CN104454011A - Coal face rock burst alarming method based on images - Google Patents

Abstract

The invention discloses an image-based ground pressure alarm method for coal mining face. The alarm method installs a camera near the relevant area of the coal mining face to analyze the change of video image data and the working condition of equipment in real time, and finds abnormalities in the monitoring area. If the combination of abnormal information meets the alarm conditions, a rock burst alarm signal will be issued. This alarm method fully considers the various characteristics of rock burst in the coal mining face, is simple to implement, can accurately determine the rock burst in the coal mining face, effectively improves the rescue efficiency, and is conducive to the management personnel to take corresponding measures in time , to avoid serious accidents such as gas explosions caused by roadway blockage and gas accumulation, and to avoid or reduce casualties.

Description

Alarm method of rock burst in coal mining face based on image

technical field

The invention relates to an image-based rockburst alarm method for a coal mining face, and the method relates to the fields of image processing, communication and the like.

Background technique

Coal is the main energy in my country, accounting for about 70% of the primary energy. The coal industry is a high-risk industry, and accidents such as gas, flood, fire, roof, and coal dust have plagued coal mine safety production. Most of the major accidents in my country's coal mines are severe gas accidents, and the number of casualties caused by gas accidents is also the largest proportion of all coal mine accidents. Therefore, the prevention and control of gas accidents is very important.

Gas accidents include gas explosion, rock burst, gas suffocation, gas burning and other accidents. In order to avoid or reduce the occurrence of rock burst accidents, a variety of coal or rock and gas outburst prevention and control methods have been proposed, which play an important role in the safety production of coal mines. However, the existing real-time monitoring and forecasting methods of rock burst (including microseismic, acoustic emission, electromagnetic radiation, infrared radiation, etc.) have a high rate of false alarms and false negatives, which is difficult to meet the needs of coal mine safety production.

The coal mining working face in the coal mine is prone to rock burst. When the rock burst occurs, it is easy to cause the staff near the working face to be buried or trapped; in addition, the rock burst will cause blockage of the roadway near the working face, causing Poor ventilation in the roadway causes gas to accumulate, which can easily lead to gas explosions. If the rockburst in this area can be quickly and accurately judged at the first time, rescue can be organized as soon as possible to gain valuable rescue time, avoid or reduce casualties caused by landfill or trapping; and organize personnel to deal with blocked roadways in time Measures to effectively avoid accidents such as gas explosions caused by gas accumulation, and to avoid or reduce casualties caused by gas suffocation and gas explosions.

Contents of the invention

When rock burst occurs in the coal mining face, a large amount of coal and rock will rush out at high speed and accumulate in most areas of the working face, and even flow out through the air inlet and return air lanes. It may cause damage to various electronic equipment and communication lines in the area; according to this feature, the present invention proposes an image-based ground pressure alarm method for coal mining face. Through real-time analysis, it is found that the movement and accumulation of a large number of abnormal coal rocks in the monitoring area is combined with the numerical value of the methane sensor to realize accurate rockburst alarm. Specific methods include:

1. Under the coal mine, place cameras on the hydraulic support or shearer of the coal mining face, in the air inlet lane of the coal mining face, and in the return air lane of the coal mining face; detect the collected video image data in real time, and detect the camera and the working condition of the communication line, while monitoring the change of the value of the methane sensor in the nearby area; when it detects abnormal accumulation of objects in the set area in the video image of the camera, or a large number of objects moving at high speed, it is judged as data abnormality; when the data is detected Abnormalities, or data anomalies and related equipment failures occur one after another in a short period of time, and the methane concentration in the nearby area is detected to be normal, then a rock burst alarm and signal will be issued.

2. The installation position of the camera is close to the top of the roadway or the height is greater than 2 meters; manually set the camera focal length and exposure value, and turn off the camera's auto focus and auto white balance functions.

3. Set the hydraulic support of the coal mining face or the part of the monitoring range of the camera on the coal mining machine that is not blocked by the coal mining machine and the scraper. If the video image data in a certain area changes drastically, and the data change cannot be recovered, and the methane concentration in the nearby area is detected to be normal, a rock burst alarm signal will be issued.

4. Use the difference algorithm to calculate the video image collected by the camera on the hydraulic support of the coal mining face or the camera on the coal mining machine and the set area of the background image; then perform binary processing on the difference image to realize the segmentation of moving objects; when statistics If it is found that the number of pixels in the area where the object has moved exceeds the threshold and lasts for a period of time, and the methane concentration in the nearby area is detected to be normal, then a rock burst alarm signal will be issued.

