Template mark detection method and template position correction method based on single camera

Patent metadata

Field	Value
Publication number	US-12100188-B2
Application number	US-202218575791-A
Country	US
Kind code	B2
Filing date	Dec 27, 2022
Priority date	Dec 21, 2022
Publication date	Sep 24, 2024
Grant date	Sep 24, 2024

Abstract

Provided are template mark detection method and template position correction method based on single camera, including: performing image collection on mark on template by single camera, and obtaining binary image after preprocessing; performing corner detection of jagged edges on binary image to obtain corner set of jagged edges; performing edge detection and line detection sequentially on binary image to obtain set of edge line segments from coarse detection; traversing such set, and judging and retaining collinear line segments, to obtain set of collinear line segments from coarse detection; traversing corner set of jagged edges, for point-line collinearity judgment with line segments in set of collinear line segments from coarse detection, to obtain set of point-line from fine detection; and performing linear fitting on set of point-line from fine detection, and calculating an inclination angle of each straight line through an arctan function, thus completing detection of mark.

First claim

What is claimed is:

1. A template mark detection method based on a single camera, comprising:
S 1 , carrying out image collection on a mark on a template by a single camera, and obtaining a binary image after preprocessing;
S 2 , carrying out corner detection of jagged edges on the binary image to obtain a corner set of jagged edges;
S 3 , performing edge detection and line detection in sequence on the binary image to obtain a set of edge line segments from coarse detection;
S 4 , traversing the set of edge line segments from coarse detection, and judging and retaining collinear line segments therein, to obtain a set of collinear line segments from coarse detection;
S 5 , traversing the corner set of jagged edges, for point-line collinearity judgment with line segments in the set of collinear line segments from coarse detection, to obtain a set of point-line from fine detection; and
S 6 , carrying out linear fitting on the set of point-line from fine detection, and calculating an inclination angle of each of straight lines through an arctan function, thus completing detection of the mark.

2. The template mark detection method based on a single camera according to claim 1 , wherein the S1 comprises:
S11, carrying out the image collection on the mark on the template by the single camera, to obtain an initial image;
S12, converting the initial image into a gray scale image and filtering the gray scale image; and
S13, performing binarization processing on an image obtained in the S12, to obtain the binary image.

3. The template mark detection method based on a single camera according to claim 1 , wherein the S2 comprises:
performing Shi-Tomasi corner response function detection on pixel points in the binary image, to obtain the corner set of jagged edges.

4. The template mark detection method based on a single camera according to claim 1 , wherein the S3 comprises:
performing Canny edge detection on the binary image to obtain an edge pattern; and
obtaining a plurality of edge line segments in the mark using Progressive Probabilistic Hough Transform according to the edge pattern, wherein the plurality of edge line segments form the set of edge line segments from coarse detection.

5. The template mark detection method based on a single camera according to claim 1 , wherein the S4 comprises:
S41, arbitrarily selecting two line segments in the set of edge line segments from coarse detection, to obtain four endpoints;
S42, calculating an area of a quadrangle formed by the four endpoints, and calculating a slope of one of the line segments, and slopes of line segments formed by connecting two endpoints of one line segment to two endpoints of the other line segment respectively;
S43, when the area of the quadrangle is smaller than a first preset threshold and a difference value between slopes is smaller than a second preset threshold, the two line segments being collinear; and
S44, traversing the set of edge line segments from coarse detection, and retaining the collinear line segments therein, to obtain the set of collinear line segments from coarse detection.

6. The template mark detection method based on a single camera according to claim 1 , wherein the S5 comprises:
S51, arbitrarily selecting a corner from the corner set of jagged edges, the corner forming triangles with the line segments in the set of collinear line segments from coarse detection respectively;
S52, calculating areas of the triangles;
S53, when the areas of the triangles are smaller than a third preset threshold, the corner being point-line collinear with the line segments; and
S54, traversing the corner set of jagged edges, and merging point-line collinear corners and line segments to obtain the set of point-line from fine detection.

7. The template mark detection method based on a single camera according to claim 6 , wherein the S5 further comprises:
S55, selecting same collinear line segments in the set of point-line from fine detection, retaining any one thereof, and removing redundant line segments.

8. The template mark detection method based on a single camera according to claim 1 , wherein the S6 comprises:
S61, carrying out linear fitting on a corner and two endpoints of a corresponding line segment in the set of point-line from fine detection to obtain a linear equation, the corner and the corresponding line segment being collinear; and
S62, calculating an inclination angle of each of the straight lines through the arctan function according to the linear equation.

9. The template mark detection method based on a single camera according to claim 1 , wherein the mark on the template comprises one of a “*” shape and a cross shape.

10. A template position correction method based on a single camera, comprising:
S1, carrying out image collection on a mark on a template by a single camera, and obtaining a binary image after preprocessing;
S2, carrying out corner detection of jagged edges on the binary image to obtain a corner set of jagged edges;
S3, performing edge detection and line detection in sequence on the binary image to obtain a set of edge line segments from coarse detection;
S4, traversing the set of edge line segments from coarse detection, and judging and retaining collinear line segments therein, to obtain a set of collinear line segments from coarse detection;
S5, traversing the corner set of jagged edges, for point-line collinearity judgment with line segments in the set of collinear line segments from coarse detection, to obtain a set of point-line from fine detection;
S6, carrying out linear fitting on the set of point-line from fine detection, and calculating an inclination angle of each of straight lines through an arctan function; and
S7, correcting a position of the template according to a difference value between the inclination angle of each of the straight lines and a standard angle.

Assignees

• Inst Optics & Electronics Cas

Inventors

• Luo Xiangang
• Li Chengwenli
• Liu Minggang
• Yan Jian
• Wang Changtao

Classifications

• G06T2207/20076

  Probabilistic image processing · CPC title

• G06T2207/20061

  Hough transform · CPC title

• G06V10/26

  Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion · CPC title

• G06V10/443

  by matching or filtering · CPC title

• G06T7/62

  of area, perimeter, diameter or volume · CPC title