A Coal Mine Dynamic Disaster Perception and Alarm Method Based on Visible Light Binocular Vision

Abstract

This invention discloses a coal mine dynamic disaster perception and alarm method based on visible light binocular vision. The main steps include: installing a visible light binocular camera at a monitoring point in the coal mine, using a device in the coal mine with a distinct color difference from the coal and rock ejected by the disaster as the background device; when the monitoring identifies a significant change in the color of the color image, an average brightness of the color image less than a set brightness threshold, a significant change in the depth image, or a moving speed of the object causing the significant changes in color and depth images greater than a set speed threshold, then acquiring methane concentration data in the vicinity of the visible light binocular camera; when the methane concentration rapidly increases or reaches an alarm value, a coal and gas outburst alarm signal is issued; conversely, a rockburst alarm signal is issued.

Description

TECHNICAL FIELD

[0001] The present application relates to a kind of coal mine power disaster perception alarm method based on visible light binocular vision, especially based on image-based target detection technology, visual speed measurement technology, coal mine rock burst identification technology, and coal and gas outburst identification technology. BACKGROUND

[0002] Coal is the main energy of our country, and plays an important role as energy ballast. Coal industry belongs to high-risk industry, and accidents such as rock burst, coal and gas outburst, fire, flood, gas and coal dust explosion seriously threaten the safety production of coal mine. Therefore, in the process of coal production, rock burst, coal and gas outburst, fire, flood, gas and coal dust explosion and other accidents are identified in time, so as to start disaster emergency plan and emergency rescue, which plays an important role in the safety production of coal mine.

[0003] Rock burst and coal and gas outburst are typical coal mine dynamic disasters. When disaster occurs, a large amount of coal and rock is thrown to the mining working face and roadway space, causing the objects in the mining working face and roadway space to be blocked, covered or buried. Once it happens, it may cause strong damage, casualties and property loss. At present, thermal infrared and color images can be used to identify coal mine dynamic disasters. However, the thermal infrared image formed by infrared thermal imager has low information richness except temperature information, and the current price is expensive and the image resolution is relatively low. Color image has rich information, low price and high resolution. Visible light binocular camera can not only get color image but also get depth image. In addition to all the characteristics of color image, it can also identify the depth characteristics and diffusion speed of abnormal area in the monitoring area. The information is more abundant, the price is relatively low, and the resolution is high. In addition, the cantilever type heading machine, continuous miner, hydraulic anchor drill vehicle, crawler type transfer and crusher, anchor transfer machine group, heading and anchoring machine, shuttle car, belt conveyor and air duct in the underground mining working face of coal mine, as well as the coal mining machine, hydraulic support, scraper conveyor, transfer machine, switch machine, crusher in the coal mining working face, and the mobile substation in the flat roadway, all have distinct color difference with the disaster thrown coal and rock, which can be used as the background to identify the color, brightness, depth and speed characteristics of the disaster thrown coal and rock.

[0004] In view of the problems existing in the present rock burst and coal and gas outburst disaster monitoring and the environmental characteristics of underground coal mine, the present application realizes a coal mine dynamic disaster perception alarm method based on visible light binocular vision. In the process of using visible light binocular image to perceive coal mine dynamic disaster, the color, brightness, depth and speed characteristics of a large amount of coal and rock thrown to the roadway space are perceived by monitoring the color image and depth image in the monitoring area, which provides simple and distinct information of a large amount of thrown coal and rock, has the advantages of low cost, non-contact, wide monitoring range, rich information, fast processing speed and stable reliability. SUMMARY

[0005] The technical problem to be solved by the present application is to use a visible light binocular camera to monitor the color change, brightness feature, depth feature and speed feature of the monitoring area based on the equipment with distinct color difference between the coal mine underground and the disaster thrown coal rock, so as to realize the perception and alarm of rock burst and coal and gas outburst disasters based on visible light binocular vision.

