Fire operation safety supervision method using unmanned aerial vehicle video monitoring

Abstract

The invention provides a fire operation safety supervision method using unmanned aerial vehicle video monitoring. The method comprises the following steps: step 1, extracting on-site fire operation information and binding an unmanned aerial vehicle; 2, setting a fire operation safety supervision area; 3, video effective data are obtained, wherein the video effective data comprise the on-site gas cylinder, the fire extinguisher placement state and the gas cylinder carrying operation behavior of on-site operation personnel; 4, detecting a fire operation safety supervision video image; and step 5, fire operation safety actual condition judgment and risk early warning are carried out. The unmanned aerial vehicle is used for video monitoring, so that the problems of low detection accuracy, multiple monitoring blind areas and poor effect of fire operation in a field environment can be well solved, and the field unmanned aerial vehicle calls to timely stop unsafe behaviors of people and relieve unsafe states of objects by measuring and calculating whether the actual distance between a detected object and a fire point reaches a preset safe distance.

Description

Technical Field

[0001] The present invention relates to the field of HSE management information technology, and in particular to a method for safety supervision of hot work operations using unmanned aerial vehicle video monitoring. Background Art

[0002] Hot work refers to various temporary operations that can directly or indirectly produce open flames in production or construction areas with fire and explosion hazards. The production and construction sites of the oil and gas industry are flammable and explosive. Hot work is a special operation and a high-risk operation in the oil and gas industry. The unsafe behavior of people on site and the unsafe state of objects can easily lead to major safety accidents, causing huge property losses and casualties. Regarding the supervision of hot work, the management regulations require that there be special personnel to supervise hot work, but the supervision work mainly relies on the experience of the supervisors, which is highly subjective and the supervision is not strong enough, resulting in many hidden dangers in the operation process that are difficult to be discovered. At the same time, the entire hot work process cannot be fully recorded and traced. Therefore, relying on safety regulations or manual experience to check the safety hazards at the construction site can no longer meet the complex and dynamic safety management requirements of hot work construction.

[0003] With the development of Internet of Things technology and video intelligent analysis technology, the commonly used methods at this stage are mainly divided into two types: one is the method of using sensors, but due to the large environmental influence and calculation limit of production and construction sites, the real-time detection accuracy is low, and the application effect cannot meet the requirements of on-site safety management and control; some researchers have tried to use fixed cameras to automatically identify abnormal conditions during hot work operations, but due to the distribution and field of view of fixed cameras, it is difficult to realize automatic monitoring of any hot work location within the construction site. The actual application is mostly limited to the staff remotely viewing video surveillance, and the application effect is poor.

[0004] The use of drone video surveillance can solve the above problems well. By measuring whether the actual distance between the detection object and the fire point reaches the preset safety distance, a shouting module and a fire extinguishing module are added, and general image intelligent detection algorithms such as open smoke, open flame, not wearing work clothes, not wearing safety helmets, vehicle intrusion, personnel intrusion, single and double inspections, smoking, making phone calls, and falling are loaded, as well as special image intelligent detection algorithms related to fire operations such as stepping on gas cylinders, carrying gas cylinders on shoulders, no fire extinguishers at the fire operation site, and bottles and cans placed upside down at the fire operation site. By shouting through the on-site drone, unsafe behaviors of people can be stopped and unsafe states of objects can be eliminated in time. The fire extinguishing module mounted on the on-site drone can be used to timely avoid the possibility of small sparks turning into fire accidents. It enriches safety supervision means, expands safety supervision content, and optimizes safety supervision effects. It has very good practical application effects and promotion and application significance.

[0005] The patent number is 202211461239.X, and the application date is 2022-11-17. It discloses a method for safety supervision of hot work operations on construction sites using tower crane video image detection, including the following steps: extracting hot work information on the construction site; setting a hot work safety supervision area; acquiring effective data from the tower crane hook video for hot work safety supervision; detecting the tower crane video image for hot work safety supervision; and determining the actual safety situation of hot work and issuing risk warnings.

[0006] The above patent uses the camera posture data as a priori conditions to screen effective camera images, reducing the number of image processing and computing costs; uses deep learning algorithms to automatically detect and identify safety images, improves safety supervision efficiency, and realizes intelligent supervision of hot work; uses the camera data at the construction site, without adding additional processes, to monitor abnormal situations that may occur during hot work at the construction site from a low-altitude perspective, enriching the means of supervision. However, the production sites and construction sites of the oil and gas industry are often in the field and do not have high-altitude fixed building conditions such as tower cranes. In the field construction site, due to the complex and changeable environment, there are many monitoring blind spots, low detection accuracy and poor detection effect. Therefore, this method cannot meet the supervision of hot work at the field construction site. Summary of the invention

[0007] The technical problem to be solved by the present invention is to provide a method for safety supervision of hot work operations using drone video monitoring, which is aimed at supervising hot work operations at outdoor construction sites and solving the problems of low detection accuracy, many monitoring blind spots and poor results.

