Precise detection positioning device and method based on four-eye image ground penetrating radar
By integrating four-eye image, six-axis inertial measurement and GPS timing modules in ground penetrating radar, high-precision underground object positioning in environments of severe electromagnetic interference is achieved, the problem of large positioning errors in the existing technology is solved, and detection is improved Accuracy and robustness.
Patent Information
- Application Number
- CN202410857161.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-28
- Publication Date
- 2024-11-08
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing ground-penetrating radar positioning technology is difficult to achieve accurate positioning in urban areas with severe electromagnetic interference, especially when the radar vehicle is not driving in a straight line, resulting in a large deviation between the actual position of the detected object and the recorded position, which cannot meet underground abnormalities. Requirements for accurate positioning.
The precise detection and positioning device based on four-eye image ground penetrating radar is adopted. Through the distance measurement function of the four-eye vision module, combined with the six-axis inertial measurement module and the GPS timing module, the actual position of the detected object is obtained by using the weighted average fusion method. Information to achieve high-precision radar vehicle positioning.
It improves positioning accuracy and robustness, can accurately discover underground hidden bodies in complex road environments, reduces errors, enhances the understanding of the depth structure of the scene, and provides higher depth perception capabilities.
Smart Images

Figure CN118915051A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of underground object positioning, and in particular to a precise detection and positioning device and method based on a four-eye imaging ground penetrating radar. Background Art
[0002] With the full utilization of underground space, underground engineering construction is becoming more and more frequent, such as subway construction, underground pipeline corridor construction, underground commercial space construction and air defense engineering construction. These underground constructions often cause soil disturbance and groundwater changes under urban roads, thereby causing road collapse. The disasters caused by them bring immeasurable losses to life and property. Therefore, high-tech non-destructive detection technology must be used to conduct regular detection to timely discover underground cavities, water bags and poor geological bodies.
[0003] Ground penetrating radar technology is the most important non-destructive and rapid detection technology for hidden defects underground in roads. It uses vehicles to tow ground penetrating radar antennas to collect data along the road for high-resolution detection, and has achieved many gratifying results. In the process of ground penetrating radar detection data collection, the precise positioning of its antenna is the key link to accurately discover underground hidden bodies. Previous positioning technologies mainly include ground penetrating radar antennas equipped with GPS and distance measuring wheel positioning. For antennas moving at a speed of 20 to 30 km per hour, the GPS precise positioning error is greater than 1m, especially in urban areas with severe electromagnetic interference, the error will be further increased. Distance measuring wheel positioning is simple, direct and easy to implement, but due to the complexity of road conditions, it is impossible to ensure that the ground penetrating radar tractor drives straight, and it often moves in a serpentine manner, resulting in a large gap between the distance measured by the distance measuring wheel and the straight-line distance, especially the accumulation of errors, which causes the error to gradually increase.
[0004] It can be seen that the above-mentioned ranging and positioning method cannot meet the requirements of accurate positioning of underground anomalies, resulting in inaccurate positioning of holes for subsequent drilling verification or inability to carry out disaster prevention treatment normally. Summary of the invention
[0005] In order to solve the problem that the actual position of the detected object deviates from the recorded position due to the non-linear motion of the radar vehicle while driving, the present invention provides a precise detection and positioning device and method based on a four-eye imaging ground penetrating radar. Through the distance measurement function of the four-eye vision module, the relative position relationship between the radar vehicle and the target object when the object is detected is accurately calculated. By deducing this relationship, the actual position information of the detected object can be accurately obtained, thereby achieving high-precision radar vehicle positioning.
[0006] To achieve the above-mentioned purpose, the present invention provides a precise detection and positioning device based on a four-eye imaging ground penetrating radar, comprising a four-eye vision acquisition module, a radar detection module, a six-axis inertial measurement module, a GPS timing module, a microcontroller module and a power management module placed on a radar vehicle, wherein the four-eye vision acquisition module, the radar detection module, the six-axis inertial measurement module and the GPS timing module are all connected to the microcontroller module;
[0007] Among them, the four-eye vision acquisition module is used to acquire the four-eye vision image of the target object;
[0008] A radar detection module, used to collect underground location information of detected objects;
[0009] Six-axis inertial measurement module, used to collect driving information of the radar car;
[0010] GPS timing module, used to synchronize the acquisition time of the radar detection module and the four-eye vision acquisition module, and collect the location information of the radar vehicle;
[0011] The microcontroller module is used to obtain the positional relationship of the detected object on the road surface by using a weighted average fusion method based on the four-eye vision image of the target object, the underground position information of the detected object, the position information of the radar vehicle and the driving information of the radar vehicle.