[0006] The present application specifically adopts the following technical solutions to solve the above technical problems:

[0007] A coal mine dynamic disaster perception and alarm method based on visible light binocular vision, characterized in that the coal mine dynamic disaster includes rock burst, coal and gas outburst, and the perception and alarm method comprises the following steps:

[0008] Step 1: install a visible light binocular camera and a methane sensor at the monitoring point position in the coal mine underground;

[0009] Step 2: collect the visible light binocular video image in the monitoring area, and use the equipment with distinct color difference between the coal mine underground and the disaster thrown coal rock as the background equipment to monitor and identify whether the color of the color video image changes greatly;

[0010] Step 3: cycle step 2, and when the color of the image changes greatly, identify the average brightness of the color video image;

[0011] Step 4: when the average brightness of the color video image is less than the set brightness threshold, use the visible light binocular camera to monitor whether the depth image of the monitoring area with the background equipment as the background changes greatly;

[0012] Step 5: when the depth image changes greatly, monitor the moving speed of the object causing the color of the color image and the depth image to change greatly;

[0013] Step 6: when the visible light binocular camera monitors the moving speed of the object causing the color of the color image and the depth image to change greatly in the monitoring area to be greater than the set speed threshold, obtain the methane concentration data of the monitoring area of the visible light binocular camera

[0014] Step 7: when the methane concentration of the monitoring area of the visible light binocular camera is normal, send a rock burst alarm signal; when the methane concentration of the monitoring area of the visible light binocular camera rapidly increases or reaches the alarm value, send a coal and gas outburst alarm signal.

[0015] Further, the monitoring point position where the visible light binocular camera is installed includes the hydraulic support on the coal mining face, the coal mining machine, the air inlet roadway of the coal mining face, the air return roadway of the coal mining face, the heading roadway of the heading face, the heading machine, and one side or both sides of the heading roadway branch.

[0016] Further, the position of installing the visible light binocular camera is located at the top of the roadway, or close to the top of the roadway, or the height is greater than 2 meters; the focal length and exposure value of the visible light binocular camera are manually set, and the automatic focusing and automatic white balance functions of the visible light binocular camera are closed.

[0017] Further, the position of installing the visible light binocular camera includes the roof of the tunneling roadway, the two sides of the tunneling roadway close to the roof, the top of the hydraulic support of the coal mining face, the side of the hydraulic column of the hydraulic support of the coal mining face close to the roof, the roof of the roadway of the crossheading, and the two sides of the roadway of the crossheading close to the roof.

[0018] Further, the background equipment includes: a boom-type tunneling machine, a continuous miner, a hydraulic anchor rod drilling vehicle, a crawler-type transfer and crushing machine, an anchor rod transfer machine group, a tunneling and anchoring machine, a shuttle car, a belt conveyor, and a wind pipe in the tunneling face; a coal mining machine, a hydraulic support, a scraper conveyor, a transfer machine, a switch machine, and a crusher in the coal mining face; and a mobile substation in the crossheading.

[0019] Further, the black area of the color image is segmented based on the HSV color model; whether the color of the visible light binocular video image changes greatly is monitored and identified by monitoring whether the black area of the color image increases; whether the average brightness of the color image is less than the set brightness threshold is monitored and identified by monitoring the size of the average brightness in the HSV color model of the color image in the monitoring area; whether the area of the changed region of the depth image increases greatly is monitored and identified by using the ViBe algorithm; if the area increases greatly, it means that the depth image changes greatly; the moving speed of the center point of the contour of the changed region of the depth video image in the monitoring area is monitored and identified as the moving speed of the object causing the color of the color image and the depth image to change greatly; the speed threshold is set to 13 m / s.

[0020] Further, in order to increase the monitoring range of a single visible light binocular camera as much as possible while not affecting the speed measurement effect of the single visible light binocular camera, the best inclination angle of the visible light binocular camera is the maximum value that satisfies the following conditions:

[0021] (1) the best inclination angle of the visible light binocular camera in the tunneling face is the maximum inclination angle that satisfies the following conditions:

[0022] ① the monitoring area of the visible light binocular camera only covers the bottom of the roadway and the two sides of the roadway;

[0023] ② the actual moving speed of the farthest object in the monitoring area of the visible light binocular camera is equal to 1 / n times (n is a positive integer) of the set speed threshold when the object moves one pixel length in one frame of time in the depth image.