[0008] The technical solution adopted by the present invention to solve its technical problem is:

[0009] A method for safety supervision of hot work using unmanned aerial vehicle video monitoring comprises the following steps:

[0010] Step 1: Extract on-site hot work information and bind the drone;

[0011] Step 2: Set up a safety supervision area for hot work;

[0012] Step 3: Obtain effective video data, including the placement status of on-site gas cylinders and fire extinguishers and the operation behavior of on-site workers in carrying gas cylinders, including the following steps:

[0013] S1. Calculate the camera's field of view

[0014] Horizontal field of view:

[0015] ω 水平 =2*arctan(w / 2f)

[0016] Vertical field of view:

[0017] ω 垂直 =2*arctan(h / 2f)

[0018] Where f is the focal length of the camera, w is the width of the sensor size, and h is the height of the sensor size;

[0019] S2. Calculate the flight altitude ranges H1 and H2 of the drone based on the hot work safety supervision area, the field of view angle range of the drone video monitoring, and the requirement that the detection object screen accounts for 1%;

[0020]

[0021] In the formula, L is the length of the longest side of the gas cylinder or fire extinguisher on site, ω0 is the maximum value of the horizontal field of view and the vertical field of view of the camera; the ground height is defined as 0, and Z is the height of the hot work point from the ground;

[0022] S3. According to the size relationship between H1 and H2, the coordinate parameter relationship of the camera mounted on the drone is calculated as follows;

[0023] If H2 is greater than H1, the following coordinate parameter relationship is used:

[0024]

[0025] In the formula, x, y, z are the coordinate parameters of the drone camera; the ground height is defined as 0, H0 is the height of the camera from the ground; Δh is the height of the camera from the working surface;

[0026] If H2 is less than H1, adjust the camera's focal length to ensure that the detection object occupies 1% of the screen, and use the following coordinate parameter relationship:

[0027]

[0028] In the formula, x, y, z are the coordinate parameters of the drone camera; Δh is the height of the camera from the working surface; f' is the adjusted focal length; the related ω' is the field of view angle after the focal length is adjusted; the ground height is defined as 0, and H1' is the height of the camera from the ground;

[0029] S4. Calculate the coordinate position range of the drone to obtain valid data, as follows:

[0030]

[0031] In the formula, x, y, z are the coordinate parameters of the drone camera; X and Y are the coordinates of the hot work point; d is the error range of the drone hovering accuracy, d = 0.5m;

[0032] S5. When the three-dimensional coordinates of the drone are within the effective range, the images captured by the camera are transmitted back for analysis and processing; otherwise, the data captured by the camera is not processed;

[0033] S6. Calculate the drone's endurance, set a low battery alarm, and perform multiple drone relays:

[0034] C min =P 飞行 *t+P 悬停 *t

[0035] In the formula, C min The battery power limit of the drone is low; 飞行 is the flight power of the UAV; P 悬停 is the hovering power of the UAV; t is the flight time required for the UAV to fly from the hangar to the current position coordinate point above the operation point;

[0036] Step 4: Video image detection of hot work safety supervision;

[0037] Step 5: Determine the actual safety situation of hot work and issue risk warnings.

[0038] The step one comprises:

[0039] A1. Through the risk operation announcements uploaded by each unit every day, the operation permit number is bound to the on-site operation ticket;

[0040] A2. After successful binding, the safety supervision system automatically extracts the "operation permit number" of the operation permit, and then compares it with the "today's operation content" in the risk operation announcement under the operation area to extract the key information of the hot work, including "hot work time", "hot work location", "hot work level" and "hot work type";

[0041] A3. On-site personnel use the distributed handheld explosion-proof tablet computer to scan the QR code to bind the drone equipment number with the operation permit on the operation permit management platform.

[0042] The second step comprises:

[0043] B1. Filter out the hot work tasks before the current date according to the extracted hot work time information;

[0044] B2. According to the extracted hot work location information, set the hot work point spatial coordinates to (X, Y, Z);

[0045] B3. According to the level and type of hot work, and referring to the standard management specifications, and consistent with the combustible leakage and exposure range and gas detection range, the hot work impact range radius R is set to 15 meters;

[0046] B4. The circular area with the spatial coordinates (X, Y, Z) of the hot work point as the center and R as the radius is the hot work safety supervision area.

[0047] In step S2, the detection objects include open flames, flammable gas bottles and flammable liquid bottles.

[0048] The step five comprises:

[0049] C1. Preset the safe distance between different detection objects and the starting point of the fire, as well as the safe distance between different detection objects;

[0050] C2. Determine the distance between different types of detection objects and the hot spot, as well as the distance between different detection objects;

[0051] C3. Issue danger warnings when unsafe behaviors of people and unsafe conditions of objects are detected, and feedback is given to the central control supervision and management personnel. Through drone calls and on-site intervention, unsafe behaviors of people and unsafe conditions of objects can be stopped in time.

