The finishing process of special steel tubes and bars requires bundling and unbundling operations,while manual unbundling has poor environment and high strength.A method for identifying the seam and wire of the bundle was proposed based on line structured light.A binarization threshold algorithm was proposed based on the number of pixels of each luminance value of the image,and the binarization threshold was obtained by calculating the number of pixels that satisfied the luminance to reduce the influence of ambient light.For the problem that the maximum binding seam position could not be determined through the view field of a single image, a laser image mosaic algorithm based on arc vertex and endpoint features was proposed.Finally,a bundle number and state identification algorithm based on Hough transform and connected domain analysis was proposed to achieve accurate identification of the bundle characteristics of pipe and bar bundles with different sizes at the production site.The experimental results show that the proposed binarization threshold algorithm has a higher recognition rate to laser streaks compared with the traditional algorithms,and the speed is increased by 17%,the detection error of the maximum binding seam and wire identification algorithm is less than 3 mm.It fully meet the actual production needs.