基于DCT压缩域的图像边缘检测算法

目前，大多数边缘检测算法都是在像素域内进行的，而图像数据大多以压缩格式传输和存储，因此要进行边缘检测，必须先经解压缩，变成非压缩格式，再进行处理，这无疑增加了计算量，影响了处理的速度和性能。给出了一种直接在DCT压缩域中进行边缘检测的方法，通过推导出的边缘检测矩阵算子，直接对反量化后的DCT系数进行处理，进而判断出图像边缘点。与传统的像素域边缘检测方法相比，文中算法由于不需要完全解压缩，因而大大降低了计算复杂度。

Algorithm for Image Edge Detection Based on DCT Compressed Domain

Now,most of algorithms for detection of image edge are based on pixel domain.However,most of the image data are transferred and stored in their compressed format,so,for the edge detection of an image,it must be decompressed firstly,turn to non-compressed format and processed,which is increased the computational cost and is very ineffective to the overall performance of the processing.In this paper,a method which detects image edge directly in DCT compressed domain is proposed,using the introduced edge detection matrix arithmetic operators,directly computes the DCT coefficients which is quantized and then decided the edge points of the image.Compared with the traditional pixel domain based edge detection method,the computational cost was greatly reduced.