Method and system for preventing mis-touch of live equipment

Abstract

The present application relates to a kind of live equipment anti-mis-touch method and system.The method includes: obtaining camera and radar data;Determine human posture data;If human posture data is empty set, reacquire camera image and radar data;If it is non-empty set, determine the first calculation result of each key point;If any key point first calculation result is less than 0, determine the radar data in the 120 degree range close to ground in radar data and record the radar data under normal no foreign matter condition;In turn traverse the radar data in the 120 degree range close to ground, if there is less than 0.5 times radar data under normal no foreign matter condition, send early warning information;If key point first calculation result is all greater than 0, calculate second calculation result;If key point second calculation result is less than or equal to set threshold, send early warning information, if greater than set threshold, reacquire camera image and radar data.The present application can early warn personnel approach and alarm when approaching live equipment.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power equipment safety, and particularly relates to a method and system for preventing mistaken touch of live equipment. BACKGROUND

[0002] There are a large number of live equipment in the environment of power plants, substations and the like. In the actual maintenance and equipment installation process, the staff is prone to mistakenly approach the live equipment or even mistakenly touch the live equipment. The existing means for preventing mistaken touch of live equipment include setting warning signs, isolation fences, infrared and laser radar alarm devices. However, the warning signs and isolation fences are prone to be ignored by the staff objectively or subjectively, thereby losing the protection effect. The infrared and laser radar alarm devices can achieve alarm when the staff approaches the live equipment, but due to the short working distance, they cannot achieve early warning, and the infrared and laser radar are used to detect the approach of the staff by ranging, which may cause false alarms. SUMMARY

[0003] The purpose of the present application is to provide a method and system for preventing mistaken touch of live equipment to solve the problems of live equipment that cannot be warned in advance and false alarms.

[0004] To achieve the above purpose, the present application provides the following solutions.

[0005] A method for preventing mistaken touch of live equipment, comprising:

[0006] estimating an equation of a scanning plane of a laser radar in a camera coordinate system;

[0007] acquiring a camera image through a camera and radar data through a laser radar at any time;

[0008] processing the camera image by using a three-dimensional human body posture estimation algorithm to determine human body posture data;

[0009] if the human body posture data is an empty set, returning to the step of acquiring the camera image and the radar data to reacquire the camera image and the radar data;

[0010] if the human body posture data is a non-empty set, the human body posture data includes three-dimensional coordinates of a plurality of key points, and a key point first calculation result of each key point is determined according to the equation and the three-dimensional coordinates; the key point first calculation result is used to determine whether the human body key point crosses the scanning plane of the laser radar;

[0011] if any key point first calculation result is less than 0, radar data in a 120-degree range close to the ground in the radar data is determined according to a scanning start angle and a mounting angle of the laser radar, and the radar data under the normal non-foreign matter condition is recorded;

[0012] sequentially traversing the radar data within the 120-degree range close to the ground, if there is radar data less than 0.5 times the radar data under the normal no-foreign object condition in the radar data within the 120-degree range close to the ground, determining that the personnel have the risk of mis-touching the live equipment, issuing a warning information, and returning to the step of "acquiring camera images through the camera and radar data through the laser radar at any time"; if there is no radar data less than 0.5 times the radar data under the normal no-foreign object condition in the radar data within the 120-degree range close to the ground, returning to the step of "acquiring camera images through the camera and radar data through the laser radar at any time";

[0013] If all the key point first calculation results are greater than 0, calculating a key point second calculation result of each key point;

[0014] If the key point second calculation result is less than or equal to a set key point second calculation result threshold, issuing a warning information; if the key point second calculation result is greater than the set key point second calculation result threshold, returning to the step of "acquiring camera images through the camera and radar data through the laser radar at any time".