[0024] (2) The optimal inclination angle of the visible light binocular camera at the mining working face is the maximum inclination angle that simultaneously satisfies the following conditions:

[0025] ① The monitoring area of the visible light binocular camera only covers the bottom of the mining working face and the area of the hydraulic support;

[0026] ② The actual moving speed of the farthest object in the monitoring area of the visible light binocular camera is equal to 1 / n times of the set speed threshold value (n is a positive integer) when the object moves one pixel length in the depth image in one frame of time.

[0027] Further, the optimal inclination angle of the visible light binocular camera is the maximum value of the following constraint conditions:

[0028] (1) The optimal inclination angle of the visible light binocular camera at the mining working face is the maximum inclination angle that simultaneously satisfies the following conditions:

[0029]

[0030] wherein, α is the forward inclination angle of the visible light binocular camera relative to the vertical downward direction; 2θ L is the field of view angle of the visible light binocular camera along the axis direction of the roadway; H is the height of the installation position of the visible light binocular camera from the roadway floor; f is the focal length of the monocular camera in the visible light binocular camera; v is the actual moving speed of the farthest object in the monitoring area when the object moves one pixel length Δs1 in the depth image in one frame of time Δt1; v T is the set speed threshold value;

[0031] (2) The optimal inclination angle of the visible light binocular camera at the mining working face is the maximum inclination angle that simultaneously satisfies the following conditions:

[0032]

[0033] wherein, α is the forward inclination angle of the visible light binocular camera relative to the vertical downward direction; 2θ L is the field of view angle of the visible light binocular camera along the axis direction of the roadway; H is the height of the installation position of the visible light binocular camera from the roadway floor; Δw is the distance of the installation position of the visible light binocular camera from the mining working face in the vertical direction of the axis of the roadway; f is the focal length of the monocular camera in the visible light binocular camera; v is the actual moving speed of the farthest object in the monitoring area when the object moves one pixel length Δs1 in the depth image in one frame of time Δt1; v T is the set speed threshold value.

[0034] Further, when the visible light binocular camera has an angle relative to the vertical downward direction, different monitoring directions monitor different lengths of the same object in the roadway, and according to the speed ratio equal to the distance ratio in the same time, then:

[0035]

[0036] In the formula, v1 is the real moving speed of the object; v2 is the moving speed of the object identified by the visible light binocular camera; a is the front inclination angle of the visible light binocular camera relative to the vertical downward direction; and θ2 is the included angle formed by the projection point of the initial moving position of the object found in the image on the projection line of the optical axis of the visible light binocular camera on the roadway floor and the line connecting the visible light binocular camera position, and the projection line of the optical axis of the visible light binocular camera on the roadway floor.

[0037] Further, according to the mathematical relationship between the depth image collected by the visible light binocular camera and the object moving in the roadway space causing great changes in the color and depth of the image, the actual moving speed of the object is:

[0038]

[0039] Wherein, v is the moving speed of the object from the moving starting point B t to the moving end point C t after a period of time Δt; Z B is the depth of the moving starting point B t of the object monitored by the visible light binocular camera from the position O of the visible light binocular camera; Z C is the depth of the moving end point C t of the object monitored by the visible light binocular camera from the position O of the visible light binocular camera; x B ,y B are respectively the horizontal and vertical coordinates of the moving starting point of the object in the pixel coordinate system; x C ,y C are respectively the horizontal and vertical coordinates of the moving end point of the object in the pixel coordinate system; f is the focal length of the monocular visible light camera in the visible light binocular camera; L c ,W c are respectively the length and width of the depth image formed by the binocular camera; and a is the front inclination angle of the visible light binocular camera relative to the vertical downward direction.

