An Improved Wavelet Filtering Method for Welding Seam Images in a Complex Environment

The quality of welding seam images is affected by noise during the process of acquisition and propagation. The traditional image-filtering method results in the loss of the edge and detail information of the welding seam image in the denoising process. In view of these problems, in this paper, we proposed an improved wavelet filtering method for welding seam images in complex environments. Based on wavelet filtering, the proposed method processes the high-frequency and low-frequency subbands of the image. The improved filtering method can better remove the noise of welding seem images. By using a wall-climbing robot equipped with an industrial camera to capture the weld image, various traditional filtering methods and the proposed filtering method were used to denoise the welding seam image. A comparison and analysis of the experimental results revealed that the proposed an improved wavelet filtering method for weld images in complex environments is better than the traditional filtering method. This study provides guidance for accurate weld identification and defect detection.