[0015] Optionally, the equation of the scanning plane of the laser radar in the camera coordinate system comprises:

[0016] The personnel stand at the scanning plane of the laser radar, the body is parallel to the scanning plane, and the position is transformed left and right, for each standing position of the personnel, an standing position image is captured by the camera;

[0017] The standing position image is processed by a three-dimensional human pose estimation algorithm to obtain standing position human pose data;

[0018] The steps of "the personnel stand at the scanning plane of the laser radar, the body is parallel to the scanning plane, and the position is transformed left and right, for each standing position of the personnel, an standing position image is captured by the camera" and "the standing position image is processed by a three-dimensional human pose estimation algorithm to obtain standing position human pose data" are repeated multiple times to obtain multiple standing position human pose data;

[0019] A linear equation set is constructed according to the multiple standing position human pose data;

[0020] The linear equation set is solved by a singular value decomposition method to determine the equation of the scanning plane of the laser radar in the camera coordinate system.

[0021] Optionally, the key point first calculation result is:

[0022] wherein, is an x-axis coordinate of the key point, is a y-axis coordinate of the key point, is a z-axis coordinate of the key point, i is a serial number of human pose data, j is a serial number of a key point in each human pose data, a is a first coefficient, b is a second coefficient, c is a third coefficient, and d is a fourth coefficient.

[0023] Optionally, the second calculation result of the key point is:

[0024]

[0025] wherein d i,j is a distance from the key point to the scanning plane.

[0026] A mis-touch prevention system for live equipment, comprising:

[0027] an equation estimation module configured to estimate an equation of a scanning plane of a laser radar in a camera coordinate system;

[0028] a camera image and radar data acquisition module configured to acquire a camera image through a camera and radar data through a laser radar at any time;

[0029] a human pose data determination module configured to determine human pose data by processing the camera image using a three-dimensional human pose estimation algorithm;

[0030] a step return module configured to return to the step of acquiring the camera image and the radar data through the camera and the laser radar again to acquire the camera image and the radar data again if the human pose data is an empty set;

[0031] a first calculation result determination module configured to determine a first calculation result of each key point according to the equation and three-dimensional coordinates of the key point if the human pose data is a non-empty set and the human pose data includes the three-dimensional coordinates of the key point, wherein the first calculation result of the key point is used to determine whether the human key point crosses the scanning plane of the laser radar;

[0032] a radar data determination module configured to determine radar data close to the ground within a 120-degree range in the radar data and record the radar data under a normal non-foreign matter condition according to a scanning start angle and a mounting angle of the laser radar if any first calculation result of the key point is less than 0;

[0033] The first early warning module is configured to sequentially traverse the radar data within the 120-degree range close to the ground, and if there is radar data less than 0.5 times the radar data under the normal no-foreign object condition in the radar data within the 120-degree range close to the ground, it is determined that the person has a risk of miscontacting a live device, an early warning information is sent, and the step of "acquiring a camera image through a camera and acquiring radar data through a laser radar at any time" is returned; if there is no radar data less than 0.5 times the radar data under the normal no-foreign object condition in the radar data within the 120-degree range close to the ground, the step of "acquiring a camera image through a camera and acquiring radar data through a laser radar at any time" is returned.

[0034] The key point second calculation result calculation module is configured to calculate a key point second calculation result of each key point if all the key point first calculation results are greater than 0.

[0035] The second early warning module is configured to send an early warning information if the key point second calculation result is less than or equal to a set key point second calculation result threshold; and return the step of "acquiring a camera image through a camera and acquiring radar data through a laser radar at any time" if the key point second calculation result is greater than the set key point second calculation result threshold.

[0036] Optionally, the equation estimation module specifically comprises:

[0037] The standing position image acquisition unit is configured to make the person stand at the scanning plane of the laser radar, make the body parallel to the scanning plane, move left and right to change positions, and for each standing position of the person, use the camera to take a standing position image.

[0038] The standing position human body posture data determination unit is configured to process the standing position image by using a three-dimensional human body posture estimation algorithm to obtain standing position human body posture data.

[0039] The plurality of standing position human body posture data determination units are configured to repeat the steps of "making the person stand at the scanning plane of the laser radar, making the body parallel to the scanning plane, moving left and right to change positions, and for each standing position of the person, using the camera to take a standing position image" and "processing the standing position image by using a three-dimensional human body posture estimation algorithm to obtain standing position human body posture data" multiple times to obtain a plurality of standing position human body posture data.

