Camera lens defect enhancement algorithm based on improved gamma transform
【摘要】：An image enhancement algorithm based on local entropy and local standard deviation to improve gamma transform was proposed to solve the problem of camera lens blue spot defect which is difficult to be recognized by traditional methods. It can achieve the goal of enhancing the contrast of blue spot defects. The concepts of local entropy and local standard deviation are introduced in order to measure the dispersion of gray level in the neighborhood of an image. The local entropy and local standard deviation are transformed by power law with an exponent greater than 1 to obtain the gamma value required by the gamma transform. The contrast of blue spot defect in camera lens is enhanced by gamma transform. In order to highlight the position information of the blue spot defect in the image, the gray projection method is applied to the image enhanced by the improved gamma transformation to judge whether there is a defect. Finally, compared with the common image enhancement algorithms, the results show that the detection accuracy of the improved gamma transform for blue spot defects is more than 93%, which is about 10% higher than that of most common image enhancement algorithms.