为提高自动准直效率, 结合以太网, 设计出一套新型光路自动准直控制和检测方案.该方案采用基于耦合矩阵的前馈补偿及图像Jacobian矩阵的近远场串并行准直方式, 并引入局部自适应阈值和二值化算法, 同时将模糊控制算法运用到步进电机的调整过程中, 提高光束近远场图像的处理精度, 降低系统的准直时间.实验结果表明, 该准直系统远场的平均调整误差为空间滤波器小孔直径0.44%, 能够满足准直系统远场调整精度(小于小孔直径5%)的要求, 准直时间由原有的30 mins缩短至12 mins左右.

In order to improve the efficiency of automatic alignment,combined with Ethernet, a new optical path automatic alignment control and detection scheme was describled. The proposed near-far field parallel alignment method implements forward feed back compensation based on coupling matrix and Jacobian image matrix. By introducing the image local adaptive thresholding, image binaryzation, and the fuzzy stepping motor control algorithm, the precision of near-far field image detection was improved with less processing time. The results show that the average far-field alignment error is smaller than 0.44% of spatial filter pinhole diameter, which meets the alignment system requirements(5% of spatial filter pinhole diameter), the alignment time reduces from the original 30 mins to 12 mins.