[0040] The linear equation set construction unit is configured to construct a linear equation set according to the plurality of standing position human body posture data.

[0041] The equation estimation unit is configured to solve the linear equation set by using a singular value decomposition method to determine an equation of the scanning plane of the laser radar in the camera coordinate system.

[0042] Optionally, the first calculation result of the key point is:

[0043] wherein, is the x-axis coordinate of the key point, is the y-axis coordinate of the key point, is the z-axis coordinate of the key point, i is the serial number of the human body posture data, j is the serial number of the key point in each human body posture data, a is a first coefficient, b is a second coefficient, c is a third coefficient, and d is a fourth coefficient.

[0044] Optionally, the second calculation result of the key point is:

[0045]

[0046] wherein, d i,j is the distance from the key point to the scanning plane.

[0047] According to the specific embodiments provided by the present application, the following technical effects are disclosed: the present application provides a live equipment anti-mis-touch method and system, which utilizes the characteristics of large visual field range and rich information obtained by camera visual recognition technology, and combines the characteristics of high accuracy of laser radar for close-range target detection, establishes the connection between the camera coordinate system and the laser radar coordinate system, realizes the joint calibration of the camera and the laser radar, and early warns personnel to approach and get close to the live equipment, thereby solving the false alarm problem caused by the separate use of laser radar. BRIEF DESCRIPTION OF DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0049] Figure 1 Fig. 1 is a schematic diagram of a live equipment anti-mis-touch device provided by the present application, which comprises a camera and a laser radar;

[0050] Figure 2 Fig. 2 is a flowchart of a live equipment anti-mis-touch method provided by the present application;

[0051] Figure 3 Fig. 3 is a schematic diagram of a personnel standing position provided by the present application;

[0052] Figure 4 Fig. 4 is a simplified flowchart of a live equipment anti-mis-touch method provided by the present application;

[0053] Figure 5The structure diagram of the mis-touch prevention system for the live equipment provided by the application. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the application will be apparently and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all the other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the application.

[0055] The purpose of the application is to provide a mis-touch prevention method and system for live equipment, which realizes early warning for personnel approaching and alarm when approaching the live equipment, and solves the problem of false alarm caused by the single use of laser radar.

[0056] In order to make the above-mentioned purpose, characteristics and advantages of the application more apparent and easy to understand, the application will be further described in detail below with reference to the drawings and specific embodiments.

[0057] The device composed of laser radar, camera and computer and the installation mode thereof, the laser radar is placed above the equipment at a distance of 30-50 cm, the scanning surface of the laser radar is perpendicular to the ground and parallel to the live equipment; the camera is opposite to the live equipment, and the field of view includes the entire monitoring area; the camera and the laser radar are connected with the computer, as shown in Figure 1 .

[0058] Figure 2 The mis-touch prevention method for live equipment provided by the application is shown in Figure 2 , and the mis-touch prevention method for live equipment comprises the following steps.

[0059] Step 201: estimating the equation of the scanning plane of the laser radar in the camera coordinate system.

[0060] The step 201 specifically comprises: a person stands on the scanning plane of the laser radar, the body is parallel to the scanning plane, and the position is changed by moving left and right, for each standing position of the person, a standing position image is captured by using the camera; the standing position image is processed by using a three-dimensional human body posture estimation algorithm to obtain human body posture data of the standing position; the steps of “a person stands on the scanning plane of the laser radar, the body is parallel to the scanning plane, and the position is changed by moving left and right, for each standing position of the person, a standing position image is captured by using the camera” and “the standing position image is processed by using a three-dimensional human body posture estimation algorithm to obtain human body posture data of the standing position” are repeated multiple times to obtain multiple human body posture data of the standing position; a linear equation set is constructed according to the multiple human body posture data of the standing position; the linear equation set is solved by using a singular value decomposition method to determine the equation of the scanning plane of the laser radar in the camera coordinate system.

[0061] In practical applications, step 201 specifically includes:

[0062] 1) The person stands on the scanning plane of the lidar, with their body roughly parallel to the scanning plane, and then moves left and right to change their position, such as... Figure 3 As shown.

