针对近景摄影测量中对编码标记点进行精确定位并对具有唯一编码值的编码标记点进行解码识别的要求，提出了一种环状编码标记点自动检测和识别算法。该算法首先对图像进行canny边缘检测，通过一系列限制条件和计算封闭轮廓质心来过滤噪声和非编码标记点。然后，采用最小二乘椭圆拟合进行编码标记点定位，结合椭圆拟合误差判断分割编码标记点轮廓并进行填充。最后，提出一种将局部同心椭圆变换为平行直线的ALPC (Affine LOG Polar Coordinate)变换，对分割出的编码标记点进行ALPC变换，利用变换后的图像特征进行解码。实验结果表明，利用本文算法对于编码标记点的定位可以达到亚像素级别；且当相机光轴与编码标记点法线夹角小于60°时平均识别准确率能达到98.8%，当夹角达到70°时识别准确率仍然能达到90.2%；在复杂背景下的识别准确率能够达到96.46%。

On the requests of coded targets in the close-range photogrammetry for precision position and decode recognition, an algorithm to automatically detect and recognize the cyclic coded targets was proposed. Firstly, the canny edge detection was used to process images, not only the noises but also the non-coded target were filtered by means of calculating the centroids of closed contours and a series of criteria. Then the least square ellipse fitting was taken to locate the coded target, and each coded target contour could be segmented and filled when the fitting error was satisfied. Finally, a transformation method named Affine LOG Polar Coordinate (ALPC)that has the property of transforming local concentric ellipses into parallel straight lines was proposed to transform each coded target for decoding. Experimental results show that the algorithm presented in this paper can locate the coded targets to a sub-pixel level, and the average recognition accuracy rate can reach 98.8% when the angle between the camera optical axis and the normal of coded target is less than 60° . Furthermore, the recognition accuracy rate can still reach 90.2% when the angle is 70°, and the recognition accuracy rate can reach 96.46% under the complex background. In conclusion, the algorithm proposed can fully satisfy the precision and accuracy requirements of coded target location and recognition in the close-range photogrammetry.