Weld Seam Tracking and Detection Robot Based on Artificial Intelligence Technology

• DOI: 10.3390/s23156725
• Publication Date: 2023-07-28T06:10:45Z
• Authors: Jiuxin Wang; Lei Huang; Jiahui Yao; Man Liu; Yurong Du; Minghu Zhao; Yaoheng Su; Dingze Lu
• Keywords: Robot welding; Tracking (education)

The regular detection of weld seams in large-scale special equipment is crucial for improving safety and efficiency, this can be achieved effectively through the use seam tracking robots. In study, a wall-climbing robot with integrated was designed, wall climbing function realized via permanent magnet array Mecanum wheel. using DeepLabv3+ semantic segmentation model. Several optimizations were implemented to enhance deployment model on embedded devices. Mobilenetv2 used replace feature extraction network original model, convolutional block attention module mechanism introduced into encoder module. All traditional 3×3 convolutions substituted depthwise separable dilated convolutions. Subsequently, welding path fitted least squares method based results. experimental results showed that volume improved reduced by 92.9%, only being 21.8 Mb. average precision reached 98.5%, surpassing 1.4%. reasoning speed accelerated 21 frames/s, satisfying real-time requirements industrial detection. successfully realizes autonomous identification seams. This study remarkably contributes development automatic intelligent technologies.