Mine fire identification and alarm method based on image contour area features

Abstract

The application discloses a mine fire identification and alarm method based on image contour area features, which comprises the following steps: installing a camera in a mine, collecting and identifying a suspected fire image of a monitoring area in real time, and identifying and alarming a mine fire by monitoring irregular shape features of a fire flame burning image contour. The mine fire identification and alarm method fully considers the features of the fire flame burning image, is fast, accurate, simple and easy to identify, and can help a person in danger in the mine to gain valuable rescue and escape time.

Description

TECHNICAL FIELD

[0001] The present application relates to a mine fire identification and alarm method based on image contour area features, which relates to the fields of digital image processing technology and communication technology. BACKGROUND

[0002] In the coal industry, gas, fire, flood, roof, coal dust and other accidents have plagued coal mine safety production. Once a mine fire occurs, if the fire cannot be controlled in time, the scope of the fire will quickly expand, causing a large number of casualties and property losses; further, it may cause an explosion in the high-concentration gas and coal dust area underground as an ignition source, causing secondary trauma to the underground environment. Therefore, it is crucial for coal safety production to quickly identify the mine fire disaster and timely alarm, and to start the emergency plan and emergency rescue accordingly.

[0003] In the existing mine fire monitoring technology, there are various types of sensor comprehensive monitoring methods, such as temperature, smoke, gas, etc. sensors. Although this type of monitoring has the advantages of high reliability and simple operation, it is affected by the complex underground operation scene, and has a large workload of arrangement and maintenance and is greatly affected by interference sources. The infrared radiation temperature measurement equipment monitoring method can measure the temperature of the fire source, but the temperature measurement accuracy is greatly affected by the underground environmental factors and interference sources. The visual feature monitoring method is the current mainstream mine fire monitoring technology. However, the existing monitoring technology adopts many characteristic variables to determine the fire, although the information features contain a lot of information, there are also many unstable factors. Based on the comprehensive determination of fire disaster by multiple cameras and multiple frames of images, the determination time is long, and the false alarm rate and the missed alarm rate are high.

[0004] There is no natural light source such as sunlight, moonlight, starlight and lightning in the underground mine, and the main factor affecting the identification of mine fire is the mine light source. Therefore, it is necessary to study a new method for quickly identifying mine fire, shorten the determination time, and reduce the false alarm rate and the missed alarm rate of the existing monitoring technology for mine fire perception. Early detection of mine fire and timely alarm of fire disaster are important guarantees for timely emergency rescue and saving the lives of people in danger underground. SUMMARY

[0005] The present application aims to provide a mine fire identification and alarm method based on image contour area characteristics, which fully considers the image contour characteristics of fire flame burning images and the image contour characteristics of fire interference sources, can quickly distinguish fire and fire interference sources, is accurate and simple, and can save more rescue and escape time for people in distress in the mine. The fire identification and alarm method comprises installing a camera at a roadway, a fully mechanized coal mining face, a tunneling face, a belt conveyor and a conveyor belt to monitor a monitoring area in real time, determining whether a difference between an area of a convex circumscribed polygon of a suspected fire image contour and an actual area of the image contour meets a set condition to determine a fire alarm, determining whether a ratio between a number of times of determining a fire alarm in a set time and a total number of determinations meets a set condition to determine a mine monitoring area fire, and immediately sending a fire alarm signal to a monitoring terminal when a mine monitoring area fire is determined.

[0006] The working process of the fire alarm determination comprises:

[0007] Step 1: The camera collects video images of the mine monitoring area in real time.

[0008] Step 2: The video images are frame-processed, the frame images are image-filtered and denoised, image-enhanced and image-pixel binarized, and then the total pixel feature value of each frame image is calculated.

[0009] Step 3: The total pixel feature value of each frame image is determined, and if the total pixel feature value N f (f≥1) of the fth frame image is greater than a pre-set threshold value H, the fth frame image is determined to be a suspected fire image, and step 4 is performed; if the suspected fire image determination condition is not met, the monitoring is continued.

[0010] Step 4: The convex circumscribed polygon of the suspected fire image contour is calculated, and step 5 is performed.

