论文部分内容阅读
在复合材料的三维重构中需要对图像进行拼接,以解决序列图像的区域对应问题.采用一种新的图像拼接融合方法,对两幅放大倍数为1000倍的SiC/Al复合材料照片进行了拼接.该方法首先对两幅待拼接的图像进行去噪处理,采用Harris角点检测方法提取特征点,并对这些特征点采用Canny边缘检测算子进行筛选.在两幅图像重合的部分选取一定大小的模板进行基于像素的互相关匹配,以选定用于匹配的特征点,在此基础上运用SVD-ICP方法获取模板图像的偏转和平移参数,然后依据这些参数对整幅图像进行变换,建立两幅图像的统一坐标表示,完成图像的拼接.
In the three-dimensional reconstruction of composite materials, the images need to be spliced to solve the regional image correspondence problem.A new method of image mosaic fusion was used to fabricate two SiC / Al composite images with magnification of 1000 times In the method, the two images to be spliced are firstly denoised, the Harris corner detection method is used to extract the feature points and the Canny edge detection operator is used to filter these feature points.When the two images overlap, a certain Size of the template for pixel-based cross-correlation matching to select the matching feature points, based on the use of SVD-ICP method to obtain the template image deflection and translation parameters, and then based on these parameters to transform the entire image, Establish a unified coordinate representation of two images to complete the image mosaic.