A method for locating fully mechanized mining equipment in underground mines based on synchronous positioning and mapping

Abstract

The invention discloses an underground comprehensive mining equipment positioning method based on synchronous positioning and mapping, comprising: S1, arranging labels at even intervals on the side wall of the mine tunnel, and using a total station to obtain the true values ​​of multiple labels; S2, setting the comprehensive mining equipment in the mine tunnel, arranging a left locator and a right locator on the left and right sides of the comprehensive mining equipment, and positioning the left locator and the right locator using a VI-SLAM method; S3, eliminating the cumulative error generated by VI-SLAM positioning, separating the labels using color segmentation, judging whether the labels reach a preset size, and if the preset size is reached, performing posture correction according to the true value of the label; S4, when the comprehensive mining equipment moves to the left side of the tunnel side wall, the left locator performs positioning and mapping, and the right locator only performs mapping at this time; conversely, when the comprehensive mining equipment moves to the right side of the tunnel side wall, the right locator performs positioning and mapping, and the left locator only performs mapping.

Description

Technical Field

[0001] The invention belongs to the technical field of coal mine equipment, and in particular relates to an underground fully mechanized mining equipment positioning method based on synchronous positioning and composition. Background Art

[0002] At present, the underground equipment positioning system is mainly based on wireless communication systems, mainly including Zigbee, RFID, and UWB. Its accuracy is from meters to decimeters. It is mainly positioned by establishing a WAN, which requires the laying of sensor nodes in advance.

[0003] Positioning through wireless communication technology is mainly divided into ranging-based methods and non-ranging-based methods. Among them, ranging-based methods are more commonly used, and can be divided into methods based on signal arrival time (TOA), methods based on signal arrival time difference (TDOA), methods based on signal arrival angle (AOA), and methods based on signal arrival time strength (RSSI). Non-ranging-based methods include centroid positioning, APIT positioning, convex planning positioning, and DV-Hop positioning.

[0004] The positioning method based on ranging has higher accuracy than the non-ranging method, but the hardware cost is often high. The positioning method based on non-ranging usually has low accuracy and cannot meet the positioning requirements of downhole equipment. Summary of the invention

[0005] In view of the above-mentioned problems, the object of the present invention is to provide a positioning method for underground fully mechanized mining equipment based on synchronous positioning and mapping.

[0006] The technical solution of the present invention is: a method for positioning underground fully mechanized mining equipment based on synchronous positioning and mapping, comprising the following steps:

[0007] S1. Arrange tags at even intervals on the side walls of the mine tunnel, and use a total station to obtain the true values ​​of multiple tags;

[0008] S2. A fully mechanized mining device is installed in the mine tunnel. A left locator and a right locator are arranged on the left and right sides of the fully mechanized mining device. The left locator and the right locator are positioned using VI-SLAM. The fully mechanized mining device obtains the mine tunnel image through a camera. A multi-layer fusion image enhancement processing module is added to the front end of the left locator and the right locator to improve the brightness and contrast of the acquired mine tunnel image and increase the number of corner point extraction.

[0009] S3, the back ends of the left locator and the right locator are both added with a label feature posture correction module to eliminate the cumulative error caused by VI-SLAM positioning, separate the labels using color segmentation, and determine whether the labels reach a preset size. If the preset size is reached, the posture correction is performed according to the true value of the label;

[0010] S4. When the comprehensive mining equipment moves to the left side of the tunnel side wall, the left positioner performs positioning and mapping, and the right positioner only performs mapping; conversely, when the comprehensive mining equipment moves to the right side of the tunnel side wall, the right positioner performs positioning and mapping, and the left positioner only performs mapping.