[0012] Preferably, the four-eye vision acquisition module includes four monocular cameras placed on the same plane.
[0013] Preferably, the driving information includes attitude and acceleration data of the radar vehicle.
[0014] Preferably, the microcontroller module communicates with the server via the communication module, and the communication module is equipped with the MQTT protocol.
[0015] The method for accurately detecting and locating a device based on a four-eye imaging ground penetrating radar comprises the following steps:
[0016] S1. Parameter calibration:
[0017] The Zhang Zhengyou calibration method is used to obtain the internal parameters, distortion parameters and external parameters of each monocular camera of the four-eye vision acquisition module, and the camera is calibrated;
[0018] S2. Synchronous information time:
[0019] Using the GPS timing module, based on the synchronization method of the post-synchronized second pulse clock, the acquisition time of the radar detection module and the four-eye vision acquisition module is synchronized;
[0020] S3. Information collection:
[0021] Searching the time point when the underground detected object is collected in the underground position information of the detected object collected by the radar detection module, and extracting the four-eye vision image of the target object at the same time point to repair the distortion;
[0022] At the same time, the six-axis inertial measurement module is used to collect the driving information of the radar car;
[0023] S4. Position estimation:
[0024] Based on the four-eye vision image, the underground position information of the detected object and the driving information of the radar vehicle, the weighted average fusion method is used to obtain the relative position relationship between the target object and the radar vehicle, and then the position information of the radar vehicle is used to obtain the position information of the detected object on the road.
[0025] Preferably, step S4 specifically includes the following steps:
[0026] S41, Stereo Matching:
[0027] The four visual images acquired by the four monocular cameras of the four-eye vision acquisition module are respectively extracted with feature points for stereo matching. Then, the visual difference is calculated using every two visual images. The internal and external parameters of each monocular camera are combined to transform the monocular camera coordinate system into a three-dimensional coordinate system. The three-dimensional point cloud data of each set of visual images is generated and merged to obtain a dense three-dimensional point cloud.
[0028] S42, obtaining pixel coordinates of the target object in each monocular camera, and converting the pixel coordinates in each monocular camera into three-dimensional coordinates between every two monocular cameras, to obtain six three-dimensional coordinates;
[0029] S43, performing weighted averaging on the obtained six three-dimensional coordinates, dense three-dimensional point cloud, position information of the radar vehicle and driving information of the radar vehicle to obtain the relative position relationship between the target object and the radar vehicle, thereby obtaining the actual coordinates of the radar vehicle;
[0030] The weighted average calculation formula is as follows:
[0031]
[0032] Where, X avg is the result of weighted average, w i is the weight of the i-th data source, X i is the data of the i-th data source, and n is the number of data sources;
[0033] S44. The position information of the detected object on the road surface is obtained by measuring the actual position information of the target object a priori or subsequently in the field, and combining the relative position relationship between the radar vehicle and the target object when the radar vehicle detects the underground detection object.
[0034] Preferably, in step S41, firstly, the underground position information of the detected object is searched in the data collected by the radar detection module, and the detection time is recorded; and the four-eye vision image at the same time is searched in the data collected by the four-eye vision acquisition module;
[0035] Then, the six intersection planes obtained from every two viewing angles in the four-eye vision image are used;
[0036] Finally, based on the six intersection planes, stereo matching is performed through the internal and external parameters of each monocular camera obtained previously to obtain the disparity map and depth information.
[0037] Preferably, in step S42, a manual region of interest method is first used to reduce the processing area of the four-eye vision acquisition module, and then deep learning or template matching is used to obtain the pixel coordinates of the underground target object in each monocular camera.
[0038] Preferably, the pixel coordinates of the target object in each monocular camera are obtained by using the deep learning of Faster-RCNN or grayscale template matching.
[0039] Preferably, the acquired pixel coordinates are matched with the result of stereo matching to obtain six three-dimensional coordinates of the target object between six intersection planes of every two viewing angles.
