【摘要】 在磷酸二氢钾(KDP)及磷酸二氘钾(DKDP)晶体损伤研究中需要对图像中损伤点进行统计，为了解决损伤图像中弱小损伤点的低对比度导致统计不准确的问题，提出一种基于局部直方图灰度自适应增强的损伤点检测方法。首先利用图像差分去除高频背景，然后根据图像中每个像素邻域的灰度与方差和整幅图像的灰度与方差之间的差异性，筛选出待增强像素，并通过邻域最大灰度值与全局最大灰度值得出灰度调节系数，对图像进行自适应增强，最后对增强后的图像进行阈值分割及目标分离。实验结果表明该增强算法可以使弱小损伤点的信噪比得到提升，更利于检测出图像中的弱小损伤点，提高DKDP晶体损伤研究中损伤点统计的准确性。更多还原

【Abstract】 In KDP&DKDP crystal damage studies, it is necessary to count the damage points in the image. In order to solve the problem of statistical inaccuracy caused by low contrast of weak damage points in damaged images, damage points detection method based on local histogram grayscale adaptive enhancement was proposed. Firstly, the high-frequency background was removed by image difference, and secondly, according to the difference between gray scale and variance of each pixel neighborhood in the image and the gray and variance of the entire image, the pixels to be enhanced was screened out, and through the neighborhood maximum gray scale and the global maximum gray scale to get the gray adjustment coefficient, the image was adaptively enhanced, finally, the enhanced image was segmented by threshold and target separation. The experimental results show that the SNR of weak damage points can be improved by the enhancement algorithm, which is more conducive to detecting weak damage points in the image and improving the accuracy of DKDP crystal damage statistics.更多还原