基于BIRCH聚类加速的彩色图像增强算法

针对现存的大多图像增强算法增强的图像可见性丢失问题,提出了一种基于BIRCH聚类加速的彩色图像增强算法；首先,通过BIRCH聚类加速确定数据库中与输入图像直方图相似度最高的图像来提取图像特征；然后,选择最小欧氏距离的特征值进行图像融合以获取目标图像；最后,增强图像通过目标图像直方图规范化和后期处理获得;大量图像融合实验结果验证了算法的有效性,该算法扩展了图像增强的类别,解决了增强过程中可能出现的可见性丢失问题,使图像增强的适应性更强;另外,EM、CII和SSIM评估指标的结果表明该算法明显改善了增强效果。

A color image enhancement algorithm based on BIRCH cluster acceleration

As the visibility of the loss problems for most of the existing image enhancement algorithms, a color image enhancement algorithm based on BIRCH cluster acceleration is proposed. Firstly, the images in the database with highest histogram correlation with input images are identified with birch cluster acceleration. Then, the target image is obtained by fusing images based on minimum Euclidean distance between extracted features. Finally, the enhanced-image is obtained by the normalization of the target image and after-processing. The verification of the proposed algorithm is testified by many experimental results. And the proposed algorithm extends the image enhancement category, solving the visibility of the loss problems that may occur during enhancing, and making image enhancement adaptable. In addition, the results of the assessment indicators EM, CII and SSIM show that the proposed algorithm improves the enhanced effect significantly.