Research On Defect Detection Method For High Reflective Surface Based On HDRI

At present,the brightness contrast in the real visual environment is far beyond the limits of the dynamic range of imaging sensors and ordinary digital display devices.The grayscale images collected from the vision imaging system that cannot be used for display the brightness contrast required for higher dynamic range.At the same time,due to the constraints of the current level of hardware development,the images collected by ordinary CCD / CMOS sensors cannot truly record the natural scene information,and there is often a global or local overshoot on the target surface with high reflective characteristics exposure phenomenon.Therefore,how to capture,synthesize and display target surface visual information with high reflection characteristics and high brightness difference by using ordinary digital display devices has become a hot issue in the field of digital image processing.In order to solve the problem of visual inspection of target surfaces with high reflection characteristics,this paper systematically studies the High Dynamic Range Imaging(HDRI)for improving the dynamic range in the image processing and machine vision inspection.In this paper,the proposed algorithm can effectively synthesize a floating point brightness image with a wide dynamic range,which has good effects in improving color saturation and rich image details,and solving the problem of overexposure on highly reflective surfaces,bringing more for authentic sensory experience.The main research work of this paper is as follows:1.A fast camera response curve calibration method is designed.Aiming at the problem of more sampling points and longer processing time in the traditional algorithm,a fast calibration algorithm of Camera Response Function(CRF)combining least square method and S-type function model was proposed.The number of sampling points required by the fast calibration algorithm is 1/10 of the traditional algorithm,and the running time of the program is about 1/4 of the traditional algorithm.It has the advantages of small calculation amount and fast calibration speed.2.A HDRI method based on sparse coding is proposed.A weighted structured sparse coding method is used to design a sparse coding model and process the image background layer and the reflection layer separately,and finally reconstruct the output target image.In the traditional HDRI technology based on multiple exposures,the problems of large data volume and complex imaging process were effectively solved,which greatly reduced the program running time.3.An improved gradient domain adaptive tone mapping algorithm based on color correction model is proposed.The dynamic range of the HDR image is well compressed.At the same time,a color correction model is introduced,and a linear interpolation function and a color correction factor between the chromatic and achromatic colors are combined to realize the function of adaptively adjusting the chromaticity information of different regions of the image.Through a large number of experiments and tests,theresearch scheme proposed in this paper has good feasibility.The improved algorithm proposed has fast processing speed,short running time,good objective evaluation effect and meets the requirements of high reflective surface defect detection.