论文部分内容阅读
针对双目立体视觉测距系统的实时性要求,提出了一种改进的激光光斑中心定位算法,并在FPGA上进行了验证。该算法首先对图像进行阈值判定排除干扰光线,然后用Sobel算子对图像进行边缘检测得到边缘图像,最后再利用改进的圆心拟合算法进行中心定位。基于FPGA的硬件系统利用合理的流水线设计提高了数据处理的并行性,可以实时处理大量的数据,克服了传统算法在精度和速度上的不足。中心定位算法结果用于立体匹配,效果良好,完全满足设计需要。
According to the real-time requirements of binocular stereo vision ranging system, an improved laser spot centering algorithm is proposed and validated on FPGA. The algorithm first thresholds the image to exclude the interference light, and then uses the Sobel operator to detect the edge of the image to obtain the edge image. Finally, the improved center-fitting algorithm is used to locate the center. FPGA-based hardware system using a reasonable pipeline design to improve the parallelism of data processing, real-time processing of large amounts of data to overcome the lack of accuracy and speed of the traditional algorithm. The results of the centering algorithm are used for stereo matching, the effect is good, fully meet the design needs.