Vehicle wheel weld detection based on improved YOLO v4 algorithm

Abstract

In recent years, vision-based object detection has made great progress across different fields. For instance, in the field of automobile manufacturing, welding detection is a key step of weld inspection in wheel production. The automatic detection and positioning of welded parts on wheels can improve the efficiency of wheel hub production. At present, there are few deep learning based methods to detect vehicle wheel welds. In this paper, a method based on YOLO v4 algorithm is proposed to detect vehicle wheel welds. The main contributions of the proposed method are the use of k-means to optimize anchor box size, a Distance-IoU loss to optimize the loss function of YOLO v4, and non-maximum suppression using Distance-IoU to eliminate redundant candidate bounding boxes. These steps improve detection accuracy. The experiments show that the improved methods can achieve high accuracy in vehicle wheel weld detection (4.92 % points higher than the baseline model with respect to AP75 and 2.75 % points higher with respect to AP50). We also evaluated the proposed method on the public KITTI dataset. The detection results show the improved method’s effectiveness.

MCA-YOLOv7: An Improved UAV Target Detection Algorithm Based on YOLOv7

“…At present, the application scenarios based on deep learning target detection are very wide. In 2022, LT Jiao et al [18] proposed a wheel weld detection method based on the YOLOv4 algorithm, which improved the detection accuracy by optimizing the loss function and anchor frame. T Liang et al [19] proposed a traffic sign detection method for automatic driving scenes in 2022.…”

GAN-Based Image Dehazing for Intelligent Weld Shape Classification and Tracing Using Deep Learning

“…The input image is fed into the backbone layer which is responsible for convolutional down sampling to extract the features. In this work, CSPDarknet53 is used as the backbone for object detection, which has 53 convolutional layers with high accuracy [36]. The dense block contains multiple convolution layers starting from 13 × 13 × 512 as the input layer X o and finally 13 × 13 × 1024 as the output transition layer.…”

“…The path aggregation network (PANNet) is added for the aggregation of features and image segmentation, which preserves the spatial information present in the images [38]. Here, three anchors- [12,16,19,36,40,28], [36,75,76,55,72,146] and [142,110,192,243,459, 401]-were used and each bounding box predicted the offset from the top corner of each image (c x , c y ), as well as B w width, B h height and probability (confidence) score c. The governing equations for bounding boxes are as follows:…”

Automated vehicle wheelbase measurement using computer vision and view geometry

“…Due to the rapid development of artificial intelligence, many deep learning methods based on convolutional neural networks (CNNs) have been proposed. Among various of studies, deep learning methods have shown impressive performance in object recognition, target tracking, and semantic segmentation (Girshick et al 2014, Girshick 2015, Ren et al 2017, Varia et al 2018, Liang et al 2022, Li et al 2022, Wang et al 2022a, Liu et al 2022a. For example, to detect multiple objects in different regions of an image, Girshick et al (2014) proposed combining regions with CNNs to design a regionbased CNN (R-CNN).…”

“…, Varia et al (2018 proposed an algorithm based on generative adversarial networks for road extraction in images, which effectively solves the problems caused by insufficient datasets. With the YOLO algorithm, in 2022, Liang et al (2022) established an architecture to automatically detect and locate vehicular wheels and welds by using the modified YOLOv4. Note that there are various versions of the YOLO algorithm (Li et al 2022, Wang et al 2022a, and in this study the most stable and applicable algorithm, YOLOv5 (Liu et al 2022a), is employed.…”