[0040] The present invention has the following beneficial effects:
[0041] 1. Higher accuracy and robustness: The four-eye vision acquisition module has more information, so it can find corresponding points more accurately in the stereo matching process, improve the depth of the measured distance, and have stronger robustness to texture loss, lighting effects, etc.;
[0042] 2. Stronger depth perception: More visual points are used to provide more depth information, thereby improving the understanding of the scene's depth structure, and having more advantages in dealing with complex three-dimensional environments such as roads where there may be occlusions;
[0043] 3. Provides a second pulse clock synchronization method based on post-synchronization, unifying the time of radar data and four-eye vision data.
[0044] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] Figure 1 It is a principle block diagram of a precise detection and positioning device based on a four-eye imaging ground penetrating radar of the present invention;
[0046] Figure 2The present invention is a flow chart of a method for a precise detection and positioning device based on a four-eye imaging ground penetrating radar. DETAILED DESCRIPTION
[0047] In order to make the purpose, technical scheme and advantages disclosed in the embodiments of the present invention clearer, the embodiments of the present invention are further described in detail in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the embodiments of the present invention and are not intended to limit the embodiments of the present invention. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. Examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions.
[0048] It should be noted that the terms "include", "have" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products or devices.
[0049] Like reference numerals and letters denote similar items in the following drawings, and thus, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.
[0050] In the description of the present invention, it should be noted that the terms "upper", "lower", "inside", "outside", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings, or are directions or positional relationships in which the product of the invention is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as a limitation on the present invention.
[0051] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "setting", "installation" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0052] like Figure 1As shown, a precise detection and positioning device based on a four-eye imaging ground penetrating radar includes a four-eye vision acquisition module, a radar detection module, a six-axis inertial measurement module, a GPS timing module, a microcontroller module and a power management module placed on a radar vehicle, wherein the four-eye vision acquisition module, the radar detection module, the six-axis inertial measurement module and the GPS timing module are all connected to the microcontroller module; wherein the four-eye vision acquisition module is used to acquire a four-eye visual image of a target object, and the four-eye vision acquisition module includes four monocular cameras placed on the same plane. In this embodiment, the target object is a fixed object on both sides of the road, such as trees, lamp poles, etc.; the radar detection module is used to acquire the underground position information of the detected object; the six-axis inertial measurement module is used to acquire the driving information of the radar vehicle; the driving information includes the attitude and acceleration data of the radar vehicle. The GPS timing module is used to synchronize the acquisition time of the radar detection module and the four-eye vision acquisition module, and to collect the position information of the radar vehicle with an accuracy of 1 meter; the micro-control module is used to obtain the position relationship of the detected object on the road surface based on the four-eye vision image of the target object, the underground position information of the detected object, the position information of the radar vehicle and the driving information of the radar vehicle, using the weighted average fusion method.
[0053] The microcontroller module communicates with the server via the communication module, and the communication module is equipped with the MQTT protocol.
[0054] like Figure 2 As shown, the method for accurately detecting and positioning a device based on a four-eye imaging ground penetrating radar comprises the following steps:
[0055] S1. Parameter calibration:
[0056] The Zhang Zhengyou calibration method is used to obtain the internal parameters, distortion parameters and external parameters of each monocular camera of the four-eye vision acquisition module, and the camera is calibrated;
[0057] S2. Synchronous information time:
[0058] Using the GPS timing module, based on the synchronization method of the post-synchronized second pulse clock, the acquisition time of the radar detection module and the four-eye vision acquisition module is synchronized, and the synchronization start time accuracy reaches ±30ns;
[0059] S3. Information collection:
[0060] Searching the time point when the underground detected object is collected in the underground position information of the detected object collected by the radar detection module, and extracting the four-eye vision image of the target object at the same time point to repair the distortion;
[0061] At the same time, the six-axis inertial measurement module is used to collect the driving information of the radar car;
[0062] S4. Position estimation:
[0063] Based on the four-eye vision image, the underground position information of the detected object and the driving information of the radar vehicle, the weighted average fusion method is used to obtain the relative position relationship between the target object and the radar vehicle, and then the position information of the radar vehicle is used to obtain the position information of the detected object on the road.
[0064] Step S4 specifically includes the following steps:
[0065] S41, Stereo Matching:
[0066] The four visual images acquired by the four monocular cameras of the four-eye vision acquisition module are respectively extracted with feature points for stereo matching. Then, the visual difference is calculated using every two visual images. The internal and external parameters of each monocular camera are combined to transform the monocular camera coordinate system into a three-dimensional coordinate system. The three-dimensional point cloud data of each set of visual images is generated and merged to obtain a dense three-dimensional point cloud.
