A Novel Image Segment Method for Three-Dimensional Laser Scanning

Exact segment of the laser image is the key of three-dimensional laser scanning in natural lighting condition. Conventional image segment algorithms, however, can hardly deal with the influence of the system noise caused by the intense scattering under natural lighting. In this paper, an image segment algorithm based on the level set was proposed. First, the original two-dimensional profile curve was described by level set. The system model was optimized by solving the partial differential equation. The image segment was finished when the model reach global optimum. This was followed by laser-stripe thinning using the eight neighborhood algorithm. Experimental results showed that high-accuracy and good-robust were obtained using this method.