基于侧抑制原理的局部分数阶微分图像增强

摘要

针对目前分数阶微分在图像增强应用中没有充分根据图像区域特征体现侧抑制的方向性,提出从图像灰度变化的8个方向出发,即x轴的正负方向,y轴的正负方向,主对角线的正负方向,次对角线的正负方向,检测出灰度变化的主要方向,然后根据检测结果设计不同的非对称的分数阶滤波器对图像进行滤波.这样结合了图像本身的特征信息,能够达到更好的图像增强效果.仿真实验结果表明,改进后的分数阶微分滤波器对增强图像纹理细节具有很好的效果.%So far,the fractional differential based image enhancement methods do not reflect sufficiently the directionality of lateral inhibition by means of the characteristics of image area.The main eight directions of image gray value variety are detected to meet the directionality of lateral inhibition,eight directions are both the positive and negative directions of x-axis and y-axis,that of principal diagonal and secondary diagonal.Then,the differential unsymmetrical fractional filters are designed to enhance image by the detection results.As a result of using the inherent feature information of the processed image,the proposed method is more advantageous than the other image enhancement methods.The simulation results show that the improved fractional differential filter is very effective to enhance image texture detail.