[0040] Further, in order to be able to identify a higher coal and rock throwing speed, it is necessary to comprehensively analyze the field angle of view of the visible light binocular camera, the front inclination angle, the collection frame rate, and the installation position of the visible light binocular camera from the height of the roadway floor, and the maximum speed that can be identified by the visible light binocular camera when the visible light binocular camera has an angle is estimated as follows:

[0041]

[0042] In the formula, V max is the maximum speed that can be recognized by the visible light binocular camera; H is the height of the installation position of the visible light binocular camera from the roadway floor; Δt1 is the time passing through a frame; α is the forward inclination angle of the visible light binocular camera relative to the vertical downward direction; 2θ L is the field of view angle of the visible light binocular camera along the optical axis direction of the camera. BRIEF DESCRIPTION OF DRAWINGS

[0043] Figure 1 The flow chart of the coal mine dynamic disaster perception alarm method based on visible light binocular vision of the application;

[0044] Figure 2 The schematic diagram of the placement position of the visible light binocular camera at the heading face of the application;

[0045] Figure 3 The schematic diagram of the placement position of the visible light binocular camera at the mining face of the application;

[0046] Figure 4 The schematic diagram of the color change feature method of the monitoring image of the visible light binocular camera of the application;

[0047] Figure 5 The schematic diagram of the method for identifying the speed of the disaster thrown coal and rock by the visible light binocular camera of the application. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical scheme and advantages of the application more clear, the application is described in detail and completely below in combination with the drawings and specific implementation methods, and the examples should not be regarded as limiting the use range of the application.

[0049] As Figure 1 shown in the flow chart of the coal mine dynamic disaster perception alarm method based on visible light binocular vision, the flow chart comprises:

[0050] 1. Initialization (101): visible light binocular cameras are installed in key monitoring areas in the mine, and the key monitoring areas include the top of the hydraulic support at the coal mining face, the top of the coal mining machine, the air inlet roadway at the coal mining face, the air return roadway at the coal mining face, the heading roadway at the heading face, the heading machine, and one side or both sides of the branch of the heading roadway.

[0051] 2. Monitoring area setting (102): the monitoring area of the visible light binocular camera is set by taking the equipment with a distinct color difference from the disaster thrown coal and rock in the coal mine as the main background.

[0052] 3. Monitor and identify color change of color image in monitoring area (103): collect color video image in monitoring area, take the equipment with distinct color difference from disaster thrown coal and rock in underground coal mine as background equipment, monitor and identify whether color video image appears large color change of image.

[0053] 4. Large color change of image? (104): when monitoring color image in monitoring area appears large color change of image, it means that image color abnormality may appear when rock burst, coal and gas outburst disaster occurs, then execute (105), otherwise return to execute (103).

[0054] 5. Monitor and identify average brightness of color image (105): when color image appears large color change of image, then identify average brightness of color video image, and count average brightness value of color image.

[0055] 6. Average brightness less than set threshold? (106): when average brightness of color video image is less than set brightness threshold, it means that average brightness of color image area with large color change of image is less than average brightness of fire, gas and coal dust explosion, and no brightness abnormality appears, which excludes the interference of fire, gas and coal dust explosion on rock burst, coal and gas outburst disaster perception, then execute (107), otherwise return to execute (103).

[0056] 7. Monitor and identify depth image change (107): when average brightness of color image is less than set brightness threshold, then monitor and identify depth image change.

[0057] 8. Large depth image change? (108): when depth image appears large change, it means that depth abnormality appears in monitoring area, then execute (109), otherwise return to execute (103).

[0058] 9. Monitor and identify moving speed of object causing large color change of color image and depth image (109): when depth image appears large change, then monitor and identify moving speed of object causing large color change of color image and depth image in monitoring area by using visible light binocular camera.

[0059] 10. Moving speed greater than set threshold (110): when moving speed of object causing large color change of color image and depth image in monitoring area is greater than set speed threshold by using visible light binocular camera, it means that speed abnormality appears when rock burst and coal and gas outburst disaster occurs, then execute (111), otherwise return to execute (103).

[0060] 11. Methane concentration rises rapidly or reaches alarm value? (111): Obtain methane concentration data in the vicinity of the visible light binocular camera to identify whether it is rockburst or coal and gas outburst.

[0061] 12. Coal and gas outburst alarm (112): When the visible light binocular camera detects a significant change in the color of the color image in the monitored area, and the average brightness of the color video image is less than the set brightness threshold, the depth image changes significantly, the visible light binocular camera detects an object in the monitored area whose movement speed causes significant changes in the color and depth images of the color image to exceed the set threshold, and at the same time, the methane concentration in the nearby area is detected to increase rapidly or reach the alarm value, then a coal and gas outburst alarm will be triggered.