[0011] Furthermore, the multi-layer fusion image enhancement module first performs brightness judgment on the input image and only performs image enhancement on low-light images. For low-light images, the brightness channel of the input image is divided into a Retinex enhancement layer, a brightness enhancement layer, and a contrast enhancement layer. In the Retinex enhancement layer, guided filtering is used to replace Gaussian filtering for reflection component estimation, adaptive gamma correction is used for brightness enhancement in the brightness enhancement layer, and CLAHE is used for contrast enhancement in the contrast enhancement layer. After three layers of parallel processing, a fusion method based on detail information is used for image fusion, and finally a bilateral filtering algorithm is used for image denoising. The image processed by the image enhancement module can well improve the brightness and contrast of the image, which is conducive to extracting corner points.

[0012] Furthermore, the detailed information includes exposure, variance and entropy.

[0013] Furthermore, the label feature posture correction module eliminates the cumulative error caused by VI-SLAM positioning, separates the label using a color segmentation method, and determines whether the label reaches a preset size. If the preset size is reached, a reprojection error model is established based on the current estimated posture and the label true value point, and the corrected posture is calculated using the least squares method.

[0014] Furthermore, the movement direction of the comprehensive mining equipment is determined by IMU data of the left locator and the right locator.

[0015] Furthermore, the positioning result of the comprehensive mining equipment is based on the positioning result of the locator on the side of the movement direction. The locators are arranged on the left and right sides of the comprehensive mining equipment. Since the movement of the comprehensive mining equipment is reciprocating forward movement, when the movement is reversed, the support machine closer to the rear end will move forward, causing part of the terrain behind the movement of the comprehensive mining equipment to change. Therefore, when the comprehensive mining equipment moves in the opposite direction, the positioning result of the locator on the side of the movement direction shall be used as the basis.

[0016] Furthermore, the arrangement positions of the labels follow the following principle: ensuring that the camera can capture two adjacent labels at any time.

[0017] Furthermore, the label is made of fluorescent material and has strong water mist and dust penetration capabilities.

[0018] Compared with the prior art, the beneficial effects of the present invention are: relative to the existing underground positioning method, the underground comprehensive mining equipment positioning method based on synchronous positioning and composition provided by the present invention has greatly improved positioning accuracy. At the same time, the present invention mainly uses the camera and IMU for fusion positioning, and the equipment cost is relatively low. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flow chart of the VI-SLAM downhole positioning method of the present invention;

[0020] Figure 2 It is the positioning process of a single positioner;

[0021] Figure 3 It is the flow chart of the image enhancement algorithm module;

[0022] Figure 4 It is the label feature pose correction flow chart;

[0023] Figure 5 It is a bird's-eye view of the fully mechanized mining equipment environment;

[0024] Figure 6 This is a schematic diagram of the comprehensive mining equipment locator.

[0025] Among them, 1-tunnel side wall, 2-label, 3-coal wall, 4-excavation route of comprehensive mining equipment, 5-camera field of view, 6-left locator, 7-comprehensive mining equipment, 8-right locator. DETAILED DESCRIPTION

[0026] The specific implementation modes of the present invention are described in detail below in conjunction with the accompanying drawings, but it should be understood that the protection scope of the present invention is not limited by the specific implementation modes.

[0027] Example: Figure 1 The figure shows a flow chart of the underground positioning system. The IMU data is integrated to determine the direction of movement of the comprehensive mining equipment 7. Since the comprehensive mining equipment 7 will move forward after movement, if the rear locator in the direction of movement of the comprehensive mining equipment 7 is used for positioning, it is easy to cause many corner points to be mismatched or even positioning failure. Therefore, two locators are arranged on the left and right sides of the comprehensive mining equipment 7. The locator on the side in the direction of movement of the comprehensive mining equipment 7 performs positioning and mapping, and the locator on the other side only performs mapping and draws out the unchanged feature points at the far end for positioning during reverse movement. When the reverse movement of the comprehensive mining equipment 7 is detected, the two locators transfer posture. Figure 6 As shown, when the comprehensive mining equipment 7 moves to the right, the right positioner 8 is used for positioning and mapping, and the left positioner 6 is only used for mapping; when the comprehensive mining equipment 7 moves to the left, the left positioner 6 is used for positioning and mapping, and the right positioner 8 is only used for mapping; when the comprehensive mining equipment 7 moves in the opposite direction, the left and right positioners transfer posture.

