A navigation and positioning device for a water tunnel inspection robot assisted by artificial road signs and its implementation method

Abstract

The present invention proposes a navigation and positioning device for a water conveyance tunnel inspection robot assisted by artificial landmarks and an implementation method. The method installs a marker with specific information as a landmark at a fixed position in the tunnel, and obtains the position information of the robot in the tunnel by the robot identifying and detecting the content of the landmark, thereby providing assistance to the robot's navigation and path planning. The present invention uses artificial landmarks to make up for the problem of missing features in the tunnel environment and realize visual information correction and positioning. At the same time, the landmarks can also serve as branch tunnel prompt information to provide guidance for the robot to enter and exit the branch tunnel, effectively avoiding the problem of difficulty in path planning of the robot in a complex tunnel environment.

Description

Technical Field

[0001] The present invention belongs to the technical field of underwater robot tunnel navigation and positioning, and in particular relates to a navigation and positioning device for a water conveyance tunnel inspection robot assisted by artificial landmarks and an implementation method thereof. Background Art

[0002] The reliability of navigation and positioning is the premise and guarantee for robots to carry out various operational tasks. However, the water transfer tunnel environment is complex and there are many branch tunnels. How to perform accurate navigation and positioning has always been a difficult problem that needs to be solved urgently.

[0003] Since the tunnel environment shields satellite signals, it is difficult to use satellite navigation and other means for navigation and positioning. Dead reckoning can only be performed through its own inertial navigation, DVL and other sensors, which inevitably leads to cumulative errors. Considering that the tunnel environment is dark and there are very few usable features, it is also difficult to correct positioning through visual means. Summary of the invention

[0004] The purpose of the present invention is to solve the problems in the prior art and propose a navigation and positioning device for a water conveyance tunnel inspection robot assisted by artificial landmarks and an implementation method. Markers with specific information are installed at fixed positions in the tunnel as landmarks, and the robot recognizes and detects the content of the landmarks to obtain the robot's position information in the tunnel, thereby providing assistance for the robot's navigation and path planning.

[0005] The present invention is realized by the following technical scheme. The present invention proposes a navigation and positioning device for a water conveyance tunnel inspection robot assisted by artificial landmarks. The device comprises a camera, a searchlight, a visual processing computer and an artificial landmark; the camera is connected to the visual processing computer through a network port for transmitting images acquired by the camera; the searchlight is connected to the visual processing computer through a serial port, and the brightness of the searchlight is controlled by the visual processing computer; the artificial landmark is installed on the top of the tunnel for assisting navigation and positioning; the searchlight and the camera are installed on a tunnel robot for acquiring images of the artificial landmarks; the visual processing computer is a tunnel robot processing system for processing and identifying acquired images and controlling the searchlight.

[0006] The present invention also proposes a method for implementing a navigation and positioning device for a water conveyance tunnel inspection robot assisted by artificial landmarks, the method comprising the following steps:

[0007] S1: Arrange artificial road signs and install the designed artificial road signs on the top of the tunnel;

[0008] S2: Initialize the robot position, establish a coordinate system at the entrance, and set the starting position as the coordinate origin;

[0009] S3: Calculate the robot position through the SINS / DVL integrated navigation system;

[0010] S4: Determine whether the robot is close to the position of the artificial landmark. If so, the robot surfaces and turns on the top camera and searchlight to obtain images. Otherwise, return to S3.

[0011] S5: Process and identify the image, send the acquired image to the visual processing computer for preprocessing, and perform positioning, segmentation and recognition operations on the processed image;

[0012] S6: Compare and analyze the recognized coding content with the database to obtain the specific meaning of the coding; when the recognized information is the main tunnel location information, use the location information contained in the road sign code to update the current positioning of the robot; when the recognized information is the branch tunnel prompt information, prepare the motion control for entering and exiting the tunnel in advance to achieve the optimal path planning.

[0013] Furthermore, the design of the artificial road signs in S1 specifically includes:

[0014] S11: The artificial road sign consists of a frame and a code; the frame and code are made of reflective materials, and the blank space is a non-reflective background board;

[0015] S12: The border shape is composed of a triangle and a rectangle. The triangle points to the left end, which can provide a directional feature for the road sign, facilitating the subsequent image tilt correction. The width and height information of the rectangular border are known, which facilitates the positioning of the road sign.

[0016] S13: The road sign code is in the form of numbers and letters. Each code has four characters, where the first digit represents the category information and the last three digits represent the location information. The road signs are divided into two categories. The first category is used to mark the location information of the main tunnel, and the interval between adjacent road signs is set to 1 km. The second category is the prompt information for entering and exiting the branch tunnel, which is used to prompt the robot to approach the branch tunnel so as to make control preparations in advance to avoid missing the branch tunnel.