[0062] 13. Rockburst alarm (113): When the visible light binocular camera detects a significant change in the color of the color image in the monitored area, and the average brightness of the color video image is less than the set brightness threshold, the depth image changes significantly, the visible light binocular camera detects that the moving speed of the object causing significant changes in the color and depth images in the monitored area is greater than the set threshold, and at the same time the methane concentration in the nearby area is normal, then a rockburst alarm is triggered.

[0063] like Figure 2 The diagram shown illustrates the placement of the visible light binocular camera at the tunneling face, including:

[0064] The visible light binocular camera O is placed at the center axis of the tunnel roof, facing the working face. In order to better monitor disasters, the visible light binocular camera is tilted forward and downward relative to the vertical. The monitoring area of ​​the visible light binocular camera is the tunnel floor monitoring area (areas ABCD) and the two sides of the tunnel.

[0065] like Figure 3 The diagram shown illustrates the placement of the visible light binocular camera at the longwall face, including:

[0066] The visible light binocular camera O is placed on top of the hydraulic support, facing the coal mining face. In order to better monitor disasters, the visible light binocular camera is tilted forward and downward relative to the vertical. The monitoring area of ​​the visible light binocular camera is the bottom of the longwall face (areas ABCD) and the hydraulic support area.

[0067] like Figure 4 The diagram shown illustrates a method for monitoring image color change features using a visible light binocular camera, including:

[0068] When disaster occurs, a large amount of coal rock (black) is thrown out, which will cause the background equipment with a distinct color difference from the coal rock to be covered, covered or buried by the coal rock in the monitoring range of the visible light binocular camera, and then the image color in the monitoring range will be partially or entirely changed to the color of the coal rock (black) except the original black area. By setting the corresponding parameter range of the HSV color model, the black (coal rock color) area in the monitoring range is segmented in a targeted manner. Then, whether the pixel number of the foreground point has a large change is counted by using the ViBe algorithm. If there is a large change, the black area in the monitoring range is changed greatly, that is, the color change feature caused by the coal rock thrown out by the disaster is recognized.

[0069] As shown in the visible light binocular camera recognition method for the speed of the coal rock thrown out by the disaster, the method comprises the steps of: Figure 5

[0070] Since the shape of the coal rock thrown out by the rock burst and the coal and gas outburst disaster is irregular, and the accumulation condition of the coal rock thrown out is unpredictable, the maximum contour center points P and Q of the area with a large change in depth in the front and rear frames in the monitoring area of the visible light binocular camera can be used as the feature points of the depth change feature of the mining working face and the roadway space caused by the coal rock thrown out by the disaster. Then, the moving speed of the maximum contour center points of the area with a large change in depth in the front and rear frames of the depth video image can be considered as the speed of the coal rock thrown out by the disaster.

[0071] In order to realize the monitoring range of a single visible light binocular camera as large as possible while not affecting the speed measurement effect of the single visible light binocular camera, the best inclination angle of the visible light binocular camera is the maximum value that satisfies the following condition:

[0072] (1) The best inclination angle of the visible light binocular camera in the mining working face is the maximum inclination angle that satisfies the following conditions:

[0073] ①The monitoring area of the visible light binocular camera only covers the bottom of the roadway and the two sides of the roadway;

[0074] ②When the farthest object in the monitoring area of the visible light binocular camera moves a pixel length in one frame of time, the actual moving speed of the object corresponding to the object in the depth image is equal to 1 / n times (n is a positive integer) of the set speed threshold.

[0075] That is, the best inclination angle of the visible light binocular camera in the mining working face is the maximum value that satisfies the following constraint condition:

[0076]

[0077] Wherein, α is the forward inclination angle of the visible light binocular camera relative to the vertical downward direction; 2θ L ​is the field of view angle of the visible light binocular camera along the axis direction of the roadway; H is the height of the installation position of the visible light binocular camera from the floor of the roadway; f is the focal length of the monocular camera in the visible light binocular camera; v is the actual moving speed of the object at the farthest end of the monitoring area when the object moves a pixel length Δs1 in the depth image in a frame of time Δt1; v T is a set speed threshold.

