Abstract: Edge detection is the essential question of image processing and computer vision. Eye visual system for images edge information is very sensitive, and it is often as a kind of important means of description image characteristics. In the video coding standard based on motion compensation and DCT transform coding, an image is divided into non-overlapping range blocks of fixed size which are encoded. Block based image edge analysis is useful in the applications, such as image processing, blocking artifacts reduction, mode selection, content-based video retrieval, etc. An edge analysis algorithm based on block edge detection is proposed. Experimental results demonstrate that compared with other algorithms, it has its superior qualities in performance analysis and computational complexity.

Key words: edge detection, video coding, edge direction

摘要： 边缘检测是图像处理和计算机视觉中的基本问题，人眼视觉系统对图像中的边缘信息非常敏感，经常作为描述图像特征的一种重要手段。基于运动补偿和DCT变换编码的视频编码标准中，一幅图像分成大小相同不重叠的块进行编码，基于块的图像边缘分析在图像处理、去块效应滤波、模式选择以及基于内容的视频检索等方面有较广泛应用。提出了一种基于分析块边缘方向的边缘分析算法。实验结果表明，与其他算法相比，该算法在分析性能与计算复杂方面具有较明显的优越性。

关键词: 边缘检测, 视频编码, 边缘方向