基于视觉感知与梯度域的遥感图像对比度增强变分模型
Variational model for contrast enhancement of remote sensing images based on perception and gradient domain

基于视觉感知增强变分模型与梯度域增强变分模型,提出了一种新的遥感图像对比度增强变分模型.首先,定义梯度增强项为一个高斯增强函数,该函数利用高斯滤波器对图像进行预处理,以克服梯度对噪声敏感的不足,并根据图像中各点梯度信息自适应地选择保持或者放大原图像的梯度信息.然后,将梯度增强项引入到视觉感知增强模型中,以提高图像对比度并保持更多细节信息.最后,利用梯度下降流法最小化模型的能量泛函并采用数值化方法获取最优解.从全局和局部对比度增强两个方面验证了所提模型的有效性.实验结果表明,相对于现有其他增强变分模型,局部对比度增强模型能够取得更好的主观视觉效果和客观性能评价指标.
A novel variational model for contrast enhancement of remote sensing images is proposed based on the perceptual-based variational model and the gradient-based variational model. First, the Gaussian enhancement function is designed for the gradient enhancement term. This function uses the Gaussian filter to preprocess the image to overcome the shortage that the gradient is sensitive to noise, and it can keep or enhance the gradient information adaptively according to the gradient information of the source image. Then, the gradient enhancement term is introduced into the perceptual-based enhancement model to improve the contrast and preserve more detail information of the source image. Finally, the gradient descent flow method is applied to minimize the energy functional and the numerical scheme is used to acquire the optimization solution. The effectiveness of the proposed model is verified from global and local contrast enhancement. The experimental results show that compared with other existing variational enhancement models, the local contrast enhancement model can achieve better subjective visual effect and objective performance evaluation