[0017] Furthermore, the encoding method of the first type of road signs is as follows:

[0018] The first bit of the code is A, which indicates the location information of the main hole;

[0019] The last three digits are coded from 000-999, which can be used to represent the location information of 1000 road signs; 000 means the road sign is at the starting position, 001 means the road sign is 1 km away from the starting point, 002 means the road sign is 2 km away from the starting point, and so on.

[0020] Furthermore, the encoding method of the second type of road signs is as follows:

[0021] The first digit of the code is B, which means there is a branch tunnel entrance on the left side in front of the main tunnel;

[0022] The first digit of the code is C, which means there is a branch cave entrance on the right side in front of the main cave;

[0023] The first digit of the code is D, which means there are branch caves on both sides of the main cave.

[0024] Furthermore, the image processing and recognition process in S5 specifically includes:

[0025] S51: Binarize the acquired image. The binarization formula is as follows:

[0026]

[0027] Among them, g(x,y) represents the value of (x,y) in the binary image, v(x,y) represents the gray value, and T represents the threshold;

[0028] S52: using a four-directional search recursive algorithm on the binary image to search for all connected domains on the image, screening the connected domains according to the width, height, aspect ratio and border area feature information of the road sign, and determining the location of the road sign;

[0029] S53: Select the largest connected domain as the feature object, use the principal component analysis method to obtain the feature vector of the connected domain, take the largest feature vector as the main direction of the image, and complete the tilt correction of the image through the rotation matrix;

[0030] S54: Use Gaussian function to smooth the image and filter and reduce noise on the image; calculate the gradient amplitude and direction; use non-maximum method to suppress the amplitude and refine the edge; take double thresholds for the gradient and connect the edges;

[0031] S55: Use a method combining connected domain and projection to segment the coded characters of the road sign image;

[0032] S56: Recognize the segmented characters based on the template matching method. First, the segmented characters are normalized by bilinear interpolation to unify the size of the characters in the template library. The calculation formula is:

[0033] g(x,y)=f(x′,y′)=f[a(x,y),b(x,y)] (2)

[0034] Where f(x,y) represents the input image, g(x,y) represents the normalized output image, a(x,y) and b(x,y) represent the spatial mapping relationship from x to x′ and from y to y′ respectively;

[0035] Then the normalized characters are traversed and compared with the characters in the template library. The processing method is to first subtract it from the character template, and then calculate the total pixel value of the image after subtraction. If it is less than a certain threshold, it means that the character to be recognized and the template are the same character. By cyclically processing the characters to be recognized, all characters can be recognized.

[0036] The present invention uses artificial road signs to make up for the problem of missing features in the tunnel environment and realize visual information correction positioning. At the same time, the road signs can also serve as branch tunnel prompt information to provide guidance for the robot to enter and exit the branch tunnel, effectively avoiding the problem of difficult path planning for the robot in a complex tunnel environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A schematic diagram of the structure of a navigation and positioning device for a water tunnel inspection robot assisted by artificial landmarks;

[0038] Figure 2 A flow chart of the navigation and positioning method of a water tunnel inspection robot assisted by artificial landmarks;

[0039] Figure 3 Schematic diagram of the robot's process of identifying road signs;

[0040] Figure 4 It is a road sign style diagram;

[0041] Figure 5 This is a schematic diagram of the road sign coding scheme;

[0042] Figure 6 Schematic diagram of the road sign coding recognition process. DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] Combination Figure 1-6The present invention proposes a navigation and positioning device for a water conveyance tunnel inspection robot assisted by artificial landmarks, the device comprising a camera, a searchlight, a visual processing computer and an artificial landmark; the camera is connected to the visual processing computer via a network port for transmitting images acquired by the camera; the searchlight is connected to the visual processing computer via a serial port, and the brightness of the searchlight is controlled by the visual processing computer; considering that the water surface height in the water conveyance tunnel will not reach the top of the tunnel, the artificial landmark is installed at the top of the tunnel for assisting navigation and positioning; the searchlight and the camera are installed on the tunnel robot for acquiring images of the artificial landmarks; the visual processing computer is a tunnel robot processing system for processing and identifying the acquired images and controlling the searchlight.

[0045] The present invention also proposes a method for implementing a navigation and positioning device for a water conveyance tunnel inspection robot assisted by artificial landmarks, the method comprising the following steps:

[0046] S1: Arrange artificial road signs and install the designed artificial road signs on the top of the tunnel;

[0047] S2: Initialize the robot position, establish a coordinate system at the entrance, and set the starting position as the coordinate origin;

[0048] S3: Calculate the robot position through the SINS / DVL integrated navigation system;

[0049] S4: Determine whether the robot is close to the position of the artificial landmark. If so, the robot surfaces and turns on the top camera and searchlight to obtain images. Otherwise, return to S3.

