of a multiclass classification of the recognition objects on the basis of the damaged images, with an accuracy spread of 71 to 86%.

show abstract

Section: Theoretical Backgroundmentioning

confidence: 99%

Deep Learning Method for Recognition and Classification of Images from Video Recorders in Difficult Weather Conditions

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Intel real sense d435i camera, which is developed by Intel Corporation(Santa Clara, California, USA), was used to provide depth image and RGB image. Pyrealsene2 module and the OpenCV library were utilized to develop software for real-time and convenient visualization of processing results [23]. The main structure of Intel Real Sense D435i is shown in Figure 9: It integrates two IR Stereo Cameras, an IR Projector and a Color Camera.…”

Section: Fruit Detection Software Developmentmentioning

confidence: 99%

Multi-Growth Period Tomato Fruit Detection Using Improved Yolov5

Yang

Han

et al. 2022

Int. J. Robot. Autom. Technol.

View full text Add to dashboard Cite

In agricultural mechanized production, in order to ensure the efficiency of hand-eye cooperative operation of tomato picking robot, the recognition accuracy and speed of multi-growth period tomato fruit is an important basis. Therefore, in order to improve the recognition speed of multi-growth period tomato fruit while ensuring or improving the accuracy, this paper improves the Yolov5s model by adding the architecture of the lightweight mobilenetv3 model. Firstly, the deep separable convolution is replaced in the backbone network of Yolov5s, which reduces the amount of convolution operation. Secondly, the linear bottleneck inverse residual structure is fused to obtain more features in high-dimensional space and perform convolution operation in low-dimensional space. Third, the attention mechanism is inserted into the last layer of the network to highlight features and improve accuracy. The research results show that the recognition accuracy of the improved Yolov5 model remains above 98%, the CPU recognition speed is 0.88f·s-1 faster than Yolov5s, and the GPU recognition speed is 90 frames per second faster than Yolov5s. Finally, a set of the recognition software system of multi-growth period tomato fruit is designed and developed by using RealSense D435i depth camera and PYQT. The software system further verifies the feasibility of the improved Yolov5 model, and lays a foundation for the visual software design of agricultural picking robot picking recognition.

show abstract

“…Kou et al designed an anchor-free feature selection mechanism and a dense convolution structure and introduced them into the YOLOv3 network to improve the detection speed and accuracy of the algorithm in detecting steel strip surface defects [13]. Andriyanov et al proposed an intelligent system for the estimation of the spatial positions of apples based on YOLOv3 and a D415 RealSense Depth Camera, which obtained the position estimates of the apples with high accuracy in a symmetric coordinate system [14]. Tulbure et al undertook a comprehensive analysis of modern object detection models that can be used for defect detection applications in industry and analyzed the applicable detection models under different leading constraints, providing an important reference for industrial defect detection [15].…”

Section: Introductionmentioning

confidence: 99%

Surface Defect Detection Model for Aero-Engine Components Based on Improved YOLOv5

Wang

et al. 2022

Applied Sciences

View full text Add to dashboard Cite

Aiming at the problems of low efficiency and poor accuracy in conventional surface defect detection methods for aero-engine components, a surface defect detection model based on an improved YOLOv5 object detection algorithm is proposed in this paper. First, a k-means clustering algorithm was used to recalculate the parameters of the preset anchors to make them match the samples better. Then, an ECA-Net attention mechanism was added at the end of the backbone network to make the model pay more attention to feature extraction from defect areas. Finally, the PANet structure of the neck network was improved through its replacement with BiFPN modules to fully integrate the features of all scales. The results showed that the mAP of the YOLOv5s-KEB model was 98.3%, which was 1.0% higher than the original YOLOv5s model, and the average inference time for a single image was 2.6 ms, which was 10.3% lower than the original model. Moreover, compared with the Faster R-CNN, YOLOv3, YOLOv4 and YOLOv4-tiny object detection algorithms, the YOLOv5s-KEB model has the highest accuracy and the smallest size, which make it very efficient and convenient for practical applications.

show abstract

Intelligent System for Estimation of the Spatial Position of Apples Based on YOLOv3 and Real Sense Depth Camera D415

Cited by 50 publications

References 32 publications

Deep Learning Method for Recognition and Classification of Images from Video Recorders in Difficult Weather Conditions

Deep Learning Method for Recognition and Classification of Images from Video Recorders in Difficult Weather Conditions

Multi-Growth Period Tomato Fruit Detection Using Improved Yolov5

Surface Defect Detection Model for Aero-Engine Components Based on Improved YOLOv5

Contact Info

Product

Resources

About