[0011] Step 5: The area of the convex circumscribed polygon of the image contour and the actual area of the image contour are calculated, and step 6 is performed.

[0012] Step 6: If the difference between the area of the convex circumscribed polygon and the actual area of the image contour is greater than a pre-set threshold value R, a mine fire occurs, and step 7 is performed; otherwise, the monitoring is continued.

[0013] Step 7: The fire alarm is determined, and step 1 is returned to be performed.

[0014] 1. The fire identification and alarm method further includes: the convex circumscribed polygon of the image contour is the smallest circumscribed convex polygon of the image contour; the smallest circumscribed convex polygon of the image contour is the polygon formed by finding the coordinates of all the outermost convex points on the boundary of the suspected fire image contour and then connecting all the outermost convex points with straight lines in sequence.

[0015] 2. The fire identification and alarm method further includes: the outermost convex point is defined as the two outermost convex points when the straight line containing any two outer convex points on the image contour boundary intersects the contour boundary only at the two outer convex points.

[0016] 3. The fire identification and alarm method further includes: the external convex point is an inflection point on the image contour boundary; the inflection point also includes an internal concave point on the image contour boundary.

[0017] 4. The fire identification and alarm method further includes: the inflection point is any consecutive three coordinate points (x, y, y) taken on the image contour boundary. i-1 ,y i-1 ),(x i ,y i ),(x i+1 ,y i+1 If (x) i-1 ,y i-1 ),(x i ,y i Collinear; (x) i-1 ,y i-1 ),(x i+1 ,y i+1 (x) are collinear; but (x) i-1 ,y i-1 ),(x i ,y i ),(x i+1 ,y i+1 If they are not simultaneously collinear, then the coordinates of the point (x) i ,y i ) is the inflection point on the boundary of its image contour, that is, the external convex point or the internal concave point; x and y represent the coordinate values ​​of the coordinate points on the boundary of the image contour, and i represents the sequence number of the coordinate points on the boundary of the image contour.

[0018] 5. The fire identification and alarm method further includes: the step of determining the convex points and concave points on the image contour boundary includes:

[0019] Step 1: Using the formula Find the coordinates of all inflection points on the image contour boundary; where, {(x i-1 ,y i-1 ),(x i ,y i ),(xi+1 ,y i+1 )} represents any adjacent coordinate point group on the image contour boundary; k i-1 , k i+1 respectively represent the slope of the straight line on which the two coordinate points are located, and when k i-1 ≠ k i+1 , then the coordinate point (x i ,y i ) represents an inflection point coordinate on the image contour boundary;

[0020] Step 2: the actual area of the image contour is calculated by the formula ; in the formula, S d represents the actual area of the image contour; {(x1, y1) (x2, y2)…(x j ,y j )} represents all the coordinate points on the image contour boundary arranged in anticlockwise order;

[0021] Step 3: a coordinate point (x e ,y e ) on the image contour boundary is randomly selected, then the left and right adjacent coordinate points (x e-1 ,y e-1 ), (x e+1 ,y e+1 ) are directly connected to form a new contour boundary, and the area of the new contour is calculated by the formula ; in the formula, S a represents the area of the new contour; {(x1, y1) (x2, y2)…(x e-1 ,y e-1 ) (x e+1 ,y e+1 )…(x j ,y j )} represents all the coordinate points on the new contour boundary arranged in anticlockwise order;

[0022] Step 4: when S a <S d , then the coordinate point (x e ,y e ) is an outer convex point on the image contour boundary; when S a >S d , then the coordinate point (x e ,y e ) is an inner concave point on the image contour boundary.

[0023] 6. The fire identification and alarm method further comprises that the area of the convex hull polygon of the suspected fire image contour is calculated by the formula ; in the formula, A trepresents the area of the convex circumscribed polygon; {(m1, n1), (m2, n2)…(m p ,n p )} represents all the coordinate points arranged in the counterclockwise direction on the convex circumscribed polygon.

[0024] 7. The fire identification and alarm method further comprises that the actual area of the suspected fire image contour is obtained by using the formula ; in the formula, S d represents the actual area of the image contour; {(x1, y1), (x2, y2)…(x j ,y j )} represents all the coordinate points arranged in the counterclockwise direction on the contour boundary.