[0050] S5: Process and identify the image, send the acquired image to the visual processing computer for preprocessing, and perform positioning, segmentation and recognition operations on the processed image;

[0051] S6: Compare and analyze the recognized coding content with the database to obtain the specific meaning of the coding; when the recognized information is the main tunnel location information, use the location information contained in the road sign code to update the current positioning of the robot; when the recognized information is the branch tunnel prompt information, prepare the motion control for entering and exiting the tunnel in advance to achieve the optimal path planning.

[0052] The design of the artificial road signs in S1 specifically includes:

[0053] S11: The artificial road sign consists of a frame and a code; the frame and code are made of reflective materials, and the blank space is a non-reflective background board;

[0054] S12: The border shape is composed of a triangle and a rectangle. The triangle points to the left end, which can provide a directional feature for the road sign, facilitating the subsequent image tilt correction. The width and height information of the rectangular border are known, which facilitates the positioning of the road sign.

[0055] S13: The road sign code is in the form of numbers and letters. Each code has four characters, where the first digit represents the category information and the last three digits represent the location information. The road signs are divided into two categories. The first category is used to mark the location information of the main tunnel, and the interval between adjacent road signs is set to 1 km. The second category is the prompt information for entering and exiting the branch tunnel, which is used to prompt the robot to approach the branch tunnel so as to make control preparations in advance to avoid missing the branch tunnel.

[0056] The encoding of the first type of road signs is as follows:

[0057] The first bit of the code is A, which indicates the location information of the main hole;

[0058] The last three digits are coded from 000-999, which can be used to represent the location information of 1000 road signs; 000 means the road sign is at the starting position, 001 means the road sign is 1 km away from the starting point, 002 means the road sign is 2 km away from the starting point, and so on.

[0059] For example, code A050 indicates a location in the main cave 50 kilometers from the starting point.

[0060] The encoding method of the second type of road signs is as follows:

[0061] The first digit of the code is B, which means there is a branch tunnel entrance on the left side in front of the main tunnel;

[0062] The first digit of the code is C, which means there is a branch cave entrance on the right side in front of the main cave;

[0063] The first digit of the code is D, which means there are branch caves on both sides of the main cave.

[0064] The last three digits of the code represent the distance to the cave entrance. 050, 030, and 010 represent 50m, 30m, and 10m away from the cave entrance respectively.

[0065] For example, code B050 means that there is an entrance to the branch cave from the left side 50m in front of the robot;

[0066] Code C010 means that there is an entrance to the branch cave from the right side 10m in front of the robot;

[0067] Code D030 means there are branch holes on both sides 30m in front of the robot.

[0068] The image processing and identification process in S5 specifically includes:

[0069] S51: Binarize the acquired image. The binarization formula is as follows:

[0070]

[0071] Among them, g(x,y) represents the value of (x,y) in the binary image, v(x,y) represents the gray value, and T represents the threshold;

[0072] S52: using a four-directional search recursive algorithm on the binary image to search for all connected domains on the image, screening the connected domains according to the width, height, aspect ratio and border area feature information of the road sign, and determining the location of the road sign;

[0073] S53: Select the largest connected domain (i.e., the road sign border) as the feature object, use the principal component analysis (PCA) method to obtain the eigenvector of the connected domain, take the largest eigenvector as the main direction of the image, and complete the image tilt correction through the rotation matrix;

[0074] S54: Use Gaussian function to smooth the image and filter and reduce noise on the image; calculate the gradient amplitude and direction; use non-maximum method to suppress the amplitude and refine the edge; take double thresholds for the gradient and connect the edges;

[0075] S55: Use a method combining connected domain and projection to segment the coded characters of the road sign image;

[0076] S56: Recognize the segmented characters based on the template matching method. First, the segmented characters are normalized by bilinear interpolation to unify the size of the characters in the template library. The calculation formula is:

[0077] g(x,y)=f(x′,y′)=f[a(x,y),b(x,y)] (2)

[0078] Where f(x,y) represents the input image, g(x,y) represents the normalized output image, a(x,y) and b(x,y) represent the spatial mapping relationship from x to x′ and from y to y′ respectively;

[0079] Then the normalized characters are traversed and compared with the characters in the template library. The processing method is to first subtract it from the character template, and then calculate the total pixel value of the image after subtraction. If it is less than a certain threshold, it means that the character to be recognized and the template are the same character. By cyclically processing the characters to be recognized, all characters can be recognized.

