CN-121999041-A - Visual detection positioning method and system for continuous casting ladle drain

CN121999041ACN 121999041 ACN121999041 ACN 121999041ACN-121999041-A

Abstract

The invention relates to a visual detection positioning method for a continuous casting ladle drain, which belongs to the technical field of metallurgy and intellectualization, and comprises the following steps of S1, driving a visual detection module to photograph and scan the continuous casting ladle drain through a moving module, collecting a depth image and corresponding point cloud data, S2, performing visual processing on the collected depth image, S3, performing circular complementary inspection through a RANSAC algorithm when an effective circle is not detected, S4, determining a point cloud interesting ROI (region of interest) according to an image detection result, processing the ROI point cloud, S5, converting a circle center coordinate from a camera coordinate system to a robot base coordinate system through a hand-eye calibration matrix, calculating an attitude angle of a target in the robot base coordinate system, and communicating the attitude angle to a long water gap four-axis robot, thereby completing automatic long water gap installation operation. The invention also provides a visual detection positioning system for the continuous casting ladle drain.

Inventors

ZHANG HUAQIANG
Gong Guibo
LIU GUILIN
TU LIN

Assignees

中冶赛迪工程技术股份有限公司

Dates

Publication Date: 20260508
Application Date: 20251231

Claims (10)

1. A visual detection positioning method for a continuous casting ladle drain is characterized by comprising the following steps: s1, driving a visual detection module to photograph and scan a continuous casting ladle position water outlet through a moving module, and collecting depth images and corresponding point cloud data; S2, performing visual processing on the acquired depth image; S3, when the effective circle is not detected, performing circular complement detection by adopting a RANSAC algorithm; s4, determining an ROI region of interest of the point cloud according to the image detection result, and processing the ROI point cloud; S5, converting the center coordinates from a camera coordinate system to a robot base coordinate system through a hand-eye calibration matrix, calculating the attitude angle of the target in the robot base coordinate system, and communicating the attitude angle to the long water gap four-axis robot to finish the operation of automatically installing the long water gap.
2. The visual inspection and positioning method for the continuous casting ladle drain nozzle according to claim 1, wherein the step S2 of visually processing the collected depth image comprises the following steps: s21, presetting a depth range, and filtering invalid pixels; S22, threshold segmentation is carried out on the filtered image; s23, detecting the outline in the image and screening the possible circular area according to the area and the aspect ratio and the outline radius.
3. The method for visually inspecting and positioning a drain port of a continuous casting ladle as claimed in claim 1, wherein in step S21, the preset depth range is determined based on the installation position of the drain port of the continuous casting ladle and the effective inspection distance of the visual inspection module, and is used for filtering out the pixels in the non-target area which are too close or too far, and reserving the pixels in the depth range which possibly contains the drain port.
4. The visual inspection positioning method of the continuous casting ladle nozzle according to claim 1 is characterized in that in the step S23, the aspect ratio is set to be close to 1 so as to adapt to the circular profile feature of the ladle nozzle, and the profile area is not smaller than a preset threshold, wherein the preset threshold is determined according to the actual size of the continuous casting ladle nozzle and the camera imaging proportion.
5. The visual inspection positioning method of the continuous casting ladle drain nozzle according to claim 1, wherein the performing circular complement inspection by adopting the RANSAC algorithm in the step S3 comprises randomly sampling a plurality of point fitting initial circles from an image contour, calculating distances from all contour points to the initial circles, counting the number of inner points meeting a threshold value condition, and reserving the circle with the largest number of inner points as an inspection result after the preset iteration number is reached.
6. The visual inspection and positioning method for the continuous casting ladle drain nozzle of claim 1, wherein the processing of the ROI point cloud in the step S4 comprises the following steps: s41, downsampling and outlier removal, namely adopting voxel grid downsampling to reduce the number of point clouds, and removing outliers through a statistical outlier removal algorithm; S42, extracting a plane point cloud, namely extracting a surface plane of a water outlet end surface, from the optimized point cloud by using a RANSAC algorithm, extracting an edge point cloud based on normal vector analysis, and focusing on a key feature area; And S43, 3D circle fitting, namely performing 3D circle fitting on the edge point cloud by using a RANSAC algorithm to obtain the center coordinates, the radius and the normal vector of the water outlet.
7. The visual inspection positioning method of a continuous casting ladle nozzle according to claim 6, wherein the range threshold value of the RANSAC algorithm in the step S43 is set according to the accuracy and positioning requirements of point cloud data and is used for screening inner points conforming to a circular model, and the radius range is determined according to the design size of the continuous casting ladle nozzle so as to exclude non-target circular areas.
8. A continuous casting ladle drain opening visual detection positioning system is characterized by comprising a mobile module and a visual detection module based on the continuous casting ladle drain opening visual detection positioning method according to any one of claims 1-7; The movable module is connected with a tail end grabbing structure and is used for driving the visual detection module to move to a fixed photographing position of the continuous casting ladle; The visual detection module consists of a depth camera containing a visual algorithm and a steel protective shell, and is used for performing image acquisition, image filtering, contour detection and circular recognition, realizing point cloud cutting, plane extraction and 3D circle fitting, completing conversion from a camera coordinate system to a robot base coordinate system through a hand-eye matrix, calculating the attitude angle of a target in the robot base coordinate system and communicating the attitude angle to the long water gap four-axis robot.
9. An electronic device comprising a memory and a processor; The memory is used for storing a computer program; The processor is configured to implement the continuous casting ladle drain visual detection positioning method as claimed in any one of claims 1 to 7 when executing the computer program.
10. A computer readable storage medium, wherein a computer program is stored on the storage medium, and when the computer program is executed by a processor, the visual inspection and positioning method of the continuous casting ladle drain as claimed in any one of claims 1 to 7 is realized.

