3D计算中高速图形转换器算法设计

在图像处理过程中的分析图像的特征,需要达成特征提取,为后续的图像处理和分析。在目标图像的提取过程中,本文选定了特征角点特征。角点定位准确的图像匹配在三维重建中起着重要的作用。角落像素相对于外在的像素灰度值的突变或凸形状的像素点较为集中。本文分析了三个角点提取方法:苏珊角点检测方法,哈里斯角点检测方法,多尺度结合苏珊算法改进,亚像素角点检测方法,以便满足系统对精度的要求。

The Algorithm Design of High-speed Graphics Converter in 3D Calculation

The characteristics of image analysis in image processing,the need to reach the feature extraction for the subsequent image processing and analysis.In the extraction process in the target image,the selected characteristics of corner features.Corner location and accurate image matching plays an important role in the three-dimensional reconstruction.Corner of the pixel relative to the mutation of the external value of the pixel gray or convex shape of the pixel is more concentrated.This paper analyzes the three-corner extraction:SUSAN corner detection method,the Harris corner detection method of multiple scales combined with Susan algorithm improvement,subpixel corner detection method,in order to meet the requirements of the system accuracy.