Research on vision detection technology based on photometric calibration

The camera is affected by illumination changes and vignetting effect during the shooting process, resulting in the lack of brightness information, which affects the accuracy of target detection. In order to reduce the impact of illumination changes and vignetting effect on image quality, obtain images closer to the real scene, and improve the accuracy of target detection, this paper conducts photometric calibration on the image from two aspects: the calibration of CRF (Camera Response Function) and the compensation function of vignetting effect. This paper first uses the N-degree polynomial model to construct the camera response function, proposes a functional model of vignetting compensation based on the position of the pixel point, and finally combines the first two experimental results and conducts experimental verification on the data set before and after the photometric calibration on different algorithms of target detection. The experimental results show that the application of photometric calibration in the field of machine vision can improve the adaptability of machine vision to illumination changes and vignetting effect to a certain extent. The self-made data set is trained and tested, when IoU is 0.7, the mAP calculated by the test set after photometric calibration is improved by 2.3% compared with the original data set.