[0025] 8. The fire identification and alarm method further comprises that the fire determination process of the underground monitoring area comprises that the fire alarm determination is cyclically executed for a duration T seconds, when the ratio of the suspected fire image alarm number variable X satisfying the fire alarm determination condition to the total determination number variable Y of the image is greater than a set threshold S, that is, X / Y>S, it is determined that the fire occurs in the underground monitoring area, and a fire alarm signal is immediately sent to the monitoring terminal.

[0026] 9. The fire identification and alarm method further comprises that the camera comprises a visible light camera, a far infrared camera, a near infrared camera and an ultraviolet camera.

[0027] The mine fire identification and alarm method based on the number of image contour edges has the following characteristics:

[0028] 1. There is no sunlight, moonlight, starlight and lightning and other natural light sources in the underground mine, and the main influence on the mine fire identification is the mine light source. The fire identification method fully analyzes the image contour characteristics of the fire flame burning image and the image contour characteristics of the interference source of the suspected fire, the method is based on the difference information between the area of the convex circumscribed polygon of the image contour and the actual area of the image contour for fire determination, which is more conducive to distinguishing and eliminating the interference source of the suspected fire, and can reduce the false alarm and the missed alarm of the fire disaster.

[0029] 2. The method not only efficiently utilizes the characteristic information of the fire image, but also has simple and efficient algorithm construction in the image processing part, and the fire information can be quickly determined by a single frame image, the method has high feature extraction recognition degree and few variables, thereby shortening the identification time of the suspected fire and further saving more escape time for the trapped and affected personnel in the underground mine. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 The figure shows the implementation scheme of the mine fire identification and alarm method based on the image contour area characteristics.

[0031] Figure 2 Fire monitoring alarm device structure diagram.

[0032] Figure 3 Fire alarm process diagram of mine fire identification and alarm method based on image contour area feature

[0033] Figure 4 Fire determination process diagram of mine fire identification and alarm method based on image contour area feature DETAILED DESCRIPTION

[0034] Figure 1 is an embodiment example of mine fire identification and alarm method based on image contour area feature, the main components include:

[0035] 1. Storage server (101), which is in communication connection with image acquisition camera (105), is responsible for storing and forwarding real-time video image data of monitoring area provided by camera; the storage server is in communication connection with fire monitoring alarm device (106), is responsible for storing and forwarding fire alarm data provided by fire monitoring alarm device; provides monitoring terminal (102) with services of viewing on-site monitoring data of monitored area or calling historical monitoring image data.

[0036] 2. Monitoring terminal (102), which is installed on the surface of the well, is used for receiving alarm signal of the fire monitoring alarm device (106); is responsible for providing display service of underground environment monitoring data, provides real-time, historical data and fire alarm data by the storage server (101), has sound and light alarm function; production management personnel can call and query historical data stored in the storage server (101) through the monitoring terminal.

[0037] 3. Core switch (103), core management and exchange device of mine Ethernet, is responsible for management and data exchange of all devices connected to mine Ethernet, has routing function, and is connected to Internet.

[0038] 4. Ring network switch (104), underground exchange device of mine Ethernet, which is installed underground, multiple ring network switches are connected in ring network mode.

[0039] 5. A camera (105) is installed in the underground roadway, coal mining face and tunneling face, responsible for collecting real-time video images of the underground roadway and working face where fire is likely to occur; the camera includes a visible light camera, a far-infrared camera, a near-infrared camera and an ultraviolet camera; the camera has network output and analog video output functions; the network interface is directly connected to a ring network switch (104), and the video image data of the monitored area collected in real time is transmitted to a storage server (101); the analog video output port is connected to a fire monitoring and alarm device (106).

[0040] 6. A fire monitoring and alarm device (106) is responsible for receiving and processing the monitoring video image data collected by the camera (105), and built-in graphic image processing software, when the processed suspected fire image meets the set fire alarm conditions, it sends fire alarm data to the storage server (101); and sends a fire alarm signal to the monitoring terminal (102); has wired and wireless communication functions.

[0041] 7. Suspected fire area, suspected ignition point area in the real-time monitoring area of the camera, the image data is monitored and collected by the camera in real time.

