CN-122023233-A - Deformed steel bar torsion correction method and device based on image recognition

CN122023233ACN 122023233 ACN122023233 ACN 122023233ACN-122023233-A

Abstract

The application provides a deformed steel bar torsion correction method and device based on image recognition, which belong to the technical field of image recognition, and the method comprises the steps of obtaining initial deformed steel bar images at a plurality of moments, dust concentration values at a plurality of moments and illuminance values; for an initial deformed steel bar image acquired at each moment, determining an image processing parameter based on deformed steel bar specification data, a dust concentration value and an illuminance value corresponding to the initial deformed steel bar image acquired at the moment, processing the initial deformed steel bar image acquired at the moment based on the image processing parameter to obtain a target deformed steel bar image, calculating initial torsion deviation angles based on all the target deformed steel bar images, and determining a target deviation correcting parameter based on the initial torsion deviation angles and the deformed steel bar moving speed if the initial torsion deviation angles exceed a torsion angle deviation threshold value, wherein the target deviation correcting parameter is used for correcting and controlling the deformed steel bar torsion angles. The application can realize the self-adaptive accurate deviation correction of the deformed steel bar under different working conditions.

Inventors

ZHAO BO
LI XIANGYANG
XUE HAIPENG
ZHU SHIBIN
MA CHAO
ZHANG BOWEN
CAO XINGYU
ZHANG JINSONG

Assignees

河北敬业宽板科技有限公司

Dates

Publication Date: 20260512
Application Date: 20251208

Claims (10)

