串行视觉通信下失真图像颜色校正方法仿真

摘要

颜色校正是提高图像视觉效果及内容感知理解程度的重要环节.因此,在串行视觉通信下,需要进行失真图像颜色校正方法研究.但是采用当前方法对失真图像进行颜色校正时,图像间的颜色亮度差异较大,存在校正结果易出现噪点及色块的问题.为此,提出一种基于梯度区域分割的串行视觉通信下失真图像颜色校正方法.上述方法先将输入图像帧分解为包含大尺度边缘信息的基图像和含有小尺度信息的细节层,获取失真图像颜色多尺度细节函数,从调色板中提取颜色特征信息,采用统计特征校正方法对颜色单一部分进行校正,采用改进的多维概率密度转换法对颜色复杂部分进行校正,从而完成对串行视觉通信下失真图像颜色的校正.仿真证明,所提方法颜色校正精度高.可以减少串行视觉通信下颜色失真,保持原图结构特征.%A color correction method of fault image under serial visual communication is put forward based on gradient region.The input image frame is decomposed into basic image containing marginal information with large scale and detail level containing information with small scale,then multi-scale detail function of color of the fault image is obtained and feature information of color from color palette is extracted.The correction method of statistical characteristic is used to correct single part of color and the modified transfer approach of multi-scale probability density is used to correct complex part of color.Thus,the color correction is completed.Simulation proves that the method has high color correction precision.It can decrease color inconsistency under the serial visual communication and maintain structure feature of original image.