【摘要】 传统方法在进行压缩域图像均衡增强时,图像细节容易丢失,使边缘较为模糊,增强后图像清晰度不佳。为解决上述问题,提出机器学习的压缩域图像均衡增强方法。通过扫描程序采集图像信息,提取数据库中的数据;确定分类边界后,实现压缩域图像均衡强;根据正向检测和逆向检测对压缩域图像进行检测。利用法相矢量获取图像的特征点,提取图像色彩、对比、噪点等多种参数值。建立压缩域图像压缩矩阵,通过线性方程匹配特征信息,计算信噪比,实现压缩域图像均衡增强。实验结果表明,设计方法能够更好地处理细节问题,有效增强图像的清晰度。更多还原

【Abstract】 During image equalization enhancement in compressed domain, the loss of image details in traditional methods leads to image edge blur. Therefore, this paper proposes an image equalization enhancement method based on machine learning in compressed domain. Based on the scanning program, image information was collected to extract data from the database. After the classification boundary was determined, the image equalization in compressed domain was achieved. The compressed domain image was detected by forward and backward detections. The normal phase vector was used to obtain the image feature points, and extract the image color, contrast and noise. The image compression matrix in compressed domain was founded, the linear equation was matched, the signal-to-noise ratio was calculated, eventually, the image equalization enhancement in compressed domain was completed. The simulation results show that this method is superior to the traditional methods in detail processing, and has excellent image definition.更多还原