1. The method for correcting the torsion of the deformed steel bar based on the image recognition is characterized by comprising the following steps of: Acquiring initial screw-thread steel images at a plurality of moments based on a first data acquisition frequency, and acquiring dust concentration values and illuminance values at a plurality of moments based on a second data acquisition frequency, wherein the first data acquisition frequency is higher than the second data acquisition frequency, and the dust concentration value and the illuminance value at one moment correspond to the initial screw-thread steel images at a plurality of moments; For an initial deformed steel bar image acquired at each moment, determining an image processing parameter based on deformed steel bar specification data, a dust concentration value and an illuminance value corresponding to the initial deformed steel bar image acquired at the moment, and processing the initial deformed steel bar image acquired at the moment based on the image processing parameter to acquire a target deformed steel bar image; And calculating an initial torsion deviation angle based on all target deformed steel bar images, acquiring the deformed steel bar moving speed if the initial torsion deviation angle exceeds a torsion angle deviation threshold value, and determining a target deviation correcting parameter based on the initial torsion deviation angle and the deformed steel bar moving speed, wherein the target deviation correcting parameter is used for correcting and controlling the deformed steel bar torsion angle.
2. The method for correcting the torsion of the deformed steel bar based on the image recognition according to claim 1, wherein the determining the image processing parameter based on the deformed steel bar specification data, the dust concentration value and the illuminance value corresponding to the initial deformed steel bar image acquired at the moment comprises: Determining a deformed steel bar type based on deformed steel bar specification data; Determining a target working condition type based on the type of the deformed steel bar, and a dust concentration value and an illuminance value corresponding to an initial deformed steel bar image acquired at the moment; And determining an image processing parameter based on the target working condition type.
3. The method for correcting the torsion of the deformed steel bar based on the image recognition according to claim 2, wherein the determining the target working condition type based on the deformed steel bar type, the dust concentration value and the illuminance value corresponding to the initial deformed steel bar image acquired at the moment comprises: if the screw thread steel type is of a first specification type, calculating a dust interference factor based on a first dust influence coefficient and a dust concentration value corresponding to an initial screw thread steel image acquired at the moment, and calculating an illumination interference factor based on a first illumination influence coefficient and an illumination value corresponding to the initial screw thread steel image acquired at the moment; If the screw thread steel type is of a second specification type, calculating a dust interference factor based on a second dust influence coefficient and a dust concentration value corresponding to an initial screw thread steel image acquired at the moment, and calculating an illumination interference factor based on a second illumination influence coefficient and an illumination value corresponding to the initial screw thread steel image acquired at the moment; The deformed steel bar specification diameter corresponding to the first specification type is smaller than the deformed steel bar specification diameter corresponding to the second specification type; the first dust influence coefficient is smaller than the second dust influence coefficient, and the first illumination influence coefficient is smaller than the second illumination influence coefficient; and determining the target working condition type based on the dust interference factor and the illumination interference factor.
4. The method for correcting the torsion of the deformed steel bar based on the image recognition according to claim 3, wherein the determining the target working condition type based on the dust interference factor and the illumination interference factor comprises: If the dust interference factor is smaller than a dust interference threshold value and the illumination interference factor is smaller than an illumination interference threshold value, determining that the target working condition type is a first working condition type; if the dust interference factor is greater than or equal to a dust interference threshold and the illumination interference factor is less than an illumination interference threshold, determining that the target working condition type is a second working condition type; if the dust interference factor is smaller than a dust interference threshold and the illumination interference factor is larger than or equal to an illumination interference threshold, determining that the target working condition type is a third working condition type; if the dust interference factor is greater than or equal to a dust interference threshold value and the illumination interference factor is greater than or equal to an illumination interference threshold value, determining that the target working condition type is a fourth working condition type; The image processing parameters corresponding to the first working condition type, the second working condition type, the third working condition type and the fourth working condition type comprise filtering parameters, edge detection parameters and Hough transformation parameters; The filtering parameters corresponding to the first working condition type are the same as the filtering parameters corresponding to the third working condition type, the filtering parameters corresponding to the second working condition type are the same as the filtering parameters corresponding to the fourth working condition type, and the filtering parameters corresponding to the fourth working condition type are larger than the filtering parameters corresponding to the first working condition type; the edge detection parameters corresponding to the first working condition type are the same as the edge detection parameters corresponding to the second working condition type, the edge detection parameters corresponding to the third working condition type are the same as the edge detection parameters corresponding to the fourth working condition type, and the edge detection parameters corresponding to the second working condition type are larger than the edge detection parameters corresponding to the third working condition type; The Hough transform parameters corresponding to the first working condition type and the Hough transform parameters corresponding to the second working condition type are the same as the Hough transform parameters corresponding to the third working condition type, and the Hough transform parameters corresponding to the third working condition type are larger than the Hough transform parameters corresponding to the fourth working condition type.
5. The method for correcting the torsion of the deformed steel bar based on the image recognition according to claim 4, wherein the image processing parameters corresponding to the first working condition type and the second working condition type further comprise image enhancement parameters; the image enhancement parameter is based on the illuminance value and is determined by: If the illuminance value is smaller than a first illuminance threshold, determining the image enhancement parameter as a first image enhancement parameter, wherein the first image enhancement parameter comprises a first gray mapping range, a first gamma correction coefficient and a first enhancement intensity weight; If the illuminance value is larger than a second illuminance threshold, determining the image enhancement parameter as a second image enhancement parameter, wherein the second image enhancement parameter comprises a second gray mapping range, a second gamma correction coefficient and a second enhancement intensity weight; the first gray scale mapping range includes the second gray scale mapping range, the first gamma correction coefficient is greater than the second gamma correction coefficient, and the first enhancement intensity weight is greater than the second enhancement intensity weight; The first gray level mapping range and the second gray level mapping range are used for adjusting gray level distribution intervals of the initial threaded steel image, the first gamma correction coefficient and the second gamma correction coefficient are used for carrying out nonlinear gray level transformation on the initial threaded steel image, and the first enhancement intensity weight and the second enhancement intensity weight are used for adjusting operation amplitude of image enhancement operation on the initial threaded steel image.
6. The method for correcting the torsion of the deformed steel bar based on the image recognition according to claim 1, wherein the image processing parameters include a filtering parameter, an edge detection parameter and a hough transformation parameter; the processing the initial deformed steel bar image acquired at the moment based on the image processing parameters to obtain a target deformed steel bar image comprises the following steps: Converting an initial deformed steel bar image acquired at the moment into a gray image, and performing filtering processing on the gray image based on the filtering parameters; Performing edge detection on the filtered gray level image based on the edge detection parameters, extracting the contour edges of the deformed steel bar longitudinal ribs, and generating an edge image comprising the contour edges of the deformed steel bar longitudinal ribs; And converting the edge image into a black-and-white binary image, detecting the black-and-white binary image by using the Hough transformation parameters, extracting the main direction straight line corresponding to the longitudinal rib, and generating a target threaded steel image comprising the main direction straight line corresponding to the longitudinal rib.
7. The method for correcting a twist of a deformed steel bar based on image recognition according to claim 1, wherein the determining a target correction parameter based on the initial twist deviation angle and the deformed steel bar moving speed comprises: Determining a deviation correcting parameter adjusting interval based on the screw-thread steel moving speed, wherein the screw-thread steel moving speed and the deviation correcting parameter adjusting interval are in negative correlation; determining a roller adjusting angle based on the deviation correcting parameter adjusting interval and the initial torsion deviation angle; and taking the deviation correcting parameter adjusting interval and the roller adjusting angle as the target deviation correcting parameters.
8. The utility model provides a deformed steel bar twists reverse deviation correcting device based on image recognition which characterized in that includes: The device comprises a data acquisition module, a data processing module and a data processing module, wherein the data acquisition module is used for acquiring initial screw steel images at a plurality of moments based on a first data acquisition frequency and acquiring dust concentration values and illuminance values at a plurality of moments based on a second data acquisition frequency; The image processing module is used for determining image processing parameters for the initial screw-thread steel image acquired at each moment based on screw-thread steel specification data, a dust concentration value and an illuminance value corresponding to the initial screw-thread steel image acquired at the moment, and processing the initial screw-thread steel image acquired at the moment based on the image processing parameters to acquire a target screw-thread steel image; The angle deviation correcting module is used for calculating initial torsion deviation angles based on all target deformed steel bar images, acquiring deformed steel bar moving speed if the initial torsion deviation angles exceed torsion angle deviation thresholds, and determining target deviation correcting parameters based on the initial torsion deviation angles and the deformed steel bar moving speed, wherein the target deviation correcting parameters are used for correcting and controlling the deformed steel bar torsion angles.
9. An electronic device comprising a memory, a processor and a computer program stored in the memory and running on the processor, characterized in that the processor implements the steps of the method according to any of claims 1 to 7 when the computer program is executed.
10. A computer readable storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the steps of the method according to any one of claims 1 to 7.

