摘要
为解决具有高反光特性的工业零部件的视觉检测难题,提出一种基于高动态范围成像(HDRI)的高反光金属表面缺陷检测方法,并针对传统方法中标定相机响应函数速度低的问题,提出一种改进的相机响应函数快速标定方法。首先搭建试验平台采集高反光金属表面序列图像;其次设计快速标定算法标定相机响应函数,获得图像像素值与曝光量之间映射关系;进而将不同曝光时间下的图像序列合成一幅高动态范围图像;最后运用自适应多阈值分割与Haar-like特征提取算法对合成的高动态范围图像进行表面缺陷检测。实验结果表明,文中方法在保证相机响应函数标定准确度的同时,具有计算量小、标定速度快的优点,而且合成的高动态范围图像细节清晰,满足高反光金属表面缺陷检测要求。
A defect detection method for high-reflective-metal surface based on HDRI was proposed to solve the problem of visual inspection of industrial parts with high reflectivity. A fast calibration method of camera response function was also proposed according to the problem of low response rate of calibration camera in traditional method. Firstly, a test platform was built to collect high-reflective-metal surface image sequence.Then, a fast calibration algorithm was designed to calibrate the camera response function to obtain the mapping relationship between the image pixel value and the exposure amount. A high dynamic range (HDR) image was synthesized from the image sequence under different exposure time. Finally, the defect detection for high-reflective-metal surface was carried out for the synthesized HDR image by using adaptive multi-thresholds segmentation and Haar-like feature extraction algorithm. The experimental results show our method has the advantages of small calculation and fast calibration speed while ensuring the calibration accuracy of the camera response function. Moreover, the synthesized HDR images have more detailed information, which meet the requirements of defect detection for high-reflective-metal surface.
作者
冯维
刘红帝
汤少靖
孙国栋
赵大兴
FENG Wei;LIU Hong-di;TANG Shao-jing;SUN Guo-dong;ZHAO Da-xing(School of Mechanical Engineering,Hubei University of Technology,Wuhan 430068,China;Hubei Key Laboratory of Modern Manufacturing Quality Engineering,Wuhan 430068,China)
出处
《仪表技术与传感器》
CSCD
北大核心
2019年第8期112-116,共5页
Instrument Technique and Sensor
基金
国家自然科学基金项目(51675166,51805153)
精密测试及仪器国家重点实验室开放基金资助(pilab1801)
关键词
高动态范围成像
相机响应函数
高反光表面
快速标定
缺陷检测
high dynamic range imaging
camera response function
high-reflective surface
fast calibration
defect detection