5. When it is detected that a large number of objects in the video images collected by the cameras of the air inlet and return airways of the coal mining face move toward the roadway at a high speed away from the working face, and at the same time, it is detected that the methane concentration in the nearby area is normal, an impact ground will be issued. Pressure alarm signal.

6. Use the differential algorithm to calculate the video images collected by the cameras of the air inlet lane and the air return lane of the coal mining face and the saved background image of the road; perform binarization processing on the difference image to realize the segmentation of moving objects; When the statistics show that the number of pixels in the area where the object moves in the image of a road exceeds the relevant threshold and continues to increase, and the methane concentration in the nearby area is detected to be normal, a coal burst alarm signal will be issued.

7. When it is detected that multiple objects with fixed original positions in the screen are moving at high speed towards the roadway away from the working face, and at the same time, it is detected that the methane concentration in the nearby area is normal, a rock burst alarm signal will be issued.

8. Set the monitoring area and pre-set the outline shape of the fixed object in the static background image; use the difference algorithm to calculate the video image and the saved background image; perform binarization processing on the difference image obtained from the calculation to achieve Segmentation of moving objects; count the number of intersection pixels between each fixed object outline shape area and the area where the object moves. When the ratio of each intersection pixel number to each fixed object outline shape area pixel number exceeds the relevant threshold and continues to increase, monitor at the same time If the methane concentration in the nearby area is normal, a rock burst alarm signal will be issued.

9. The built-in or external digital signal processor of the camera can be used to complete the analysis and alarm of the video image at the front end of the video acquisition, or directly use the camera with the motion detection function for monitoring, and send a motion alarm when the object is detected to move If the methane concentration in the nearby area is detected to be normal, a rock burst alarm signal will be issued.

Description of drawings

Fig. 1 Schematic diagram of the installation position of the camera and the methane sensor in the coal mining face.

Fig. 2 is a schematic diagram of the system of Embodiment 1.

Fig. 3 is a schematic diagram of the workflow of video management and rock burst alarm in Embodiment 1.

Fig. 4 is a schematic diagram of the system of Embodiment 2.

Fig. 5 is a schematic diagram of the workflow of video management and rock burst alarm in Embodiment 2.

Fig. 6 Schematic diagram of the video image analysis and detection process of the coal mining face camera.

Fig. 7 is a schematic diagram of the detection process of the camera video image in the air inlet lane and the air return lane.

Fig. 8 is a schematic diagram of the detection process for detecting that multiple objects with fixed original positions are all moving at high speed in the direction of the roadway away from the working face.

Detailed ways

The location of the camera is shown in Figure 1, including:

1. The hydraulic support of the coal mining face or the camera (101) on the coal mining machine.

2. The camera (102) in the air return lane of the coal mining face.

3. The camera (103) in the air inlet lane of the coal mining face.

4. Corner methane sensor on coal mining face (104)

5. Methane sensor for coal mining face (105)

6. Methane sensor in air inlet lane of coal mining face (106)

7. Methane sensor for air return lane in coal mining face (107)

8. The methane sensor (108) at the air inlet lane of the serial working face near the air distribution port for series ventilation

Implementation 1:

The system shown in Figure 2 mainly includes:

1. Video recognition server (201). The video recognition server is responsible for processing the video images of various cameras, and sends out alarm and power-off locking signals by analyzing data changes and fault information.

2. Monitoring host (202); with sound and light alarm function, production managers can view live video images, sensor data changes and equipment failures through the monitoring host, and can manually issue alarms and power-off locking signals to cut off all non-intrinsically safe underground coal mines. Power supply of electrical equipment, issue a dispatch order to notify the withdrawal of underground coal mine operators. And can retrieve historical monitoring data from the storage server.

3. Storage server (203); responsible for collecting and storing camera signals, various sensor signals and equipment failure signals, and providing query and retrieval services for users.

4. Network switch (204); responsible for the management and data exchange of all equipment connected to the mining Ethernet.

5. Underground switch (205); responsible for the access and data exchange of substations, with a flameproof enclosure, meeting the explosion-proof requirements of underground coal mines.

6. Substation (206); responsible for the access and data exchange of cameras and sensors, with a flameproof enclosure, which meets the explosion-proof requirements of underground coal mines.