Description

Deformed steel bar torsion correction method and device based on image recognition Technical Field The application belongs to the technical field of image recognition, and particularly relates to a deformed steel bar torsion correction method and device based on image recognition. Background The screw-thread steel is a core steel bar material in the construction engineering, the longitudinal rib structure on the surface of the screw-thread steel can obviously enhance the bonding adhesive force with concrete, and the form integrity of the longitudinal rib directly determines the mechanical property and the subsequent construction quality of the steel bar. In the process of screw-thread steel rolling production, the screw-thread steel is affected by factors such as the deviation of the axis parallelism of the upper roller and the lower roller, the uneven distribution of rolling force and the like, the phenomenon of twisting of the longitudinal ribs easily occurs in the finished screw-thread steel, the form of the longitudinal ribs deviates from the design standard, the anchoring effect of the reinforced steel bars and the concrete is reduced, and the potential safety hazard of the structure is also caused, so that the detection and the correction of the twisting of the longitudinal ribs are key links in screw-thread steel production. At present, image recognition technology is adopted to assist detection aiming at detection and correction means of screw steel longitudinal rib torsion in the industry. However, the conventional deformed steel bar torsion deviation correcting technology based on image recognition has two major limitations in practical application, on one hand, the prior art adopts fixed image processing parameters to process deformed steel bar images under different working conditions, when the rolling environment fluctuates, the problems of incomplete longitudinal rib profile extraction and error recognition caused by the interference of the environment easily occur, the calculated deviation of torsion angles is higher, and the detection precision is affected, and on the other hand, the deviation correcting module adopts a deviation exceeding threshold, namely a fixed logic for singly adjusting corresponding parameters, the influence of dynamic change in the deformed steel bar production process on deviation correcting response is not considered, secondary torsion is easily caused due to overlarge adjustment amplitude, or deviation accumulation is easily caused due to adjustment lag, and the self-adaptive accurate deviation correction under different working conditions cannot be realized. Disclosure of Invention The application aims to provide a deformed steel bar torsion correction method and device based on image recognition so as to realize self-adaptive and accurate deformed steel bar correction under different working conditions. In a first aspect of the embodiment of the present application, a method for correcting a twisted steel torsion based on image recognition is provided, including: Acquiring initial screw-thread steel images at a plurality of moments based on a first data acquisition frequency, and acquiring dust concentration values and illuminance values at a plurality of moments based on a second data acquisition frequency; For an initial deformed steel bar image acquired at each moment, determining an image processing parameter based on deformed steel bar specification data, a dust concentration value and an illuminance value corresponding to the initial deformed steel bar image acquired at the moment, and processing the initial deformed steel bar image acquired at the moment based on the image processing parameter to acquire a target deformed steel bar image; And calculating an initial torsion deviation angle based on all the target deformed steel bar images, acquiring the deformed steel bar moving speed if the initial torsion deviation angle exceeds a torsion angle deviation threshold value, and determining a target deviation correcting parameter based on the initial torsion deviation angle and the deformed steel bar moving speed, wherein the target deviation correcting parameter is used for correcting and controlling the deformed steel bar torsion angle. In a second aspect of the embodiment of the present application, there is provided a deformed steel bar torsion correction device based on image recognition, including: The system comprises a data acquisition module, a data processing module and a data processing module, wherein the data acquisition module is used for acquiring initial screw steel images at a plurality of moments based on a first data acquisition frequency and acquiring dust concentration values and illuminance values at a plurality of moments based on a second data acquisition frequency; The image processing module is used for determining image processing parameters for the initial screw-thread steel image acquired at each moment based on screw-thread steel