[0063] 2) For each standing position k of the person, the camera takes one image Ic. k .

[0064] 3) Process the image using the SimpleBaseline3D algorithm or other 3D human pose estimation algorithms. k Obtain human posture data The lth key point

[0065] 4) Repeat steps 1)-2) M times to obtain M standing human posture data.

[0066] 5) Construct a system of linear equations:

[0067]

[0068] 6) The equation of the scanning plane of the lidar in the camera coordinate system is obtained by using the singular value decomposition method: ax + by + cz + d = 0; where a is the first coefficient, b is the second coefficient, c is the third coefficient, and d is the fourth coefficient; ax + by + cz + d = 0 is the equation of the scanning plane of the lidar in the camera coordinate system.

[0069] Step 202: At any given time, acquire camera images using the camera and acquire radar data using the LiDAR.

[0070] In practical applications, image I is acquired from the camera at time t. t Data acquired from LiDAR in, for The distance in the angular direction, i belongs to n, where n is the angular resolution of the lidar.

[0071] Step 203: Process the camera image using a 3D human pose estimation algorithm to determine human pose data.

[0072] In practical applications, the SimpleBaseline3D algorithm or other 3D human pose estimation algorithms are used to process image I. t Obtain human posture data H t The i-th data contains the 3D coordinates of 17 key points. And the jth key point Contains three coordinates, H t The number of elements in the middle is N t That is, 0 < i ≤ N t .

[0073] Step 204: If the human posture data is empty, return to step 202 to reacquire camera images and radar data.

[0074] Step 205: If the human posture data is a non-empty set, the human posture data includes the three-dimensional coordinates of multiple key points, and the key point first calculation result of each key point is determined according to the equation and the three-dimensional coordinates.

[0075] The key point first calculation result is:

[0076] Wherein, The x-axis coordinate of the key point is x, The y-axis coordinate of the key point is y, The z-axis coordinate of the key point is z, i is the serial number of the human posture data, and j is the serial number of the key point in each human posture data.

[0077] In actual application, for all elements in H t Each key point in The key point first calculation result is calculated:

[0078] Step 206: If any of the key point first calculation results is less than 0, the radar data within the 120-degree range close to the ground in the radar data is determined according to the scanning start angle and the installation angle of the laser radar, and the radar data under the normal non-foreign object condition is recorded.

[0079] In actual application, the data acquired by the laser radar at time t According to the scanning start angle and the installation angle of the laser radar, the data within the 120-degree range close to the ground in r t The data under the normal non-foreign object condition is recorded in advance:

[0080] ​​Step 207: sequentially traversing the radar data in the 120-degree range close to the ground, if there is radar data less than 0.5 times the radar data under the normal no-foreign object condition in the 120-degree range close to the ground, it is determined that the personnel have the risk of mis-touching the live equipment, a warning information is issued, and the step 202 is returned; if there is no radar data less than 0.5 times the radar data under the normal no-foreign object condition in the 120-degree range close to the ground, the step 202 is returned.

[0081] In practical application, sequentially traversing the radar data in the 120-degree range close to the ground, if there is radar data less than 0.5 times the radar data under the normal no-foreign object condition in the 120-degree range close to the ground, it is determined that the personnel have the risk of mis-touching the live equipment, a warning information is issued, and the step 202 is returned; if there is no radar data less than 0.5 times the radar data under the normal no-foreign object condition in the 120-degree range close to the ground, the step 202 is returned. t ', if the data less than 0.5 times the corresponding normal data is found, that is: then it is judged that the personnel have the risk of mis-touching the live equipment, and the step 202 is returned; if not, the step 202 is returned.

[0082] Step 208: if all the first calculation results of the key points are greater than 0, the second calculation result of each key point is calculated.

[0083] The second calculation result of the key point is:

[0084] Wherein, d i,j is the distance of the key point to the scanning plane.

[0085] In practical application, for all elements t in H , the second calculation result of the key point is calculated:

[0086] If d i,j is less than or equal to the set threshold value δ, a warning information is issued, and the step 202 is executed; if d i,j is greater than the set threshold value δ for all i, j, the step 202 is executed.

