摘要
以往检测技术因为在检测前采用有效方法减小系统误差,导致检测结果精度较低。研究人工智能技术的光学高精度检测技术,采用软阈值方法去除采用光学探测器检测光学元件损伤点中心位置的系统误差,获取去除系统误差的光学元件损伤点训练样本图像,基于灰度—梯度对二维最大熵原理分割光学元件损伤点训练样本图像中损伤点特征,利用多目标优化图像特征识别算法提取分割后光学元件损伤点特征并发送至支持向量机中,选取支持向量机预测的前d个窗口特征与待检测光学元件损伤点图像共同发送至二级支持向量机中,二级支持向量机将最终检测结果输出至显示窗口实现高精度光学元件损伤点检测。实验结果表明,采用该方法检测光学元件损伤点精度高,满足精密光学元件损伤点高精度检测要求。
In the past,the effective methods are used to reduce the system error before detection,resulting in low accuracy of detection results.The optical high-precision detection technology of artificial intelligence technology is researched.The soft threshold method is used to remove the system error of damage center point of optical elements detected by optical detector.And the training sample image of optical element damage point is obtained to remove the system error.Based on the gray-gradient pair,the two-dimensional maximum entropy principle is used to segment the damage point characteristics in training sample image of optical element damage point,and the multi-objective optimized image feature recognition algorithm is used to extract damage point feature of optical element after segmentation and send it to the support vector machine.The window features before d supporting vector machine and the damage point image of optical element to be detected are selected and sent to the secondary support vector machine.The secondary support vector machine outputs the final detection result to the display window to realize high-precision optical element damage detection.Experimental results show that this method has high accuracy in detecting damage points of optical elements and meets the requirements of high-precision detection of damage points of precision optical elements.
作者
王超
刘艳辉
张玉霞
王保民
WANG Chao;LIU Yanhui;ZHANG Yuxia;WANG Baomin(Randan University,Handan Hebei 056005,China)
出处
《激光杂志》
北大核心
2020年第9期47-51,共5页
Laser Journal
基金
邯郸市科技局规划项目(No.1721202044-3)。
关键词
人工智能
技术
光学
高精度
检测技术
研究
artificial intelligence
technology
optics
high precision
detection technology
research
