基于多尺度灰度变换的图像增强研究

Image enhancement using multi-scale gray-level transformation

为了提高图像增强效果,有效地保持原始图像细节,提出了一种基于多尺度灰度变换的图像增强方法。利用梯度域递归双边滤波对原始图像进行多尺度分解。基于小波变换,将分解层的子带分别作灰度变换。根据变换后的各个子带重构得到分解层的增强结果,并在其基础上实现图像的整体增强.对比直方图均衡化、灰度变换,提出的方法增强效果更好,并且保图像细节。利用客观性能指标对增强结果进行评价。实验结果表明,提出的方法有效,并具有结构简单、计算复杂度低的特点。

In order to improve the performance of image enhancement, effectively preserve original image details, an image enhancement method using multi-scale gray-level transformation was presented. The orig- inal image was decomposed into multi-scale layers using gradient domain based recursive bilateral filtering. Based on wavelet transform, the subcinctures of decomposed layers were transformed respectively in the gray-level. The overall image enhancement was realized based on the enhancement results of decomposed layers, which were reconstituted with the transformed subcinctures. Compared with histogram equalization and gray-level transformation, the proposed method performed better in image enhancement, and preserved details, The performance of the enhancement results was evaluated by the objective indicator. The experimental results show that the proposed method has excellent effect, simple structure and lower computational complexity.