[0028] Specifically, the locator positioning process is as follows: Figure 2 As shown, the following steps are included:

[0029] Step S1, arranging tags 2 at even intervals on the side wall 1 of the mine tunnel, and obtaining the true values ​​of multiple tags 2 using a total station;

[0030] Specifically, the tags 2 are evenly spaced on the sidewall 1 of the tunnel, and the arrangement should ensure that the camera can capture two adjacent tags 2 at any time. The true value of the tag 2 is obtained using a total station for subsequent posture correction;

[0031] Step S2, a fully mechanized mining device 7 is set in the mine tunnel, and a left locator 6 and a right locator 8 are arranged on the left and right sides of the fully mechanized mining device 7. The left locator 6 and the right locator 8 are positioned using the VI-SLAM method. The fully mechanized mining device 7 obtains the mine tunnel image through a camera, and a multi-layer fusion image enhancement processing module is added to the front end of the left locator 6 and the right locator 8 to improve the brightness and contrast of the acquired mine tunnel image and increase the number of corner point extraction;

[0032] Specifically, Figure 3 This is the image enhancement process. Traditional image enhancement algorithms mostly use a single method to enhance images. Although each method has its corresponding advantages, its disadvantages are also obvious. The image enhancement module of the present invention first judges the brightness of the input image and only performs image enhancement on low-light images. For low-light images, the brightness channel of the input image is divided into a Retinex enhancement layer, a brightness enhancement layer, and a contrast enhancement layer. In the Retinex enhancement layer, guided filtering is used to replace Gaussian filtering for reflection component estimation. In the brightness enhancement layer, adaptive gamma correction is used for brightness enhancement. In the contrast enhancement layer, CLAHE is used for contrast enhancement. Then, a fusion method based on detail information is used for image fusion. Finally, a bilateral filtering algorithm is used for image denoising. The image processed by the image enhancement module can greatly improve the brightness and contrast of the image, which is conducive to extracting corner points.

[0033] Step S3, the rear ends of the left locator 6 and the right locator 8 are both added with a tag feature posture correction module to eliminate the cumulative error caused by VI-SLAM positioning, separate the tag 2 using color segmentation, and determine whether the tag 2 reaches the preset size. If it reaches the preset size, the posture correction is performed according to the true value of the tag 2;

[0034] Specifically, first determine whether the initialization has been completed. If not, perform visual SFM and visual inertial navigation alignment. After successful initialization, the front end tracks the feature points according to the optical flow method to obtain matching, and the back end solves and optimizes the camera pose in a sliding window manner. Finally, the camera pose is obtained, which is the pose of the comprehensive mining equipment.

[0035] Since the label color is fixed, the label in the image is extracted using the color segmentation method to determine whether it reaches the predetermined size. If it reaches the predetermined size, posture correction is performed; Figure 4 As shown in the figure, p is a 3D label point, which has been obtained by the total station to obtain the true value, p1 is the projection point of point p according to the current camera posture, and p2 is the corner point extracted from the image at the current moment. Due to the cumulative error of VI-SLAM, p1 and p2 are not the same point, so the reprojection error error of point p is established. p =||p1-p2||2, the reprojection error of all label feature points is The corrected posture is obtained by minimizing the reprojection error using the least squares method;

[0036] Step S4, when the comprehensive mining equipment 7 moves to the left side of the tunnel side wall 1, the left positioner 6 performs positioning and mapping, and the right positioner 8 only performs mapping; conversely, when the comprehensive mining equipment 7 moves to the right side of the tunnel side wall 1, the right positioner 8 performs positioning and mapping, and the left positioner 6 only performs mapping.

