Abstract:
The computer-based traditional point cloud reconstruction algorithm is not suitable for complicated environment due to its complexity and high cost. In recent years, the embedded point cloud reconstruction method has been receiving more and more attentions, but it also has some shortcomings such as complexity, low computing speed, immature theory and few applications. Aiming at these shortcomings, this paper proposes a three-dimension point cloud reconstruction method by utilizing two-dimension fault information and third-dimension length information. Two-dimension sampling point is firstly obtained by collecting the measurement information of the light curtain sensor at a certain time in the measurement process. And then, Two-dimension sampling point is further converted into three-dimension space coordinates, in which two-dimension sampling points construct the coordinates of each sampling point and two-dimension slice information. The lengths of the object information and the two-dimension point cloud information are obtained at all times. By means of the length of the object information, three-dimension point cloud can be reconstructed in the entire measurement process via two-dimension fault coordinate phase shift. The object of three-dimension point cloud can be displayed, whose contour information has a high degree of recognition. If the object is tilted, the geometric information is used to correct the fault information. This proposed correction method is simple and effective for reducing the operation time of processor and satisfying the rapid application of embedded system. Finally, a dual light curtain measurement system based on FPGA is developed in this paper to verify the proposed algorithm. The experimental results show that the three-dimension fault point cloud reconstruction method can accurately display the point cloud outline of the object. Compared with the computer-based three-dimension point cloud reconstruction, the proposed algorithm is simple, effective, and higher execution efficient.