7. Video camera (207); a digital network camera is adopted, equipped with an explosion-proof shell meeting the explosion-proof requirements of coal mines.

8. Methane sensor (208); the methane sensor is a full range or high and low concentration methane sensor, and has an automatic alarm function.

The working process of video management and rock burst alarm is shown in Figure 3:

1. (301) The camera collects video images, digitizes and compresses the video, and transmits the compressed video signal to the substation through a network cable.

2. (302) The video signal and the methane sensor signal are transmitted to the downhole switch through the substation.

3. (303) The underground switch transmits the signal to the network switch on the well, and the network switch distributes the video signal and the methane sensor signal to the storage server, the monitoring terminal and the video recognition server.

4. (304) The video recognition server analyzes and detects the video image and the working conditions of the equipment in real time, and when the alarm condition is satisfied, it sends out an alarm and a power-off locking signal.

5. (305) The monitoring host displays the on-site video, methane sensor data and equipment working conditions in real time, and receives the alarm signal from the video recognition server. When the alarm conditions are met, it will automatically sound and light alarm; Sensor data, alarm conditions and equipment conditions, when the video and data acquisition hardware is damaged, the historical on-site video and sensor data are called. Production management personnel can manually send out alarm and power-off locking signals, cut off the power supply of all non-intrinsically safe electrical equipment in the coal mine, and issue dispatch instructions to notify the evacuation of underground coal mine operators.

Implementation 2:

The system shown in Figure 4 mainly includes:

1. Storage server (401); responsible for collecting and storing video signals, alarm signals, methane sensor signals and equipment working conditions, and providing query and retrieval services for users.

2. Monitoring host (402); with sound and light alarm function, production management personnel can view live video images, alarm conditions, methane sensor data and equipment conditions through the monitoring host, and can manually send out alarms and power-off locking signals to cut off all abnormal Intrinsically safe electrical equipment power supply, issued a dispatch order to notify the withdrawal of underground coal mine workers. And can retrieve historical monitoring data from the storage server.

3. Monitoring the standby machine (403); when the monitoring host breaks down, the monitoring standby machine works.

4. Network switch (404); responsible for the management and data exchange of all equipment connected to the mining Ethernet.

5. Underground switch (405); responsible for the access and data exchange of substations, with a flameproof enclosure, which meets the explosion-proof requirements of underground coal mines.

6. Substation (406); responsible for the access and data exchange of cameras and sensors, with a flameproof enclosure, which meets the explosion-proof requirements of underground coal mines.

7. Video identification alarm device (407); the main processor selects DSP chip to digitize and compress the analog video signal collected by the camera, transmit the compressed video signal to the substation through the network cable, and analyze the video image at the same time Recognition, which can recognize image changes and object movement in the set area, and automatically output an alarm signal when the image changes and object movement reach the set alarm indicators. The video recognition alarm device and the video camera (408) are placed together in an explosion-proof casing meeting the explosion-proof requirements of coal mines.

8. Camera (408): An analog camera is used to output a standard analog video signal, and it is placed together with the video recognition alarm device (407) in an explosion-proof casing that meets the explosion-proof requirements of coal mines.

9. Methane sensor (409); the methane sensor is a full range or high and low concentration methane sensor, and has an automatic alarm function.

The working process of video management and rock burst alarm in implementation plan 2 is shown in Figure 5:

1. (501) The camera collects video data, transmits the video analog signal to the video recognition alarm device, and the video recognition alarm device digitizes and compresses the video, and transmits the compressed video signal and mobile alarm signal to the substation.

2. (502) The video recognition alarm device analyzes and recognizes video images, and can identify image changes and object movements in the set area. When the image changes and object movements reach the set alarm indicators, an alarm signal is automatically output to the substation.

3. (503) Each substation sends the collected signals to the underground switch through the mine Ethernet ring network.

4. (504) The network switch receives the data forwarded by the downhole switch, and distributes the video signal, sensor signal and alarm signal to the storage server, monitoring host and backup host.

5. (505) The storage server stores the video signal, sensor signal and alarm signal.

6. (506) The monitoring host displays live video, sensor data, alarm conditions and equipment working conditions in real time. When the alarm conditions are met, it will automatically sound and light alarm; production management personnel can view live real-time video, alarm conditions and equipment working conditions, When the video and data acquisition hardware is damaged, the historical on-site video and data are called. Production management personnel can manually send out rock burst alarms and power-off lockout signals to cut off the power supply of all non-intrinsically safe electrical equipment in the coal mine, and issue dispatch instructions to notify the evacuation of underground coal mine operators.