[0037] The specific models of the above electronic components are not specially specified, and ordinary products available on the market can be selected as long as they can meet the use requirements of the present invention.

[0038] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and does not limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the protection scope of the present invention.

Claims

1. A method for positioning underground fully mechanized mining equipment based on synchronous positioning and mapping, characterized in that: The following steps are involved: S1, arranging tags (2) at even intervals on the side wall (1) of a mine tunnel, and obtaining true values ​​of a plurality of tags (2) using a total station; S2. A fully mechanized mining device (7) is installed in a mine tunnel, and a left locator (6) and a right locator (8) are arranged on the left and right sides of the fully mechanized mining device (7). The left locator (6) and the right locator (8) are positioned using a VI-SLAM method. The fully mechanized mining device (7) obtains an image of the mine tunnel through a camera, and a multi-layer fusion image enhancement processing module is added to the front end of the left locator (6) and the right locator (8) to improve the brightness and contrast of the obtained mine tunnel image and increase the number of corner points extracted; The multi-layer fusion image enhancement processing module first judges the brightness of the input image and only performs image enhancement on low-light images. For low-light images, the brightness channel of the input image is divided into a Retinex enhancement layer, a brightness enhancement layer, and a contrast enhancement layer. In the Retinex enhancement layer, guided filtering is used to replace Gaussian filtering for reflection component estimation. In the brightness enhancement layer, adaptive gamma correction is used for brightness enhancement. In the contrast enhancement layer, CLAHE is used for contrast enhancement. After three layers of parallel processing, a fusion method based on detail information is used for image fusion. Finally, a bilateral filtering algorithm is used for image denoising. The image processed by the multi-layer fusion image enhancement processing module can well improve the brightness and contrast of the image, which is conducive to extracting corner points. S3, adding a label feature posture correction module to the rear ends of the left locator (6) and the right locator (8), eliminating the cumulative error generated by VI-SLAM positioning, using color segmentation to separate the label (2), and judging whether the label (2) reaches a preset size. If the preset size is reached, the posture correction is performed according to the true value of the label (2); S4. When the comprehensive mining equipment (7) moves toward the left side of the tunnel side wall (1), the left positioner (6) performs positioning and mapping, and the right positioner (8) only performs mapping; conversely, when the comprehensive mining equipment (7) moves toward the right side of the tunnel side wall (1), the right positioner (8) performs positioning and mapping, and the left positioner (6) only performs mapping.

2. A method for positioning underground fully mechanized mining equipment based on synchronous positioning and mapping as claimed in claim 1, characterized in that: The detailed information includes exposure, variance and entropy.

3. The underground fully mechanized mining equipment positioning method based on synchronous positioning and mapping as claimed in claim 1, characterized in that: The label feature posture correction module eliminates the cumulative error generated by VI-SLAM positioning, separates the label (2) using a color segmentation method, and determines whether the label (2) reaches a preset size. If the preset size is reached, a reprojection error model is established based on the current estimated posture and the label true value point, and the corrected posture is calculated using a least squares method.

4. The method for positioning underground fully mechanized mining equipment based on synchronous positioning and mapping as claimed in claim 1, characterized in that: The movement direction of the comprehensive mining equipment (7) is determined by IMU data of the left positioner (6) and the right positioner (8).

5. The method for positioning underground fully mechanized mining equipment based on synchronous positioning and mapping as claimed in claim 1, characterized in that: The positioning result of the fully mechanized mining equipment (7) is based on the positioning result of the positioner on the moving direction side.

6. The method for positioning underground fully mechanized mining equipment based on synchronous positioning and mapping as claimed in claim 1, characterized in that: The arrangement positions of the tags (2) follow the following principle: ensuring that the camera can capture two adjacent tags (2) at any time.

7. The method for positioning underground fully mechanized mining equipment based on synchronous positioning and mapping as claimed in claim 1, characterized in that: The label (2) is made of fluorescent material and has strong water mist and dust penetration capabilities.

Images