[0067] In step S41, firstly, the underground position information of the detected object is searched in the data collected by the radar detection module, and the detection time is recorded; and the four-eye vision image at the same time is searched in the data collected by the four-eye vision acquisition module;
[0068] Then, the six intersection planes obtained from every two viewing angles in the four-eye vision image are used;
[0069] Finally, based on the six intersection planes, stereo matching is performed through the internal and external parameters of each monocular camera obtained previously to obtain the disparity map and depth information.
[0070] S42, obtaining pixel coordinates of the target object in each monocular camera, and converting the pixel coordinates in each monocular camera into three-dimensional coordinates between every two monocular cameras, to obtain six three-dimensional coordinates;
[0071] In step S42, the manual region of interest method is first used to reduce the processing area of the four-eye vision acquisition module, and then the pixel coordinates of the underground target object in each monocular camera are obtained by deep learning or template matching.
[0072] Use Faster-RCNN deep learning or grayscale template matching to obtain the pixel coordinates of the target object in each monocular camera.
[0073] The obtained pixel coordinates are matched with the results of stereo matching to obtain six three-dimensional coordinates of the target object between the six intersection planes of each two viewing angles.
[0074] S43, performing weighted averaging on the obtained six three-dimensional coordinates, dense three-dimensional point cloud, position information of the radar vehicle and driving information of the radar vehicle to obtain the relative position relationship between the target object and the radar vehicle, thereby obtaining the actual coordinates of the radar vehicle;
[0075] The weighted average calculation formula is as follows:
[0076]
[0077] Where, X avg is the result of weighted average, w i is the weight of the i-th data source, X i is the data of the i-th data source, and n is the number of data sources;
[0078] In this embodiment, data source n=9, and data source X i They are six three-dimensional coordinates, the radar vehicle coordinates under dense three-dimensional point cloud, the position coordinates of the radar vehicle and the coordinates from the starting point under the action of driving information; the above nine data sources all indicate the position information of the radar vehicle from its perspective at this time.
[0079] It should be noted that the three-dimensional coordinates obtained by four-eye matching and the driving information of the radar car are the main discriminant data, and the dense three-dimensional point cloud and the GPS position information of the radar car are the reference data. Therefore, the weight w is set i When the weight of the main discriminant data is increased.
[0080] S44. The position information of the detected object on the road surface is obtained by measuring the actual position information of the target object a priori or subsequently in the field, and combining the relative position relationship between the radar vehicle and the target object when the radar vehicle detects the underground detection object.
[0081] Therefore, the present invention adopts the above-mentioned precise detection and positioning device and method based on four-eye imaging ground penetrating radar, realizes a vehicle-mounted two-dimensional or three-dimensional ground penetrating radar antenna, and when dragging and moving detection on the road, it can obtain information with the help of the four-eye vision acquisition module arranged on the antenna, thereby realizing precise positioning using intelligent technology.
[0082] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.
Claims
1. A precise detection and positioning device based on a four-eye imaging ground penetrating radar, characterized in that: It includes a four-eye vision acquisition module, a radar detection module, a six-axis inertial measurement module, a GPS timing module, a microcontroller module and a power management module placed on the radar vehicle, and the four-eye vision acquisition module, the radar detection module, the six-axis inertial measurement module and the GPS timing module are all connected to the microcontroller module; Among them, the four-eye vision acquisition module is used to acquire the four-eye vision image of the target object; A radar detection module, used to collect underground location information of detected objects; Six-axis inertial measurement module, used to collect driving information of the radar car; GPS timing module, used to synchronize the acquisition time of the radar detection module and the four-eye vision acquisition module, and collect the location information of the radar vehicle; The microcontroller module is used to obtain the positional relationship of the detected object on the road surface by using a weighted average fusion method based on the four-eye vision image of the target object, the underground position information of the detected object, the position information of the radar vehicle and the driving information of the radar vehicle.
2. According to claim 1, a precise detection and positioning device based on a four-eye imaging ground penetrating radar is characterized in that: The four-eye vision acquisition module includes four monocular cameras placed on the same plane.
3. The precise detection and positioning device based on four-eye imaging ground penetrating radar according to claim 1 is characterized in that: Driving information includes the attitude and acceleration data of the radar car.
