机器视觉中激光光带两步提取算法研究

Extraction of Structured Light Method in Machine Vision

在基于线激光投射的三维探测系统中,如何快速精确地从复杂背景中快速提取出投射在障碍物表面的激光光带,并在光条图像中提取光条中心位置是实时精密探测出障碍物的关键问题。论文对当前已有的线激光光条提取算法进行综述,并比较了现有算法的优缺点。最后,提出了一种两步提取算法。第一步是将RGB与HSV色彩进行空间变换初步处理提取出激光光带,第二步结合阈值法与重心法两算法优点精确地确定光条中心位置。实验说明,激光光条中心的两步提取算法精准、快速,达到了像素级,这种算法的抗噪声能力也很强。

In the line structured light three-dimensional measurement system based on trigonometry, how quickly and accurately from the complicated background of rapid extraction of structured light, and extracting the light image light strip Center location is key to real-time precision distance measuring problems. Based on the analysis of line-structured laser light source selection, environmental noise and other factors affecting the light extraction based on to the already existing line- structured laser light extraction method are reviewed, and the strengths and weaknesses of existing algorithms are compared. Finally, this paper presents a preliminary extraction of structured light HSV color space, then uses the threshold and light Center of gravity method to determine two-step extraction algorithm. This algorithm converts the RGB space into HSV space for initial processing to extract structured light, then the thresholds and focus are combined to accurately determine centre position of the light. Experimental results show that the algorithm can accurately extract stripe Center, and have a strong a- bility to withstand noise, sub-pixel precision.