[0080] The above is a detailed introduction to the artificial landmark-assisted water tunnel inspection robot navigation and positioning device and implementation method proposed in the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for implementing a navigation and positioning device for a water conveyance tunnel inspection robot based on artificial landmark assistance, characterized in that: The device includes a camera, a searchlight, a visual processing computer and an artificial road sign; the camera is connected to the visual processing computer through a network port to transmit images acquired by the camera; the searchlight is connected to the visual processing computer through a serial port, and the brightness of the searchlight is controlled by the visual processing computer; the artificial road sign is installed on the top of the tunnel to assist navigation and positioning; the searchlight and the camera are installed on a tunnel robot to acquire images of the artificial road sign; the visual processing computer is a tunnel robot processing system, which is used to process and identify the acquired images and control the searchlight; The method comprises the following steps: S1: Arrange artificial road signs and install the designed artificial road signs on the top of the tunnel; S2: Initialize the robot position, establish a coordinate system at the entrance, and set the starting position as the coordinate origin; S3: Calculate the robot position through the SINS / DVL integrated navigation system; S4: Determine whether the robot is close to the position of the artificial landmark. If so, the robot surfaces and turns on the top camera and searchlight to obtain images. Otherwise, return to S3. S5: Process and identify the image, send the acquired image to the visual processing computer for preprocessing, and perform positioning, segmentation and recognition operations on the processed image; S6: Compare and analyze the recognized code content with the database to obtain the specific meaning of the code; when the recognized information is the main tunnel location information, use the location information contained in the road sign code to update the current positioning of the robot; when the recognized information is the branch tunnel prompt information, make good motion control in and out of the tunnel in advance to achieve the optimal path planning; The design of the artificial road signs in S1 specifically includes: S11: The artificial road sign consists of a frame and a code; the frame and code are made of reflective materials, and the blank space is a non-reflective background board; S12: The border shape is composed of a triangle and a rectangle. The triangle points to the left end, which can provide a directional feature for the road sign, facilitating the subsequent image tilt correction. The width and height information of the rectangular border are known, which facilitates the positioning of the road sign. S13: The road sign code is in the form of numbers and letters. Each code has four characters, of which the first digit represents the category information and the last three digits represent the location information. The road signs are divided into two categories. The first category is used to mark the location information of the main tunnel, and the interval between adjacent road signs is set to 1 km. The second category is the prompt information for entering and exiting the branch tunnel, which is used to prompt the robot to approach the branch tunnel position so as to make control preparations in advance to avoid missing the branch tunnel. The encoding of the first type of road signs is as follows: The first bit of the code is A, which indicates the location information of the main hole; The last three digits are 000-999, which can be used to represent the location information of 1000 road signs; 000 means the road sign is at the starting position, 001 means the road sign is 1 km away from the starting point, 002 means the road sign is 2 km away from the starting point, and so on; The encoding method of the second type of road signs is as follows: The first digit of the code is B, which means there is a branch tunnel entrance on the left side in front of the main tunnel; The first digit of the code is C, which means there is a branch cave entrance on the right side in front of the main cave; The first digit of the code is D, which means there are branch caves on both sides of the main cave.

2. The method according to claim 1, characterized in that: The image processing and identification process in S5 specifically includes: S51: Binarize the acquired image. The binarization formula is as follows: Among them, g(x,y) represents the value of (x,y) in the binary image, v(x,y) represents the gray value, and T represents the threshold; S52: using a four-directional search recursive algorithm on the binary image to search for all connected domains on the image, screening the connected domains according to the width, height, aspect ratio and border area feature information of the road sign, and determining the location of the road sign; S53: Select the largest connected domain as the feature object, use the principal component analysis method to obtain the feature vector of the connected domain, take the largest feature vector as the main direction of the image, and complete the tilt correction of the image through the rotation matrix; S54: Use Gaussian function to smooth the image and filter and reduce noise on the image; calculate the gradient amplitude and direction; use non-maximum method to suppress the amplitude and refine the edge; take double thresholds for the gradient and connect the edges; S55: Use a method combining connected domain and projection to segment the coded characters of the road sign image; S56: Recognize the segmented characters based on the template matching method. First, the segmented characters are normalized by bilinear interpolation to unify the size of the characters in the template library. The calculation formula is: g(x,y)=f(x′,y′)=f[a(x,y),b(x,y)] (2) Where f(x,y) represents the input image, g(x,y) represents the normalized output image, a(x,y) and b(x,y) represent the spatial mapping relationship from x to x′ and from y to y′ respectively; Then the normalized characters are traversed and compared with the characters in the template library. The processing method is to first subtract it from the character template, and then calculate the total pixel value of the image after subtraction. If it is less than a certain threshold, it means that the character to be recognized and the template are the same character. By cyclically processing the characters to be recognized, all characters can be recognized.

Images