摘要
为使数码相机直接拍摄电子显示屏获得更清晰的摄屏图像,提出一种基于液晶点物理结构与原始像素点映射的空间重构算法。该算法通过改进的尺度金字塔标定采样的液晶结构特征,并利用基于区域关系的向量内积插值算法恢复颜色通道。算法将图像重构至原始尺寸,在保留原始图像信息的同时过滤高频噪声,避免缩放产生摩尔纹和镜头畸变。实验表明,与传统图像分割算法相比,基于尺度金字塔标定的摄屏图像特征算法运行时间更短,样本提取特征准确度达到99.99%。其重构图像与原始图像的直方图相似度达到85%,远高于摄屏图像与原始图像36%的相似度,同时消除了镜头畸变与缩放类摩尔纹。
A space reconstruction algorithm based on the physical structure of liquid crystal dots and the mapping of original pixels is studied for the digital camera to directly capture the image of the electronic display screen.The algorithm recovers the color channel by using the improved scale pyramid to calibrate the oversampled LCD structure features and the vector interpolation algorithm based on the region relationship.The algorithm reconstructs the image to the original size,filters the high-frequency noise while retaining the original image information,and avoids the moire lines and lens distortion caused by zooming.Experiments show that compared with the traditional image segmentation algorithm,the running time of the feature algorithm is shorter,and its accuracy is better than that of the traditional image segmentation algorithm.The histogram similarity between the reconstructed image and the original image is 85%,which is much higher than the similarity of 36%between the screen image and the original image.At the same time,the lens distortion and zooming moire lines are eliminated.
作者
陈申渭
马汉杰
冯杰
许佳立
CHEN Shen-wei;MA Han-jie;FENG Jie;XU Jia-li(School of Information Science and Technology,Zhejiang Sci-Tech University,Hangzhou 310018,China;Yuhang District Branch of Hangzhou Public Security Bureau,Hangzhou 311100,China)
出处
《软件导刊》
2020年第4期234-237,共4页
Software Guide
基金
国家自然科学基金项目(61501402)。
关键词
图像金字塔
图像滤波
图像重构
image pyramid
image filtering
image reconstruction