[0078] (2) The optimal inclination angle of the visible light binocular camera at the coal mining face is the maximum inclination angle that simultaneously satisfies the following conditions:

[0079] ① The monitoring area of the visible light binocular camera only covers the bottom of the coal mining face and the hydraulic support area;

[0080] ② The actual moving speed of the object at the farthest end of the monitoring area of the visible light binocular camera when the object moves a pixel length in the depth image in a frame of time is equal to 1 / n times the set speed threshold (n is a positive integer).

[0081] That is, the optimal inclination angle of the visible light binocular camera at the coal mining face is the maximum value of the following constraint conditions:

[0082]

[0083] wherein, α is the forward inclination angle of the visible light binocular camera relative to the vertical downward direction; 2θ L is the field of view angle of the visible light binocular camera along the axis direction of the roadway; H is the height of the installation position of the visible light binocular camera from the floor of the roadway; Δw is the distance of the installation position of the visible light binocular camera from the coal mining face in the vertical direction of the axis of the roadway; f is the focal length of the monocular camera in the visible light binocular camera; v is the actual moving speed of the object at the farthest end of the monitoring area when the object moves a pixel length Δs1 in the depth image in a frame of time Δt1; v T is a set speed threshold.

[0084] When the visible light binocular camera has an inclination angle relative to the vertical downward direction, the length of the same object in the roadway monitored in different monitoring directions is different, and according to the speed ratio equal to the distance ratio in the same time, then:

[0085]

[0086] In the formula, v1 is the real moving speed of the object; v2 is the moving speed of the object recognized by the visible light binocular camera; α is the forward inclination angle of the visible light binocular camera relative to the vertical downward direction; θ2 is the included angle formed by the projection point of the initial moving position of the object found in the image on the projection line of the floor of the roadway on the optical axis of the visible light binocular camera and the line connecting the visible light binocular camera position and the projection line of the floor of the roadway on the optical axis of the visible light binocular camera.

[0087] Then, according to the mathematical relationship between the depth image collected by the visible light binocular camera and the object moving with the color and depth image of the tunnel space changing greatly, the actual moving speed of the object is:

[0088]

[0089] Wherein, v is the moving speed of the object from the moving starting point B t to the end point C t after a period of time Δt; Z B is the depth of the moving starting point B t of the object from the position of the visible light binocular camera O; Z C is the depth of the moving end point C t of the object from the position of the visible light binocular camera O; x B ,y B are respectively the horizontal and vertical coordinates of the moving starting point of the object in the pixel coordinate system; x C ,y C are respectively the horizontal and vertical coordinates of the moving end point of the object in the pixel coordinate system; f is the focal length of the monocular visible light camera in the visible light binocular camera; L c ,W t are respectively the length and width of the depth image pixel size formed by the binocular camera; and α is the front inclination angle of the visible light binocular camera relative to the vertical downward.

[0090] In order to identify a higher coal and rock throwing speed, it is necessary to combine the field angle of the visible light binocular camera, the front inclination angle, the frame rate and the installation position of the visible light binocular camera from the height of the tunnel floor for comprehensive analysis. The maximum speed estimation method that can be identified by the visible light binocular camera with inclination angle is as follows:

[0091]

[0092] In the formula, V max is the maximum speed that can be identified by the visible light binocular camera; H is the height of the installation position of the visible light binocular camera from the tunnel floor; Δt1 is the time after a frame; α is the front inclination angle of the visible light binocular camera relative to the vertical downward; and 2θ L is the field angle of the visible light binocular camera along the camera optical axis direction.

