Up to now, micro milling has been recognized as the most promising technique to mitigate the surface damage of large-aperture KDP crystal used in laser fusion facility. However, there are quantities of damages on the surface of large-aperture KDP crystal to be mitigated. Furthermore, the efficiency and accuracy of manual tool setting is relatively low, and it is easy to damage KDP crystal surface due to mis-operation. Therefore, an automatic tool setting technology for KDP crystal surface damage mitigation based on machine vision is proposed in this research. The automatic tool setting includes fast pre tool setting achieved by image ranging and final tool setting achieved by micro-chip detection. In the process of fast pre tool setting, real-time image of the micro milling cutter is obtained, and the distance between the tool tip and the KDP crystal surface is obtained by image ranging. Finally, the tool tip is quickly positioned 20 μm away from the KDP crystal surface. In the process of final tool setting, the micro milling cutter is retreated to a distance of about 50 μm away from the KDP crystal surface before detecting micro-chips. Therefore, the micro-chips covered by the micro milling cutter will be exposed and appear in the view field of microscope, so as to improve the micro-chip recognition rate. The results showed that the total time of automatic tool setting is 130 s, and the accuracy of automatic tool setting is better than 1.359 μm. Compared with manual tool setting, the automatic tool setting based on machine vision proposed in this research improves the safety, efficiency, and accuracy of independently developed mitigation system for KDP crystal surface damage.

迄今为止，微铣削已被认为是减轻激光聚变设备中使用的大口径KDP晶体表面损伤的最有前途的技术。但是，大孔径KDP晶体表面的损伤程度可以减轻。此外，手动工具设置的效率和准确性相对较低，并且容易由于误操作而损坏KDP晶体表面。因此，本研究提出了一种基于机器视觉的KDP晶体表面损伤自动对刀技术。自动工具设置包括通过图像测距实现的快速预工具设置以及通过微芯片检测实现的最终工具设置。在快速进行预刀具设定的过程中，可以获得微型铣刀的实时图像，通过图像测距获得刀尖与KDP晶体表面之间的距离。最后，将刀尖快速定位在距KDP晶体表面20μm的位置。在最终的工具设置过程中，在检测微芯片之前，将微铣刀退回到距KDP晶体表面约50μm的距离。因此，被微铣刀覆盖的微芯片将被暴露并出现在显微镜的视野中，从而提高了微芯片的识别率。结果表明，自动对刀时间为130 s，自动对刀精度达到1.359μm。与手动对刀相比，本研究提出的基于机器视觉的自动对刀提高了安全性，效率，