[0042] Figure 2 is a schematic diagram of the structure of the fire monitoring and alarm device, the main components of the structure of the fire monitoring and alarm device include:

[0043] Core processor, graphics processor, storage unit, clock unit, power unit, USB interface unit, video image acquisition module, wireless communication unit, network interface unit, SD card interface unit.

[0044] 1. Core processor (201), using Broadcom BCM2837 processor, using ARM Cortex-A53 architecture, 64-bit quad-core 1.2GHz.

[0045] 2. Graphics processor (202), using Dual Core VideoCore IV GPU processor.

[0046] 3. Storage unit (203), using 1GB LPDD2 memory.

[0047] 4. Clock unit (204), using 19.2MHz crystal oscillator.

[0048] 5. Power unit (205), using AC / DC module, input 100V-240VAC, output 12VDC, used for device power supply.

[0049] 6. USB interface unit (206), supporting 4 USB interfaces.

[0050] 7. Video image acquisition module (207), converting the mode video signal into digital video data, inputting the analog video output port of the camera (105), transmitting the digital video data to the core processor (201) through the USB port, supporting multi-channel video acquisition.

[0051] 8. Communication module (208), responsible for converting the RS485 communication interface into a USB communication interface, connecting other data monitoring devices.

[0052] 9. Wireless communication unit (209), supporting 802.11b / g / n protocol, used for wireless communication with monitoring devices supporting wireless communication.

[0053] 10. Network interface unit (210), responsible for accessing the mine Ethernet, connecting the ring network switch (104).

[0054] 11. SD card unit (211), used for storing system files, library files, monitoring program files, etc., using Linux system management, built-in OpenCV library for video data processing, using a Micro SD card not less than 4GB.

[0055] The fire alarm flowchart of the mine fire identification and alarm method based on the image contour area feature is shown in Figure 3 , which includes:

[0056] 1. (301) Set the internal timer time T of the fire monitoring and alarm device to zero and start timing, set the total number of image determination variables Y = 0 for executing the fire alarm determination condition, set the number of fire alarm variables X = 0 for the suspected fire image meeting the fire determination alarm condition, and execute step (302).

[0057] 2. (302) Further set the total number of image determination variables Y = Y + 1 for executing the fire alarm determination condition, and the determination variable is accumulated by one every time the fire alarm determination condition is executed, and step (303) is executed.

[0058] 3. (303) The fire monitoring and alarm device processes the image and determines the suspected fire image, and when the set fire alarm determination condition is met, step (304) is executed, otherwise step (302) is returned.

[0059] 4. (304) The suspected fire image meets the fire alarm determination condition every time, and the fire alarm variable is accumulated by one, that is, X = X + 1, and a fire warning is sent to the monitoring terminal, and step (305) is executed.

[0060] 5. (305) The workflow of fire alarm judgment is executed in a cycle within a set time, and it is judged whether the judgment timer time value T is greater than the set time threshold t. When the time value T > t, step (306) is executed, otherwise, step (302) is returned to be executed.

[0061] 6. (306) It is further judged whether the ratio of the fire image alarm number X satisfying the fire alarm judgment condition to the total judgment number Y of the image executing the fire alarm judgment condition is greater than the set threshold S. When X / Y > S, step (307) is executed, otherwise, step (301) is returned to be executed.

[0062] 7. (307) The fire monitoring alarm device sends a fire alarm signal to the storage server (101) and the monitoring terminal (102).

[0063] The fire judgment process example of the mine fire identification and alarm method based on the image contour area feature is shown in the figure. Figure 4 The process includes:

[0064] 1. (401) The camera is installed at the roadway, fully mechanized working face, heading working face, belt conveyor and conveying belt, and real-time video images of the monitoring area are collected and uploaded to the fire monitoring alarm device for image recognition processing, and further step (402) is executed.

[0065] 2. (402) The fire monitoring alarm device pre-processes the monitoring video image. After the monitoring video image is subjected to image framing, image denoising, image enhancement and image pixel binarization, the total pixel feature value of each frame image is calculated, and further step (403) is executed.