Claims

1. A coal mine power disaster perception alarm method based on visible light binocular vision, characterized in that: The coal mine dynamic disaster includes rock burst and coal and gas outburst, and the sensing and alarming method comprises the following steps: Step 1: installing a visible light binocular camera and a methane sensor at a monitoring point position in a coal mine underground; Step 2: collecting visible light binocular video images in a monitored area, taking equipment with a distinct color difference from disaster thrown coal and rock in the coal mine underground as background equipment, and monitoring and identifying whether a color image appears a great color change; Step 3: repeating step 2, when the color image appears a great color change, color video image average brightness identification is performed; Step 4: when the color video image average brightness is less than a set brightness threshold value, whether a depth image of the monitored area with the background equipment as the background monitored by the visible light binocular camera appears a great change is monitored; Step 5: when the depth image appears a great change, the moving speed of an object causing the color image color and the depth image to appear a great change is monitored; Step 6: when the visible light binocular camera monitors that the moving speed of the object causing the color image color and the depth image to appear a great change in the monitored area is greater than a set speed threshold value, methane concentration data of the visible light binocular camera monitored area is acquired; Step 7: when the methane concentration of the visible light binocular camera monitored area is normal, a rock burst alarm signal is sent; when the methane concentration of the visible light binocular camera monitored area rapidly increases or reaches an alarm value, a coal and gas outburst alarm signal is sent; The HSV color model is used to segment a black color region of the color image, whether the color image appears a great color change is monitored and identified by monitoring whether the black color region of the color image increases, whether the color image average brightness is less than the set brightness threshold value is monitored and identified by monitoring the size of the average brightness of the HSV color model of the color image in the monitored area, whether the depth image change area appears a great increase is monitored and identified by using the ViBe algorithm, if the depth image change area appears a great increase, it is indicated that the depth image appears a great change, the moving speed of the depth image change area contour center point in the front and back frames of the depth video image in the monitored area is monitored as the moving speed of the object causing the color image color and the depth image to appear a great change, and the speed threshold value is set to 13 m / s.

2. The visible light-based binocular vision-based coal mine power disaster perception and alarm method according to claim 1, characterized in that: The monitoring point position where the visible light binocular camera is installed comprises a coal mining face hydraulic support, a coal mining machine, an air inlet roadway of the coal mining face, an air return roadway of the coal mining face, a heading roadway of the heading face, a heading machine, and one side or both sides of a heading roadway branch.

3. The visible light-based binocular vision-based coal mine power disaster perception and alarm method according to claim 1, characterized in that: The position where the visible light binocular camera is installed is located at the roadway top, close to the roadway top, or higher than 2 meters, the focal length and the exposure value of the visible light binocular camera are manually set, and the automatic focusing and the automatic white balance functions of the visible light binocular camera are closed.

4. The visible light-based binocular vision-based coal mine power disaster perception and alarm method according to claim 1, characterized in that: The position where the visible light binocular camera is installed comprises the heading roadway roof, the heading roadway two sides close to the roof, the coal mining face hydraulic support top, the coal mining face hydraulic support hydraulic column side close to the roof, the roadway roof of the flat roadway, and the roadway two sides of the flat roadway close to the roof.

5. The visible light-based binocular vision based coal mine power disaster perception and alarm method according to claim 1, characterized in that: The background device comprises: a boom-type heading machine, a continuous miner, a hydraulic anchor rod drilling vehicle, a crawler-type transfer crusher, an anchor rod transfer machine set, a heading and anchoring machine, a shuttle car, a belt conveyor, and a ventilation shaft in a heading face; a coal mining machine, a hydraulic support, a scraper conveyor, a transfer machine, a switch machine, and a crusher in a coal mining face; and a mobile substation in a crossheading.

6. The visible light-based binocular vision based coal mine power disaster perception and alarm method according to claim 1, characterized in that: In order to realize the maximum monitoring range of a single visible light binocular camera and at the same time not affect the speed measurement effect of the single visible light binocular camera, the optimal inclination angle of the visible light binocular camera is the maximum value satisfying the following conditions: (1) The optimal inclination angle of the visible light binocular camera in the heading face is the maximum inclination angle satisfying the following conditions: The visible light binocular camera only covers the bottom and sides of the tunnel. The visible light binocular camera monitors the farthest object in the region, and the actual moving speed of the corresponding object in the depth image when moving one pixel length in one frame of time is equal to 1 / n times of the set speed threshold, n is a positive integer. (2) The optimal inclination angle of the visible light binocular camera in the coal mining face is the maximum inclination angle satisfying the following conditions: The visible light binocular camera only covers the bottom of the longwall face and the hydraulic support area; The visible light binocular camera monitors the farthest object in the region, and the actual moving speed of the corresponding object in the depth image is equal to 1 / n times of the set speed threshold value when it moves a pixel length in one frame of time, n is a positive integer.

