针对传统光轴角测量系统中人为因素干扰导致的重复性能降低等不足，设计一种基于图像处理技术与高精度电控转台相结合的自动测量方案。分析了照明光斑光强分布特性，提出了一种基于图像区域分割的最优非线性加权的特征点提取算法。该方法根据光强与灰度的线性关系，将光强量化到3个区间。然后分析灰度因素在特征点定位中的影响; 针对不同区间应用局部最优加权法则以实现局部最优化处理。最后，利用光轴偏转角与特征点在像面上位移的对应关系，计算得到光轴角测量值。开展了光斑特征点重复性实验并分析了光轴角检测精度，结果表明: 该系统在不同光强、不同角度条件下重复性好，测角精度优于60.18″。与激光光源的光轴角测量实验对比，该算法在光斑光强分布不均的条件下仍能完成稳定且精确的光轴角测量。

An automatic measurement scheme based on image processing techniques and a high-precision electronic control turntable was designed to improve disadvantages of repetition performance degradation in traditional optic axial angle measurement systems due to human factors. The distribution characters of illumination light spots and light intensity in the system were analyzed. A feature point extraction algorithm for the optimal nonlinear weighting was proposed based on image area segmentation.According to the linear relationship between light intensity and gray level,the light intensity was quantified to three sections.The influences of grey level on feature point location was analyzed. Aiming at different sections, the optimal regional weighting rule was used to implement local optimal processing. Finally, Optic axial angle measurement value was calculated by using the corresponding relationship of displacement between optical axis deflection angle and feature point on image plane. The experiments for optical spot repeatability was performed and the measuring accuracy of optical axial angles was analyzed. The result shows in the experiment that the system under different light intensity and different angles has good repeatability and the angle measurement accuracy is superior to 60.18′′. Compared with optic axial angle measurement experiment by lasers, the algorithm can finish stable and accurate optic axial angle measurement under uneven distribution of light spots and light intensity.