[0087] Step 209: if the second calculation result of the key point is less than or equal to the set second calculation result threshold of the key point, a warning information is issued; if the second calculation result of the key point is greater than the set second calculation result threshold of the key point, the step 202 is returned.

[0088] Figure 4 The simplified flowchart of the live equipment anti-mis-touching method provided by the application is as shown in Figure 4 .

[0089] 1. Estimate the equation of the scanning plane of the laser radar in the camera coordinate system.

[0090] 2. Obtain the image from the camera and obtain the data from the laser radar. ​

[0091] 3. Processing the image using a three-dimensional human pose estimation algorithm to obtain pose data.

[0092] 4. Determining whether the key points are out of bounds, if so, executing step 6, if not, executing step 5.

[0093] 5. Determining whether the key points are in the warning range, if so, issuing a warning, if not, executing step 2.

[0094] 6. Determining whether the laser radar data is abnormal, if so, issuing an alarm, if not, executing step 2.

[0095] Figure 5 The structure diagram of the charged equipment anti-mis-touch system provided by the application is shown in Figure 5 A charged equipment anti-mis-touch system, comprising:

[0096] An equation estimation module 501 is configured to estimate the equation of the scanning plane of the laser radar in the camera coordinate system.

[0097] The equation estimation module 501 specifically comprises: a standing position image acquisition unit configured to make a person stand at the scanning plane of the laser radar, with the body parallel to the scanning plane, and move left and right to change positions, and for each standing position of the person, use a camera to take a standing position image; a standing position human pose data determination unit configured to process the standing position image using a three-dimensional human pose estimation algorithm to obtain standing position human pose data; a plurality of standing position human pose data determination units configured to repeat the steps of "making a person stand at the scanning plane of the laser radar, with the body parallel to the scanning plane, and move left and right to change positions, and for each standing position of the person, use a camera to take a standing position image" and "processing the standing position image using a three-dimensional human pose estimation algorithm to obtain standing position human pose data" multiple times to obtain a plurality of standing position human pose data; a linear equation set construction unit configured to construct a linear equation set according to the plurality of standing position human pose data; and an equation estimation unit configured to solve the linear equation set using a singular value decomposition method to determine the equation of the scanning plane of the laser radar in the camera coordinate system.

[0098] A camera image and radar data acquisition module 502 is configured to acquire a camera image through a camera and radar data through a laser radar at any time.

[0099] A human pose data determination module 503 is configured to process the camera image using a three-dimensional human pose estimation algorithm to determine human pose data.

[0100] The step returns to the step of "acquiring camera images through a camera and radar data through a laser radar at any time" to reacquire the camera images and the radar data if the human body posture data is an empty set.

[0101] The key point first calculation result determination module 505 is configured to determine a key point first calculation result of each key point according to the equation and the three-dimensional coordinates of the key points if the human body posture data is a non-empty set and the human body posture data includes the three-dimensional coordinates of the key points.

[0102] The key point first calculation result is:

[0103] wherein, is an x-axis coordinate of the key point, is a y-axis coordinate of the key point, is a z-axis coordinate of the key point, i is a serial number of the human body posture data, j is a serial number of the key point in each human body posture data, a is a first coefficient, b is a second coefficient, c is a third coefficient, and d is a fourth coefficient.

[0104] The radar data determination module 506 is configured to determine radar data in a 120-degree range close to the ground in the radar data according to a scanning start angle and a mounting angle of the laser radar and record the radar data under a normal non-foreign matter condition if any key point first calculation result is less than 0.

[0105] The first warning module 507 is configured to sequentially traverse the radar data in the 120-degree range close to the ground, determine that a person has a risk of mistakenly touching a live device and issue a warning information if there is radar data less than 0.5 times the radar data under the normal non-foreign matter condition in the radar data in the 120-degree range close to the ground, and return to the step of "acquiring camera images through a camera and radar data through a laser radar at any time"; and return to the step of "acquiring camera images through a camera and radar data through a laser radar at any time" if there is no radar data less than 0.5 times the radar data under the normal non-foreign matter condition in the radar data in the 120-degree range close to the ground.

