Halcon棋盘格畸变矫正

LeonDL168

于 2024-07-24 08:45:19 发布

阅读量98

点赞数 1

分类专栏： Halcon 文章标签： opencv 计算机视觉人工智能图像处理算法

原文链接：https://blog.csdn.net/m0_51559565/article/details/136586345

版权

Halcon 专栏收录该内容

203 篇文章 9 订阅

订阅专栏

前言

通常我们使用的相机都是带有一定的畸变的，只是畸变的大小不同。对于我们常见的测量项目来说，对测量精度的要求很高，通常从相机镜头上进行优化，但是即使镜头优化或者成本原因，造成图像仍有一定的畸变时，我们就可是使用棋盘格标定板，进行畸变矫正

1.程序

* Image Acquisition 01: Code generated by Image Acquisition 01
read_image (Image, 'C:/Users/1/Desktop/work/3.8/10.bmp')
*****************************标定矫正过程**************************************
WidthOfGrid := 0.17
NumSquares := 110
create_rectification_grid (WidthOfGrid, NumSquares, 'rectification_grid.ps')
MinContrast := 25
Radius := 10
find_rectification_grid (Image, GridRegion, MinContrast, Radius)
reduce_domain (Image, GridRegion, ImageReduced)

SigmaSaddlePoints := 2
Threshold := 3
saddle_points_sub_pix (ImageReduced, ‘facet’, SigmaSaddlePoints, Threshold, Row, Col)
gen_cross_contour_xld (SaddlePoints, Row, Col, 20, 1)
dev_display (Image)
dev_display (SaddlePoints)

SigmaConnectGridPoints := 0.9
MaxDist :=5
GridSpacing := 45
dev_set_color (‘red’)
connect_grid_points (ImageReduced, ConnectingLines, Row, Col, SigmaConnectGridPoints, MaxDist)
gen_grid_rectification_map (ImageReduced, ConnectingLines, Map, Meshes, GridSpacing,0, Row, Col, ‘bilinear’)
map_image (ImageReduced, Map, ImageMapped)

存储标定参数*
write_image (Map, ‘tiff’, 0, ‘C:/Users/1/Desktop/work/3.8/calib.tiff’)
read_image (Map, ‘C:/Users/1/Desktop/work/3.8/calib.tiff’)

读取图像测试
read_image (Image, ‘C:/Users/1/Desktop/work/3.8/2.bmp’)
map_image (Image, Map, ImageMapped)

2.采集标定图像

在这里插入图片描述
此次使用激光雕刻的棋盘格，单个格子长宽2mm，图像采集像素值为单格43像素。
标定前，可以根据所需要的视野大小进行标定，当然最佳的方式是标定板应该大于视野范围，以确定边缘畸变的量，同时标定板应该摆放正，如果没有摆放正的情况会出现矫正图像有倾斜的现象。

3.裁剪标定板区域

*实际棋盘格宽度
WidthOfGrid := 0.3
*棋盘格单向数量
NumSquares := 110
*创建标定板
create_rectification_grid (WidthOfGrid, NumSquares, 'rectification_grid.ps')
*最小二值化的值
MinContrast := 25
Radius := 10
find_rectification_grid (Image, GridRegion, MinContrast, Radius)
reduce_domain (Image, GridRegion, ImageReduced)

4.提取网格

*提取每个网格的交点
*预设参数一般不需要修改。但是需要保证每个格子与格子间的角都要找到，哪里没找到，哪里就会矫正缺损
SigmaSaddlePoints := 2
Threshold := 3
saddle_points_sub_pix (ImageReduced, 'facet', SigmaSaddlePoints, Threshold, Row, Col)
gen_cross_contour_xld (SaddlePoints, Row, Col, 20, 1)
dev_display (Image)
dev_display (SaddlePoints)
*生产网格线
SigmaConnectGridPoints := 0.9
MaxDist :=5
*设置矫正后的单个像素值
GridSpacing := 43
dev_set_color ('red')
connect_grid_points (ImageReduced, ConnectingLines, Row, Col, SigmaConnectGridPoints, MaxDist)
gen_grid_rectification_map (ImageReduced, ConnectingLines, Map, Meshes, GridSpacing,0, Row, Col, 'bilinear')
*矫正图像
map_image (ImageReduced, Map, ImageMapped)

注意：当棋盘格上有细小的污点时，会严重影响实际的畸变矫正的效果，所以采集到图像时，需要使用PS或者画图软件将污点去除，在进行运算

5.保存矫正数据和验证

5.1保存数据

******************************************存储标定参数*****************
write_image (Map, 'tiff', 0, 'C:/Users/1/Desktop/work/3.8畸变矫正/calib.tiff')
read_image (Map, 'C:/Users/赵伟/Desktop/work/3.8畸变矫正/calib.tiff')