Description

Visual detection positioning method and system for continuous casting ladle drain Technical Field The invention belongs to the technical field of metallurgy intelligence, and relates to a visual detection and positioning method and a visual detection and positioning system for a continuous casting ladle drain. Background In the continuous casting production process of steel smelting, the installation of a long water gap between a continuous casting ladle and a tundish is a key link for guaranteeing the purity of molten steel and preventing secondary oxidation. At present, steel factories at home and abroad commonly adopt a manual operation semi-automatic equipment mode to finish the installation operation of a long water gap, and the typical process is that an operator manually controls a hydraulic or pneumatic driven installation arm to convey the long water gap to a target area through naked eyes to observe the relative position relation between a ladle bottom water gap and a tundish, and the space posture of the installation arm is adjusted by experience to finish the operation of butt joint, sealing and locking of the long water gap and the ladle bottom water gap, and finally, the installation quality is confirmed through visual inspection. However, severe conditions such as high temperature, dust, molten steel splashing, harmful gas and the like exist in the continuous casting ladle position area, an operator needs to continuously operate in a short-distance high-temperature environment, so that the occupational health problem is easily caused, visual judgment errors can be caused by environmental interference, and the positioning accuracy is insufficient and the quality stability is poor. Therefore, aiming at the continuous casting ladle position long nozzle installation link, an automatic technical scheme which can adapt to high-temperature severe environment, realize high-precision positioning, improve the operation efficiency and reduce the safety risk is needed so as to solve the inherent defects of the existing manual operation mode. Disclosure of Invention In view of the above, the invention aims to solve the problems of low positioning precision, poor environmental adaptability, low efficiency, insufficient safety and the like in the existing continuous casting ladle drain opening manual positioning and long water opening installation processes, and provides a continuous casting ladle drain opening visual detection positioning method and system, which realize full-automatic high-precision positioning in a high-temperature and dusty environment and provide reliable position and attitude data for automatic installation of a long water opening robot. In order to achieve the above purpose, the present invention provides the following technical solutions: In one aspect, the invention provides a visual detection positioning method for a continuous casting ladle drain, which comprises the following steps: s1, driving a visual detection module to photograph and scan a continuous casting ladle position water outlet through a moving module, and collecting depth images and corresponding point cloud data; S2, performing visual processing on the acquired depth image; S3, when the effective circle is not detected, performing circular complement detection by adopting a RANSAC algorithm; s4, determining an ROI region of interest of the point cloud according to the image detection result, and processing the ROI point cloud; S5, converting the center coordinates from a camera coordinate system to a robot base coordinate system through a hand-eye calibration matrix, calculating the attitude angle of the target in the robot base coordinate system, and communicating the attitude angle to the long water gap four-axis robot to finish the operation of automatically installing the long water gap. Further, as described in step S2, the visual processing of the acquired depth image includes the following steps: s21, presetting a depth range, and filtering invalid pixels; S22, threshold segmentation is carried out on the filtered image; s23, detecting the outline in the image and screening the possible circular area according to the area and the aspect ratio and the outline radius. Further, in step S21, the preset depth range is determined based on the installation position of the drain port of the continuous casting ladle and the effective detection distance of the visual detection module, and is used for filtering out the pixels of the non-target area which are too close or too far, and reserving the pixels in the depth interval possibly including the drain port. Further, in step S23, the aspect ratio is set to be close to 1 to adapt to the circular contour feature of the drain opening, and the contour area is not smaller than a preset threshold, where the preset threshold is determined according to the actual size of the drain opening of the continuous casting ladle and the camera imaging proportion. Further, in step S3