[0106] The key point second calculation result calculation module 508 is configured to calculate a key point second calculation result of each key point if all the key point first calculation results are greater than 0.

[0107] The key point second calculation result is: wherein, d i,j is a distance from the key point to the scanning plane.

[0108] The second early warning module 509 is configured to issue a warning information if the second key point calculation result is less than or equal to a set second key point calculation result threshold value; and return to the step of "acquiring a camera image through a camera and acquiring radar data through a laser radar at any time" if the second key point calculation result is greater than the set second key point calculation result threshold value.

[0109] The various embodiments are described in a progressive manner in the specification, and each embodiment focuses on the difference from other embodiments. The same or similar parts between the various embodiments can be referred to each other. For the system disclosed by the embodiments, the description is relatively simple because it corresponds to the method disclosed by the embodiments. The relevant parts can be referred to the description of the method.

[0110] The principles and implementation manners of the present application are described by using specific examples in the specification. The above description of the embodiments is only used to help understand the method of the present application and its core idea. Meanwhile, for the general technical personnel in the field, the specific implementation manners and application ranges can be changed according to the idea of the present application. In conclusion, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method of preventing a mistaken touch of a live device, characterized by, The method comprises the following steps: estimating the equation of the scanning plane of the laser radar in the camera coordinate system; acquiring camera images through the camera and radar data through the laser radar at any time; processing the camera images by using a three-dimensional human pose estimation algorithm to determine human pose data; if the human pose data is an empty set, returning to the step of acquiring camera images and radar data through the camera and the laser radar at any time to reacquire camera images and radar data; if the human pose data is a non-empty set, the human pose data comprising three-dimensional coordinates of a plurality of key points, determining a key point first calculation result of each key point according to the equation and the three-dimensional coordinates; the key point first calculation result is used to determine whether the human key point has crossed the scanning plane of the laser radar; if any key point first calculation result is less than 0, determining radar data in a 120-degree range close to the ground in the radar data according to the scanning start angle and the installation angle of the laser radar and recording the radar data under the normal non-foreign matter condition; sequentially traversing the radar data in the 120-degree range close to the ground, if there is radar data less than 0.5 times the radar data under the normal non-foreign matter condition in the radar data in the 120-degree range close to the ground, determining that the personnel have the risk of miscontacting live equipment, issuing a warning information, and returning to the step of acquiring camera images and radar data through the camera and the laser radar at any time; if there is no radar data less than 0.5 times the radar data under the normal non-foreign matter condition in the radar data in the 120-degree range close to the ground, returning to the step of acquiring camera images and radar data through the camera and the laser radar at any time; if all the key point first calculation results are greater than 0, calculating a key point second calculation result of each key point; if the key point second calculation result is less than or equal to a set key point second calculation result threshold, issuing a warning information; if the key point second calculation result is greater than the set key point second calculation result threshold, returning to the step of acquiring camera images and radar data through the camera and the laser radar at any time.

2. The method for preventing a mistake touch on the live device according to claim 1, wherein The equation of the scanning plane of the laser radar in the camera coordinate system comprises the following steps: the personnel stand at the scanning plane of the laser radar, the body is parallel to the scanning plane, and the position is changed by moving left and right; for each standing position of the personnel, a standing position image is captured by using the camera; the standing position image is processed by using a three-dimensional human pose estimation algorithm to obtain standing position human pose data; the steps of "the personnel stand at the scanning plane of the laser radar, the body is parallel to the scanning plane, and the position is changed by moving left and right; for each standing position of the personnel, a standing position image is captured by using the camera" and "the standing position image is processed by using a three-dimensional human pose estimation algorithm to obtain standing position human pose data" are repeated multiple times to obtain a plurality of standing position human pose data; a linear equation set is constructed according to the plurality of standing position human pose data; Solve the linear equations by singular value decomposition method to determine the equation of the scanning plane of the laser radar in the camera coordinate system.