7. The visible light-based binocular vision based coal mine power disaster perception and alarm method according to claim 1, characterized in that: The optimal inclination angle of the visible light binocular camera is the maximum value of the following constraint conditions: (1) The optimal inclination angle of the visible light binocular camera in the heading face is the maximum inclination angle satisfying the following conditions: wherein, is the front tilt angle of the visible light binocular camera relative to the vertical downward direction; 2 is the field of view angle of the visible light binocular camera along the axis direction of the roadway; is the height of the installation position of the visible light binocular camera from the floor of the roadway; is the focal length of the monocular camera in the visible light binocular camera; is the object at the farthest end of the monitoring area, and the corresponding object actual moving speed is moving a pixel length moving a pixel length is the set speed threshold value; (2) The optimal inclination angle of the visible light binocular camera in the coal mining face is the maximum inclination angle satisfying the following conditions: wherein, is the front tilt angle of the visible light binocular camera relative to the vertical downward direction; 2 is the field of view angle of the visible light binocular camera along the axis direction of the roadway; is the height of the installation position of the visible light binocular camera from the floor of the roadway; is the distance of the installation position of the visible light binocular camera from the coal mining face in the vertical direction of the axis of the roadway; is the focal length of the monocular camera in the visible light binocular camera; is the object at the farthest end of the monitoring area, and the corresponding object actual moving speed in the depth image is moved by one pixel length over time is the set speed threshold value.

8. The visible light-based binocular vision method for coal mine dynamic disaster perception and alarm according to claim 1, characterized in that: When the visible light binocular camera has an inclination angle relative to the vertical downward direction, different monitoring directions monitor the same object in the roadway with different lengths, and according to the speed ratio equal to the distance ratio in the same time, then: wherein, is the real moving speed of the object; is the moving speed of the object identified by the visible light binocular camera; is the front-leaning angle of the visible light binocular camera relative to the vertical downward direction; is the included angle formed by the projection point of the initial moving position of the object found in the image on the projection line of the tunnel floor of the optical axis of the visible light binocular camera and the line connecting the visible light binocular camera position and the optical axis of the visible light binocular camera.

9. The visible light-based binocular vision based coal mine power disaster perception and alarm method according to claim 1, characterized in that: According to the mathematical relationship between the depth image collected by the visible light binocular camera and the color and depth image of the object moving in the roadway space, the actual moving speed of the object is: wherein, is the moving speed of the object from the moving start point after a period of time to the moving end point ; is the moving start point of the object monitored by the visible light binocular camera is the depth of the moving start point of the object from the visible light binocular camera ; is the moving end point of the object monitored by the visible light binocular camera is the depth of the moving end point of the object from the visible light binocular camera ; are respectively the horizontal and vertical coordinates of the moving start point of the object in the pixel coordinate system; are respectively the horizontal and vertical coordinates of the moving end point of the object in the pixel coordinate system; is the focal length of the monocular visible light camera in the visible light binocular camera; are respectively the length and width of the pixel size of the depth image formed by the binocular camera; is the front tilt angle of the visible light binocular camera relative to the vertical downward direction.

10. The visible light-based binocular vision based coal mine power disaster perception and alarm method according to claim 1, characterized in that: In order to be able to identify a higher coal and rock throwing speed, it is necessary to comprehensively analyze the field angle of the visible light binocular camera, the front inclination angle, the frame rate, and the installation position distance of the visible light binocular camera from the height of the roadway floor. The maximum speed estimation method of the visible light binocular camera with an inclination angle is as follows: In the formula, is the maximum speed that the visible light binocular camera can recognize; is the height of the visible light binocular camera installation position from the roadway floor; is the time passed by one frame; is the forward inclination angle of the visible light binocular camera relative to the vertical downward direction; is the field of view angle of the visible light binocular camera in the direction of the camera optical axis.

Images