视觉测量中环状编码标记点检测与识别方法

针对视觉测量中环状编码标记点检测识别易受成像角度、光照等因素影响的问题,提出了一种基于编码环采样的解码识别方法。使用Canny算子对图像进行边缘检测,提取图像轮廓结构信息,基于尺寸和形状约束,滤除非标记点轮廓;使用Otsu算法进一步提取标记点边缘环状邻域,通过最小二乘椭圆拟合求解椭圆参数,通过椭圆参数约束,再次滤除非标记点轮廓;基于标记点椭圆参数,对编码环进行连续采样,解算获得编码值。试验结果表明,所提出的检测识别算法受成像角度、光照等因素影响小,可实现编码标记点的精确定位与准确识别。

Detection and Recognition of Circular Coded Targets in Vision Measurement

To realize the detection and recognition of circular coded targets in vision measurement that can be easily affected by imaging angle and illumination,a novel method based on coded ring sampling is proposed.Canny detector was used to extract the image edges.With geometric constraints,such as size and shape,non-target edges were filtered.Otsu algorithm was employed to further obtain the circular neighborhood of targets,and their ellipse parameters were solved by least square ellipse fitting.Based on the fitted ellipse parameters,non-target edges were filtered further,and the coded values of targets are solved by sampling their coded rings.Experimental results demonstrate that the proposed method can locate and identify the coded targets accurately,and it is robust to the change of imaging angle and illumination.