CN-122023294-A - Method for detecting angle of wirecord fabric image with complex background

CN122023294ACN 122023294 ACN122023294 ACN 122023294ACN-122023294-A

Abstract

The invention discloses a method for detecting the angle of a wirecord fabric image with a complex background, which comprises the following steps of roughly extracting a cord fabric area to obtain a rough cord fabric area, wherein the area is mainly used for obtaining position information of the cord fabric area in a picture, respectively generating small measuring moments for detecting edges on a vertical edge and a bevel edge of a tire according to the obtained position information, carrying out gray projection on the small measuring moments, obtaining a first derivative and a second derivative of a projection result, obtaining a point with the maximum first derivative as a boundary point, fitting the points on the bevel edge and the vertical edge to form straight lines, and obtaining an included angle between the two straight lines through slope calculation after obtaining the two fitting straight lines. The invention can shield complex background interference in the picture, and the edge of the fitted curtain cloth is more accurate, thereby improving the stability and accuracy of angle detection.

Inventors

YANG MU
ZHENG YE
ZHANG DONG

Assignees

钛玛科(江苏)工业科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260119

Claims (8)

1. The angle detection method for the wirecord fabric image with the complex background is characterized by comprising the following steps of: Step S1, roughly extracting a curtain cloth region, namely roughly extracting the curtain cloth region by a global threshold method, and carrying out threshold segmentation and post-treatment on the curtain cloth region to obtain a rough curtain cloth region, wherein the region is mainly used for obtaining the position information of the curtain cloth region in a picture; step S2, according to the position information obtained in the step S1, respectively generating small measuring moments for detecting edges on the vertical side and the inclined side of the tire, wherein the direction of the small measuring moment on the vertical side is the horizontal direction, and the direction of the small measuring moment on the inclined side is the standard value of a preset curtain cloth; step S3, gray projection is carried out on the small measuring moment, horizontal projection is carried out on the small measuring moment on the vertical side, and projection is carried out on the small measuring moment on the inclined side according to a preset standard value; s4, obtaining a first derivative and a second derivative of the projection result, wherein the point with the maximum first derivative is the boundary point; step S5, the small measuring moment is scratched by the edge according to a certain step length, and step S3 and step S4 are executed every time the small measuring moment advances until the small measuring moment reaches the edge of a straight line or an oblique line and cannot advance; Step S6, fitting points on the bevel edge and the vertical edge to form straight lines respectively; And S7, obtaining two fitting straight lines, and calculating the included angle between the two straight lines through slope.
2. The method for detecting angle of steel wire cord fabric image with complex background according to claim 1, wherein in step S1, the global thresholding method extracts the cord fabric region by setting a fixed gray threshold, and the image is divided into cord fabric and background.
3. The method for detecting the angle of the wirecord fabric image with the complex background according to claim 2, wherein the method is characterized by comprising the steps of threshold segmentation to obtain a binary image: 1. If the pixel gray value is more than or equal to the threshold value, marking as curtain cloth; 2. if the pixel gray value < threshold→marked as background.
4. The method for detecting the angle of the wirecord fabric image with the complex background according to claim 1, wherein in the step S1, the post-processing is performed: S11, performing corrosion on the binary image, performing expansion, removing noise points and filling holes; S12, screening the connected domain with the largest area to obtain a rough curtain cloth area.
5. The method for detecting angle of complex background steel cord image as set forth in claim 1, wherein in step S4, S41, calculating projection, namely projecting an image in a horizontal or vertical direction to obtain a projection curve; S42, calculating a first derivative of a projection curve by using a central difference method, and reflecting the change rate of the pixel value; s43, obtaining a second derivative, namely, re-deriving the first derivative curve and assisting in judging the authenticity of the extremum; s44, locating the boundary, namely finding the point with the largest absolute value of the first derivative, and corresponding to the steepest change of the projection curve, namely the boundary point.
6. The method for detecting the angle of a complex background steel cord image according to claim 1, wherein in the step S7, the angle of the steel cord is the angle θ between the cord and the horizontal or vertical direction, and the angle is represented as the slope of the cord line in the image.
7. The method for detecting angle of a complex background steel cord image according to claim 6, wherein a straight line is fitted to edge points of steel cord in the image, such as y=k px x+b, where k px is a slope in a pixel coordinate system.
8. The method for detecting angle of complex background wirecord fabric image of claim 7, wherein the camera calibration is performed to obtain pixel size equivalent, and the slope of pixel coordinates is converted into actual physical slope, i.e. actual slope k = K px calculating the angle between the cord and the horizontal direction from the actual slope k arctan(k)× Λx is the actual width corresponding to 1 pixel in the horizontal direction, and λy is the actual height corresponding to 1 pixel in the vertical direction.