The detection process of the video image of the camera in the coal mining face is shown in Figure 6:

1. (601) Set the monitoring area A within the video monitoring range, and retrieve the set area data each time the recognition server is started.

2. (602) Restore the standard compressed video stream to image frames.

The currently captured video images are:

F = {f(x, y), x ∈ M, y ∈ N, MN}

The image resolution is M×N, (x, y) is the coordinate of any point in the video image, and f(x, y) is the gray value of the point (x, y) in the video image.

3. (603) The monitoring area of the video image and the background image at intervals of 5 frames is calculated using a difference algorithm, and the background image is automatically updated at intervals of a period of time.

Let the background image be:

B = {b(x, y), x∈M, y∈N, MN}

Let the current video image and each corresponding pixel of the background image perform a differential operation to obtain a differential image G:

The difference image after difference operation is:

G = {g(x, y), x ∈ M, y ∈ N, MN}

The calculation formula is:

$\mathrm{<span\; class="notranslate">\; g</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; ;</span>& <span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; A</span>}\\ \mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span>& <span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&NotElement;</span>\mathrm{<span\; class="notranslate">\; A</span>}\end{array}\right.$

4. (604) Carry out binarization processing to difference image and realize the segmentation of moving object, realize the segmentation of moving object; If the gray scale change of certain pixel exceeds setting threshold value T, then set the gray scale value of this pixel to It is set to 255, indicating the area where the object moves, otherwise it is 0, indicating the area where the object does not move.

The image expression after binarization is:

D = {d(x, y), x ∈ M, y ∈ N, MN}

The calculation formula is:

5. (605) Carry out numerical statistics on the pixels in the region where the object moves, and enter the early warning state if the number of pixels exceeds the threshold L (606);

The statistical calculation formula is:

$\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 255</span>}}\underset{\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 255</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span>$

6. (607) If no video data stream output is detected in the early warning state, an alarm is triggered (609);

7. (608) If the warning state lasts for 5 seconds and exceeds the threshold L, an alarm is triggered (609);

8. (610) In the early warning state, if the number of pixels in the area where the object moves is detected to be lower than the threshold L, the early warning is released (611);

The detection process of camera video images in the air inlet lane and the air return lane is shown in Figure 7 (some expression formulas can refer to the description in Figure 6 above):

1. (701) Restore the standard compressed video stream to an image frame.

2. (702) The difference algorithm is used to calculate the video image and the background image every 5 frames, and the background image is updated at intervals.

3. (703) Perform binarization processing on the difference image to realize segmentation of moving objects.

4. (704) Count the number of pixels in the area where the object moves, and enter the early warning state (705) if the number of pixels exceeds the threshold L1;

5. (706) In the pre-warning state, if no alarm is triggered within 10 seconds, the pre-warning state is automatically released (707).

6. (708) In the early warning state, if no video data stream output is detected, an alarm is triggered (713);

7. (709) In the early warning state, determine whether the number of pixels in the area where the object moves in the video image continues to increase, and if it increases, add 1 to the counter (711), otherwise the counter is cleared and returns (710).

8. (712) If the counter exceeds 12, it means that the number of pixels in the region where the object moves in the video image is detected to increase continuously for 1 second, and an alarm is triggered (713);

The implementation method of detecting that multiple objects with fixed original positions are all moving at high speed in the direction of the roadway away from the working face is shown in Figure 8:

1. (801) Set the monitoring area A within the video monitoring range, if it has been set, the set area data will be called every time the recognition server is started.

2. (802) Set the outline shape of the fixed object in the static background image, and multiple object outline shape areas B ₁ to B ₂ can be set, and the outline shape data of the set fixed object is called every time the recognition server is started. The numbers of shape pixels are s ₁ to s _, respectively.

3. (803) Restore the standard compressed video stream to image frames.

4. (804) The difference algorithm is used to calculate the video image and the background image every 5 frames, and the background image is updated at intervals.

5. (805) Binarize the difference image to realize the segmentation of the moving object area.

6. (806) Count the intersection pixel numbers p ₁ -p ₂ of each fixed object outline shape area and the area where the object moves.