3. The method for preventing a mistake touch of the live device according to claim 1, wherein The first calculation result of the key points is: wherein, is an x-axis coordinate of the key point, is a y-axis coordinate of the key point, is a z-axis coordinate of the key point, i is a serial number of human pose data, j is a serial number of key points in each human pose data, a is a first coefficient, b is a second coefficient, c is a third coefficient, and d is a fourth coefficient.

4. The method for preventing a mistake touch of the live device according to claim 3, wherein The second calculation result of the key points is: where d i,j is the distance of the key point to the scan plane.

5. A live equipment mistaken touch prevention system characterized by, The method comprises the steps of: An equation estimation module is configured to estimate the equation of the scanning plane of the laser radar in the camera coordinate system; A camera image and radar data acquisition module is configured to acquire camera images by a camera and radar data by a laser radar at any time; A human posture data determination module is configured to process the camera images by a three-dimensional human posture estimation algorithm to determine human posture data; A step return module is configured to return to the step of acquiring camera images and radar data by a camera and a laser radar at any time if the human posture data is an empty set; A first calculation result determination module is configured to determine a first calculation result of each key point according to the equation and the three-dimensional coordinates of the key points if the human posture data is a non-empty set and comprises three-dimensional coordinates of multiple key points; the first calculation result of the key points is used to determine whether a human key point crosses the scanning plane of the laser radar; A radar data determination module is configured to determine radar data close to the ground within a 120-degree range in the radar data according to the scanning start angle and the installation angle of the laser radar if any first calculation result of the key points is less than 0, and record the radar data under normal conditions without foreign matter; A first warning module is configured to traverse the radar data close to the ground within the 120-degree range in sequence, determine that a person has a risk of miscontacting live equipment if there is radar data less than 0.5 times the radar data under normal conditions without foreign matter in the radar data close to the ground within the 120-degree range, issue a warning message, and return to the step of acquiring camera images by a camera and a laser radar at any time; if there is no radar data less than 0.5 times the radar data under normal conditions without foreign matter in the radar data close to the ground within the 120-degree range, return to the step of acquiring camera images by a camera and a laser radar at any time; A second calculation result calculation module is configured to calculate a second calculation result of each key point if all the first calculation results of the key points are greater than 0; A second warning module is configured to issue a warning message if the second calculation result of the key points is less than or equal to a set second calculation result threshold value of the key points; If the second calculation result of the key points is greater than the set second calculation result threshold value of the key points, return to the step of acquiring camera images by a camera and a laser radar at any time.

6. The miscontact prevention system for live equipment according to claim 5, wherein The equation estimation module specifically comprises: A standing position image acquisition unit is configured to make a person stand at the scanning plane of the laser radar, with the body parallel to the scanning plane, and move left and right to change positions; for each standing position of the person, an image of the standing position is captured by a camera. A standing position human body posture data determination unit is configured to process the standing position image by using a three-dimensional human body posture estimation algorithm to obtain standing position human body posture data. A plurality of standing position human body posture data determination units are configured to repeat the steps of "a person stands at the scanning plane of the laser radar, the body is parallel to the scanning plane, and the left and right movements change positions, for each standing position of the person, a standing position image is captured by using a camera" and "a standing position human body posture data is obtained by processing the standing position image by using a three-dimensional human body posture estimation algorithm" multiple times to obtain a plurality of standing position human body posture data. A linear equation set construction unit is configured to construct a linear equation set according to the plurality of standing position human body posture data. An equation estimation unit is configured to solve the linear equation set by using a singular value decomposition method to determine the equation of the scanning plane of the laser radar in the camera coordinate system.

7. The miscontact prevention system for live equipment according to claim 5, wherein The first calculation result of the key points is: wherein, is an x-axis coordinate of the key point, is a y-axis coordinate of the key point, is a z-axis coordinate of the key point, i is a serial number of human pose data, j is a serial number of key points in each human pose data, a is a first coefficient, b is a second coefficient, c is a third coefficient, and d is a fourth coefficient.

8. The miscontact prevention system for live equipment according to claim 7, characterized by The second calculation result of the key points is: where d i,j is the distance of the key point to the scan plane.

Images