Description

Method for detecting angle of wirecord fabric image with complex background Technical Field The invention relates to a detection method, in particular to a detection method for angle of a wirecord fabric image with a complex background. Background In the production process of the tire, the angle of the wirecord fabric is required to be detected, the traditional method for calculating the angle of the wirecord fabric is to extract the wirecord fabric area by utilizing the gray level difference of the wirecord fabric and the background through a global threshold segmentation method, then extract the outline of the wirecord fabric area, further find the two boundaries of the wirecord fabric, and then obtain the angle of the wirecord fabric. For images of wirefabrics with complex background or uneven gray scale (the wirefabric images are shown in fig. 1-3, but the background is different, the right black diamond is the wirefabric, and the rest is the background of the conveyor belt), the angle area of the wirefabric cannot be extracted by directly using the method of the global threshold value due to background interference, or the angle error is larger. Disclosure of Invention The invention provides a method for detecting the angle of a wirecord fabric image with a complex background, which aims to solve the problems that the angle area and the angle error of the wirecord fabric are difficult to be directly extracted by using a global threshold value. The invention provides the following technical scheme: A method for detecting the angle of a wirecord fabric image with a complex background comprises the following steps: Step S1, roughly extracting a curtain cloth region, namely roughly extracting the curtain cloth region by a global threshold method, and carrying out threshold segmentation and post-treatment on the curtain cloth region to obtain a rough curtain cloth region, wherein the region is mainly used for obtaining the position information of the curtain cloth region in a picture; step S2, according to the position information obtained in the step S1, respectively generating small measuring moments for detecting edges on the vertical side and the inclined side of the tire, wherein the direction of the small measuring moment on the vertical side is the horizontal direction, and the direction of the small measuring moment on the inclined side is the standard value of a preset curtain cloth; step S3, gray projection is carried out on the small measuring moment, horizontal projection is carried out on the small measuring moment on the vertical side, and projection is carried out on the small measuring moment on the inclined side according to a preset standard value; s4, obtaining a first derivative and a second derivative of the projection result, wherein the point with the maximum first derivative is the boundary point; step S5, the small measuring moment is scratched by the edge according to a certain step length, and step S3 and step S4 are executed every time the small measuring moment advances until the small measuring moment reaches the edge of a straight line or an oblique line and cannot advance; Step S6, fitting points on the bevel edge and the vertical edge to form straight lines respectively; And S7, obtaining two fitting straight lines, and calculating the included angle between the two straight lines through slope. In step S1, the global thresholding method extracts the curtain area by setting a fixed gray threshold, and divides the image into a curtain and a background. Further, threshold segmentation is performed to obtain a binary image: 1. If the pixel gray value is more than or equal to the threshold value, marking as curtain cloth; 2. if the pixel gray value < threshold→marked as background. In step S1, post-processing: S11, performing corrosion on the binary image, performing expansion, removing noise points and filling holes; S12, screening the connected domain with the largest area to obtain a rough curtain cloth area. In the step S4 of the process, S41, calculating projection, namely projecting an image in a horizontal or vertical direction to obtain a projection curve; S42, calculating a first derivative of a projection curve by using a central difference method, and reflecting the change rate of the pixel value; s43, obtaining a second derivative, namely, re-deriving the first derivative curve and assisting in judging the authenticity of the extremum; s44, locating the boundary, namely finding the point with the largest absolute value of the first derivative, and corresponding to the steepest change of the projection curve, namely the boundary point. In step S7, the angle of the wirecord fabric refers to the included angle theta between the cord and the horizontal or vertical direction, and the included angle theta is represented as the slope of the cord line in the image. Straight lines are fitted to the steel cord edge points in the image, such as y=k px x+b, where k px is the slo