7. (807) If the ratio of all p _i to the number of pixels s _i of the set fixed object outline area satisfies greater than R%, then enter the early warning state (808);

where i=1,2,...,n

8. (809) In the early warning state, if no alarm is triggered within 10 seconds, the early warning state (810) is automatically released.

9. (811) In the early warning state, if no video data stream output is detected, an alarm is triggered (816);

10. (812) In the pre-warning state, if it is detected that the moving area in the monitoring area continues to increase, the counter is incremented by 1 (814), otherwise the counter is cleared and returns (813).

11. (815) If the counter exceeds 12, it means that the movement area in the monitoring area is detected to increase for 1 second, and an alarm is triggered (816).

Claims (9)

1. An image-based coal mining face rock burst alarm method is characterized in that: under the coal mine, on the hydraulic support of the coal mining face or on the coal mining machine, in the air inlet lane of the coal mining face, at the back of the coal mining face Place the camera in the wind lane; detect the collected video image data in real time, detect the working conditions of the camera and communication lines, and monitor the change of the methane sensor value in the nearby area; when an abnormal object appears in the set area in the camera video image When a large number of objects are piled up or moving at high speed, it is judged as data anomaly; when data anomalies are detected, or data anomalies and related equipment failures occur one after another in a short period of time, and the methane concentration in the nearby area is detected to be normal, a rock burst will be issued. Alarms and signals. 2. The alarm method according to claim 1, characterized in that: the installation position of the camera is close to the top of the roadway or the height is greater than 2 meters; the focal length and exposure value of the camera are manually set, and the automatic focus and automatic white balance functions of the camera are turned off. 3. The alarm method according to claim 1, characterized in that: set the hydraulic support of the coal mining face or the part of the camera monitoring range on the coal mining machine that is not blocked by the coal mining machine and the scraper. Under the condition of normal work and normal video communication, when the video image data in the set area is detected to change drastically, and the data change is irreversible, and the methane concentration in the nearby area is detected to be normal, a rock burst alarm signal will be issued. 4. The alarm method as claimed in claim 3, characterized in that: the video image collected by the camera on the coal mining face hydraulic support or the coal shearer and the setting area of the background image are calculated using a difference algorithm; Carry out binarization processing to realize the segmentation of moving objects; when the number of pixels in the area where the object moves exceeds the threshold and lasts for a period of time, and the methane concentration in the nearby area is detected to be normal, a rock burst alarm signal is issued. 5. The alarm method according to claim 1, characterized in that: when it is detected that there are a large number of objects moving at a high speed to the roadway away from the direction of the working face in the video images collected by the cameras of the air inlet lane and the air return lane of the coal mining face, At the same time, if the methane concentration in the nearby area is detected to be normal, a coal and gas outburst alarm signal will be issued. 6. The method for reporting to the police as claimed in claim 5, characterized in that: the video image collected by the camera of the air inlet lane and the air return lane of the coal mining face and the preserved road background image are calculated using a difference algorithm; Value image is binarized to realize the segmentation of moving objects; when the statistics show that the number of pixels in the area where the object moves in a certain image exceeds the relevant threshold and continues to increase, and at the same time it is detected that the methane concentration in the nearby area is normal, a coal mine will be sent. Rockburst warning signal. 7. The alarm method according to claim 1, characterized in that: when it is detected that multiple objects with fixed original positions in the picture are moving at high speed to the roadway away from the working face, and the methane concentration in the nearby area is detected to be normal at the same time, then A rockburst alarm signal is issued. 8. The alarm method as claimed in claim 7, characterized in that: the monitoring area is set and the contour shape of the fixed object in the static background image is set in advance; the difference algorithm is used to calculate the video image and the saved background image; Carry out binarization processing on the difference image obtained from the calculation to realize the segmentation of moving objects; count the number of intersection pixels between the contour shape area of each fixed object and the area where the object moves, when the number of pixels of each intersection and the contour shape area of each fixed object If the pixel number ratio exceeds the relevant threshold and continues to increase, and the methane concentration in the nearby area is detected to be normal, a rockburst alarm signal will be issued. 9. The alarm method according to claim 1, characterized in that: the digital signal processor built in or external to the camera can be used to complete the analysis and alarm of the video image at the front end of the video acquisition, or directly use the camera with the motion detection function The camera monitors and sends out a motion alarm signal when it detects that the object is moving, and at the same time monitors that the methane concentration in the nearby area is normal, then sends out a rock burst alarm signal.