[0066] 3. (403) The fire monitoring alarm device judges the total pixel feature value of the frame image frame by frame. If the total pixel feature value N f of the f frame image is greater than H, step (404) is executed; if the suspected fire image judgment condition is not met, step (401) is returned to be executed.

[0067] 4. (404) The f frame image is judged as a suspected fire image, and step (405) is executed.

[0068] 5. (405) The minimum circumscribed convex polygon of the suspected fire image contour is calculated: according to the formula All the coordinates of the inflection points on the suspected fire image contour boundary of the image are calculated. All the coordinates of the outer convex points in the inflection point coordinates are further calculated, and the actual area of the image contour is calculated by using the formula Any inflection point coordinate (x e ,y e ) on the image contour boundary is taken, and then the left and right adjacent inflection point coordinates (x e-1 ,y e-1),(x e+1 ,y e+1 ) directly straight line connection, constitute a new contour boundary, then use the formula to find the area of the new contour; when the determination S a <S d , then the inflection point (x e ,y e ) is the outer convex point on the image contour boundary. Further find all the outermost convex point coordinates in the outer convex point coordinates: take two outer convex point coordinates (x b ,y b ), (x c ,y c ) on the image contour boundary, the slope of the straight line on which the two outer convex point coordinates is , then the equation of the straight line passing through the two outer convex points is , then if the remaining outer convex points on the image contour boundary do not satisfy the equation except the two outer convex points (x b ,y b ), (x c ,y c ), then determine that the outer convex point (x b ,y b ), (x c ,y c ) is the outermost convex point coordinate on the image contour boundary. Then connect the outermost convex points obtained in sequence by straight lines to obtain the minimum circumscribed convex polygon of the image contour, and execute step (406).

[0069] 6. (406) find the area A t of the minimum circumscribed convex polygon of the image contour according to the formula ; find the actual area S d of the suspected fire image contour according to the formula , then execute step (407).

[0070] 7. (407) when the fire monitoring and alarm device monitors the difference A t -S d between the area of the minimum circumscribed convex polygon of the image contour and the actual area of the image contour R, execute step (408), otherwise return to execute step (401).

[0071] 8. (408) perform fire alarm processing, and return to execute step (401).

Claims

1. A mine fire identification and alarm method based on image contour area features, characterized in that: The camera is installed at a roadway, a fully mechanized coal mining face, a tunneling face, a belt conveyor and a conveying belt to monitor a monitoring area in real time, fire alarm determination is performed according to whether a difference between an area of a convex circumscribed polygon of a suspected fire image contour and an actual area of the image contour satisfies a set condition, underground monitoring area fire determination is performed according to whether a ratio between a number of times of determination as fire alarm in a set time and a total number of determination times satisfies a set condition, and a fire alarm signal is immediately sent to a monitoring terminal when it is determined that a fire occurs in the underground monitoring area; The working process of the fire alarm determination comprises: Step 1: a camera acquires video images of an underground monitoring area in real time; Step 2: frame pre-processing is performed on the video images, frame image filtering and denoising, image enhancement and image pixel binarization are performed, and total pixel feature values of single frame images are calculated respectively; Step 3: Determine the total value of the frame image pixel features frame by frame, if the total value of the frame image pixel features of the first frame image is greater than the pre-set threshold value total value of the frame image pixel features , , then determine the first frame image as a suspected fire image, and execute Step 4; if the suspected fire image determination condition is not met, then continue monitoring; ​​ Step 4: a convex circumscribed polygon of a suspected fire image contour is obtained, and step 5 is executed; Step 5: an area of the convex circumscribed polygon of the image contour and an actual area of the image contour are obtained, and step 6 is executed; Step 6: When the difference between the area of the convex hull polygon and the actual area of its image contour is greater than a pre-set threshold value, a mine fire situation occurs, and step 7 is executed; otherwise, return to continuous monitoring. , a mine fire situation occurs, and step 7 is executed; otherwise, return to continuous monitoring. Step 7: fire alarm is performed, and step 1 is returned to be executed; The convex circumscribed polygon of the image contour is a minimum circumscribed convex polygon of the image contour; the minimum circumscribed convex polygon of the image contour is a polygon formed by sequentially connecting all outermost convex point coordinates in outer convex point coordinates on a boundary of the suspected fire image contour; The outer convex point is an inflection point on the boundary of the image contour; the inflection point further includes an inner concave point on the boundary of the image contour; The determination steps of the outer convex point and the inner concave point on the boundary of the image contour comprise: Step 1: Using the formula Find the coordinates of all inflection points on the image contour boundary; where, {(x i-1 ,y i-1 ), (x i ,y i ), (x i+1 ,y i+1 )} represents any group of adjacent coordinate points on the image contour boundary; k i-1 k i+1 Let represent the slopes of the lines containing the two coordinate points, and when k i-1 ≠k i+1 When, then the coordinate point (x i ,y i () represents the coordinates of the inflection point on the image contour boundary; Step 2: find the actual area of the image contour by using the formula , where S d represents the actual area of the image contour; and { (x1, y1) (x2, y2)... (x j , y j )} represents all the coordinate points arranged in reverse clockwise order on the boundary of the image contour. Step 3: If a coordinate point (x e ,y e ) is a corner point on the contour of an arbitrary image, then the two adjacent coordinate points (x e-1 ,y e-1 ), (x e+1 ,y e+1 ) are directly connected by a straight line to form a new contour, and the area of the new contour is calculated by the formula ; where S a represents the area of the new contour; and { (x1, y1) (x2, y2) … (x e-1 ,y e-1 ) (x e+1 ,y e+1 ) … (x j ,y j )} represents all the coordinate points on the new contour in anticlockwise order. Step 4: When judging S a <S d , then the inflection point (x e ,y e ) is the outer convex point on the image contour boundary; when judging S a >S d , then the inflection point (x e ,y e ) is the inner concave point on the image contour boundary.

