A defect detection method of gear end-face based on modified YOLO-V3

Most deep-learning-based object detection algorithms exhibit low speeds and accuracy in gear surface defect detection due to their high computational costs and complex structures. To solve this problem, a lightweight model for gear surface defect detection, namely STMS-YOLOv5, is proposed in this paper. Firstly, the ShuffleNetv2 module is employed as the backbone to reduce the giga floating-point operations per second and the number of parameters. Secondly, transposed convolution upsampling is used to enhance the learning capability of the network. Thirdly, the max efficient channel attention mechanism is embedded in the neck to compensate for the accuracy loss caused by the lightweight backbone. Finally, the SIOU_Loss is adopted as the bounding box regression loss function in the prediction part to speed up the model convergence. Experiments show that STMS-YOLOv5 achieves frames per second of 130.4 and 133.5 on the gear and NEU-DET steel surface defect datasets, respectively. The number of parameters and GFLOPs are reduced by 44.4% and 50.31%, respectively, while the mAP@0.5 reaches 98.6% and 73.5%, respectively. Extensive ablation and comparative experiments validate the effectiveness and generalization capability of the model in industrial defect detection.

Section: Related Workmentioning

confidence: 99%

STMS-YOLOv5: A Lightweight Algorithm for Gear Surface Defect Detection

Yan,

Zhang,

Bai

et al. 2023

“…Among these models, R-CNN, SSP-NET and Faster R-CNN have two detection stages, with high accuracy but much slower computing speed than YOLO and SSD models with primary structures. YOLO (You Look Only Once) includes YOLO, YOLOv3 [27][28][29][30][31], YOLOv4 [32] and YOLOv5 [33]. Other methods are favored by researchers because they could directly train the target position in single-stage operation.…”

Section: Introductionmentioning

confidence: 99%

A Real-Time Zanthoxylum Target Detection Method for an Intelligent Picking Robot under a Complex Background, Based on an Improved YOLOv5s Architecture

Huang

et al. 2022

The target recognition algorithm is one of the core technologies of Zanthoxylum pepper-picking robots. However, most existing detection algorithms cannot effectively detect Zanthoxylum fruit covered by branches, leaves and other fruits in natural scenes. To improve the work efficiency and adaptability of the Zanthoxylum-picking robot in natural environments, and to recognize and detect fruits in complex environments under different lighting conditions, this paper presents a Zanthoxylum-picking-robot target detection method based on improved YOLOv5s. Firstly, an improved CBF module based on the CBH module in the backbone is raised to improve the detection accuracy. Secondly, the Specter module based on CBF is presented to replace the bottleneck CSP module, which improves the speed of detection with a lightweight structure. Finally, the Zanthoxylum fruit algorithm is checked by the improved YOLOv5 framework, and the differences in detection between YOLOv3, YOLOv4 and YOLOv5 are analyzed and evaluated. Through these improvements, the recall rate, recognition accuracy and mAP of the YOLOv5s are 4.19%, 28.7% and 14.8% higher than those of the original YOLOv5s, YOLOv3 and YOLOv4 models, respectively. Furthermore, the model is transferred to the computing platform of the robot with the cutting-edge NVIDIA Jetson TX2 device. Several experiments are implemented on the TX2, yielding an average time of inference of 0.072, with an average GPU load in 30 s of 20.11%. This method can provide technical support for pepper-picking robots to detect multiple pepper fruits in real time.

“…(2) The regression-based target detection framework is dominated by the “you only look once” (yolo) series [ 18 , 19 , 20 , 21 ] and the single shot multibox detector (SSD) [ 22 ], which streamlines the feature extraction process to obtain a faster speed, but with accuracy slightly lacking in the same period of development. Combined with specific module design, this one-stage approach can often be efficiently applied to defect detection [ 23 ].…”

Section: Introductionmentioning

confidence: 99%

A Lightweight Deep Network for Defect Detection of Insert Molding Based on X-ray Imaging

Wang

Huang

2021

Aiming at the abnormality detection of industrial insert molding processes, a lightweight but effective deep network is developed based on X-ray images in this study. The captured digital radiography (DR) images are firstly fast guide filtered, and then a multi-task detection dataset is constructed using an overlap slice in order to improve the detection of tiny targets. The proposed network is extended from the one-stage target detection method of yolov5 to be applicable to DR defect detection. We adopt the embedded Ghost module to replace the standard convolution to further lighten the model for industrial implementation, and use the transformer module for spatial multi-headed attentional feature extraction to perform improvement on the network for the DR image defect detection. The performance of the proposed method is evaluated by consistent experiments with peer networks, including the classical two-stage method and the newest yolo series. Our method achieves a mAP of 93.6%, which exceeds the second best by 3%, with robustness sufficient to cope with luminance variations and blurred noise, and is more lightweight. We further conducted ablation experiments based on the proposed method to validate the 32% model size reduction owing to the Ghost module and the detection performance enhancing effect of other key modules. Finally, the usability of the proposed method is discussed, including an analysis of the common causes of the missed shots and suggestions for modification. Our proposed method contributes a good reference solution for the inspection of the insert molding process.