Abstract: Aiming at the problem of low contrast in weld images of large metal components in the nuclear industry, an image enhancement algorithm based on point cloud data and improved power law transformation function was proposed.Firstly, a three-dimensional point cloud image of the weld was obtained by scanning with a line structured light camera, and an improved piecewise power law transformation function was used to nonlinearly enhance the point cloud image, so that defects above and below the surface of the base metal were significantly enhanced. Subsequently, the min-max normalization method was used to normalize the point cloud z-coordinates to the range 0, 255 and convert them into depth images. Finally, the contrast-limited histogram equalization method was used to enhance the depth image to balance the brightness while improving the image contrast. The results show that this algorithm can effectively reduce the impact of uneven ambient light sources on images. While maintaining the balance of image brightness, it can still highlight image details and improve image contrast. Compared with histogram equalization, Laplace transform, gamma transform and contrast-limited histogram equalization algorithms, the algorithm in this paper performs better in terms of edge strength, mean pixel and information entropy.