欢迎访问《图学学报》
分享到：
摘 要：针对现有红外图像存在噪声干扰与边缘模糊等问题，提出一种基于方向高频子带各层间系数相关性的红外图像增强算法。利用各层间小波Contourlet系数相关性的不同，合理对图像进行去噪和边缘优化处理。提出一种适用于文章算法的阈值选择方法，从而得到更为精确的去噪和边缘优化效果。仿真结果表明，实验结果与预期目标相符，能获得相比传统增强算法更高的PSNR值与PAI值。#br# 关 键 词：红外图像；系数相关性；小波Contourlet；去噪；边缘优化
Abstract: The existing infrared image has the problems of noise and edge blur. This article proposes an infrared image enhancement algorithm using wavelet Contourlet intrastate coefficients dependency. Reasonable denoising and edge optimization process is used to the wavelet Contourlet coefficient between different layers. A threshold selection method is proposed for the algorithm to obtain a more accurate de-noising and edge optimization results. Simulation results are consistent with target effects. The PSNR and PAI values obtained are higher than other traditional enhancement algorithm.#br# Key words: infrared image; coefficient of correlation; wavelet Contourlet; de-noising; Edge optimization
摘要： 针对现有红外图像存在噪声干扰与边缘模糊等问题，提出一种基于方向高频子带各层间系数相关性的红外图像增强算法。利用各层间小波Contourlet系数相关性的不同，合理对图像进行去噪和边缘优化处理。提出一种适用于文章算法的阈值选择方法，从而得到更为精确的去噪和边缘优化效果。仿真结果表明，实验结果与预期目标相符，能获得相比传统增强算法更高的PSNR值与PAI值。