2. The mine fire identification and alarm method based on image outline area features according to claim 1, characterized in that: The outermost convex point is represented as when a straight line on which any two outer convex points on the boundary of the image contour are located only intersects the contour boundary at the two outer convex points, the two outer convex points are referred to as the outermost convex points.

3. The mine fire identification and alarm method based on image outline area features according to claim 1, characterized in that: The inflection point is any continuous three coordinate points (x i-1 ,y i-1 ), (x i ,y i ), (x i+1 ,y i+1 ) on the image contour boundary, if (x i-1 ,y i-1 ), (x i ,y i ) are collinear; (x i-1 ,y i-1 ), (x i+1 ,y i+1 ) are collinear; but (x i-1 ,y i-1 ), (x i ,y i ), (x i+1 ,y i+1 ) are not collinear at the same time, then the coordinate point (x i ,y i ) is an inflection point on the image contour boundary, namely an outer convex point or an inner concave point; the x, y represent the value of the coordinate point on the image contour boundary, and i represents the serial number of the coordinate point on the image contour boundary.

4. The mine fire identification and alarm method based on image outline area features according to claim 1, characterized in that: The area of the convex hull polygon of the suspected fire image contour is calculated using the formula A = 1 / 2 * |∑(xiyi+1-xi+1yi)| t A represents the area of the convex hull polygon;{ (m1, n1), (m2, n2)... (m p , p n )} represents all coordinate points arranged in the counterclockwise direction on the convex hull polygon.

5. The mine fire identification and alarm method based on image outline area features according to claim 1, characterized in that: The actual area of the suspected fire image contour is calculated using the formula S = A d The actual area of the suspected fire image contour is calculated using the formula j j The actual area of the suspected fire image contour is calculated using the formula​ 6. The mine fire identification and alarm method based on image outline area feature according to claim 1, characterized in that: Satisfying the set condition to make the downhole monitoring area fire determination includes the duration The work of fire alarm determination is executed in a 1-second cycle, and when the number of suspected fire image alarms satisfying the fire alarm determination condition is greater than the total number of determination times of the image The ratio is greater than the set threshold When , it is determined that a fire occurs in the downhole monitoring area, and a fire alarm signal is immediately sent to the monitoring terminal.

7. The mine fire identification and alarm method based on image outline area features according to claim 1, characterized in that: The camera comprises a visible light camera, a far infrared camera, a near infrared camera and an ultraviolet camera.

Images