4. The precise detection and positioning device based on four-eye imaging ground penetrating radar according to claim 1 is characterized in that: The microcontroller module communicates with the server via the communication module, and the communication module is equipped with the MQTT protocol.
5. A method for accurate detection and positioning of a device based on a four-eye imaging ground penetrating radar according to any one of claims 1 to 4, characterized in that: The following steps are involved: S1. Parameter calibration: The Zhang Zhengyou calibration method is used to obtain the internal parameters, distortion parameters and external parameters of each monocular camera of the four-eye vision acquisition module, and the camera is calibrated; S2. Synchronous information time: Using the GPS timing module, based on the synchronization method of the post-synchronized second pulse clock, the acquisition time of the radar detection module and the four-eye vision acquisition module is synchronized; S3. Information collection: Searching the time point when the underground detected object is collected in the underground position information of the detected object collected by the radar detection module, and extracting the four-eye vision image of the target object at the same time point to repair the distortion; At the same time, the six-axis inertial measurement module is used to collect the driving information of the radar car; S4. Position estimation: Based on the four-eye vision image, the underground position information of the detected object and the driving information of the radar vehicle, the weighted average fusion method is used to obtain the relative position relationship between the target object and the radar vehicle, and then the position information of the radar vehicle is used to obtain the position information of the detected object on the road.
6. The method of the precise detection and positioning device based on the four-eye imaging ground penetrating radar according to claim 5 is characterized in that: Step S4 specifically includes the following steps: S41, Stereo Matching: The four visual images acquired by the four monocular cameras of the four-eye vision acquisition module are respectively extracted with feature points for stereo matching. Then, the visual difference is calculated using every two visual images. The internal and external parameters of each monocular camera are combined to transform the monocular camera coordinate system into a three-dimensional coordinate system. The three-dimensional point cloud data of each set of visual images is generated and merged to obtain a dense three-dimensional point cloud. S42, obtaining pixel coordinates of the target object in each monocular camera, and converting the pixel coordinates in each monocular camera into three-dimensional coordinates between every two monocular cameras, to obtain six three-dimensional coordinates; S43, performing weighted averaging on the obtained six three-dimensional coordinates, dense three-dimensional point cloud, position information of the radar vehicle and driving information of the radar vehicle to obtain the relative position relationship between the target object and the radar vehicle, thereby obtaining the actual coordinates of the radar vehicle; The weighted average calculation formula is as follows: Where, X avg is the result of weighted average, w i is the weight of the i-th data source, X i is the data of the i-th data source, and n is the number of data sources; S44. The position information of the detected object on the road surface is obtained by measuring the actual position information of the target object a priori or subsequently in the field, and combining the relative position relationship between the radar vehicle and the target object when the radar vehicle detects the underground detection object.
7. The method of the precise detection and positioning device based on the four-eye imaging ground penetrating radar according to claim 6 is characterized in that: In step S41, firstly, the underground position information of the detected object is searched in the data collected by the radar detection module, and the detection time is recorded; and the four-eye vision image at the same time is searched in the data collected by the four-eye vision acquisition module; Then, the six intersection planes obtained from every two viewing angles in the four-eye vision image are used; Finally, based on the six intersection planes, stereo matching is performed through the internal and external parameters of each monocular camera obtained previously to obtain the disparity map and depth information.
8. The method of the precise detection and positioning device based on the four-eye imaging ground penetrating radar according to claim 6 is characterized in that: In step S42, the manual region of interest method is first used to reduce the processing area of the four-eye vision acquisition module, and then the pixel coordinates of the underground target object in each monocular camera are obtained by deep learning or template matching.
9. The method of the precise detection and positioning device based on the four-eye imaging ground penetrating radar according to claim 8, characterized in that: Use Faster-RCNN deep learning or grayscale template matching to obtain the pixel coordinates of the target object in each monocular camera.
10. The method of the precise detection and positioning device based on the four-eye imaging ground penetrating radar according to claim 9, characterized in that: The obtained pixel coordinates are matched with the results of stereo matching to obtain six three-dimensional coordinates of the target object between the six intersection planes of each two viewing angles.
Citation Information
Patent Citations
Four-eye visual positioning system-based positioning method
CN109269473A
High-precision map data acquisition device and method
CN112556654A
Navigation map making and using method and system based on radar image
CN115015911A
Three-dimensional measurement technical method based on four-eye structured light
CN116164667A
Overground and underground integrated intelligent mobile detection system
CN117724089A