[0052] The step C2 comprises:

[0053] H1, calculate the horizontal coverage length L1 of the camera

[0054]

[0055] H2, calculate the length L2 that the camera can cover vertically

[0056]

[0057] The rectangle with L1 as the length and L2 as the width is the monitoring area of ​​the camera;

[0058] H3, based on the camera's horizontal coverage length L1, vertical coverage length L2 and image size (w j 、h j ), get the actual size of the image;

[0059] H4, use drones to determine the distance between different detection objects and the fire point;

[0060] H5. Determine the difference between the actual distance of different detection objects from the ignition point and the preset safety distance, and the difference between the actual distance between different detection objects and the preset safety distance: if the distance is greater than the preset safety distance, the detection object is safely placed and no danger warning is issued; if the distance is greater than the preset safety distance, the detection object is unsafely placed and a danger warning is issued;

[0061] H6. Based on the image intelligent detection algorithm, determine whether there are any unsafe behaviors of people or unsafe conditions of objects at the scene; if there are no unsafe behaviors of people or unsafe conditions of objects, no danger alarm will be issued; if there are one or more unsafe behaviors of people or unsafe conditions of objects, danger alarms will be issued according to the type.

[0062] The step H3 comprises:

[0063]

[0064] The step H4 comprises:

[0065] S1. Use the tilt sensor on the drone to keep the drone camera perpendicular to the ground.

[0066] S2. According to the pixel position of the center of the hot spot detection frame and the pixel position information of the center of other detection object frames, the distance D1 of different detection objects from the hot spot is determined:

[0067]

[0068] S3. Determine the distance D2 between different detection objects based on the pixel position information of the center point of the detection frame between the detection objects:

[0069]

[0070] The unsafe behaviors mentioned include open smoke, open flame, not wearing work clothes, not wearing safety helmets, vehicle intrusion, personnel intrusion, single or double person inspection, smoking, making phone calls and falling.

[0071] The unsafe conditions of the objects include stepping on gas cylinders, carrying gas cylinders on shoulders, no fire extinguishers at the hot work site, and bottles and cans placed upside down at the hot work site.

[0072] The beneficial effects of the present invention are:

[0073] 1. By utilizing drone video surveillance, the problems of low accuracy, many blind spots and poor effect in hot work detection under on-site environment can be effectively solved. By calculating whether the actual distance of the detection object from the hot work point reaches the preset safety distance, adding a shouting module and a fire extinguishing module, loading common image intelligent detection algorithms such as open smoke, open flame, not wearing work clothes, not wearing safety helmets, vehicle intrusion, personnel intrusion, single or double inspection, smoking, making phone calls, and falling, as well as special image intelligent detection algorithms related to hot work such as stepping on gas cylinders, carrying gas cylinders on shoulders, no fire extinguishers at the hot work site, and bottles and cans placed upside down at the hot work site, the on-site drone shouting can promptly stop unsafe behaviors of people and eliminate unsafe conditions of objects.

[0074] 2. Through the fire-fighting module mounted on the on-site drone, the possibility of small sparks turning into fire accidents can be avoided in time, the safety supervision methods can be enriched, the safety supervision content can be expanded, and the safety supervision effect can be optimized, which has very good practical application effects and promotion and application significance.

[0075] 3. The length of the gas cylinder or fire extinguisher on site is set in the formula to ensure that the screen occupancy rate of the detection object is required to be 1%, thereby ensuring the recognition accuracy of the corresponding intelligent detection algorithms such as stepping on gas cylinders, carrying gas cylinders on shoulders, no fire extinguishers at the hot work site, and bottles and cans placed upside down at the hot work site. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 It is a flow chart of the present invention.

[0077] Figure 2 It is a schematic diagram of the on-site deployment location of drones, image acquisition methods, and distance measurement principles. DETAILED DESCRIPTION

[0078] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0079] Example 1

[0080] like Figures 1 to 2 As shown, a method for safety supervision of hot work using drone video monitoring includes the following steps:

[0081] Step 1: Extract on-site hot work information and bind the drone;

[0082] Step 2: Set up a safety supervision area for hot work;

[0083] Step 3: Obtain effective video data, including the placement status of on-site gas cylinders and fire extinguishers and the operation behavior of on-site workers in carrying gas cylinders, including the following steps:

[0084] S1. Calculate the camera's field of view

[0085] Horizontal field of view:

[0086] ω 水平 =2*arctan(w / 2f)

[0087] Vertical field of view:

[0088] ω 垂直 =2*arctan(h / 2f)

[0089] Where f is the focal length of the camera, w is the width of the sensor size, and h is the height of the sensor size;

[0090] S2. Calculate the flight altitude ranges H1 and H2 of the drone based on the hot work safety supervision area, the field of view angle range of the drone video monitoring, and the requirement that the detection object screen accounts for 1%;

[0091]

[0092] In the formula, L is the length of the longest side of the gas cylinder or fire extinguisher on site, ω0 is the maximum value of the horizontal field of view and the vertical field of view of the camera; the ground height is defined as 0, and Z is the height of the hot work point from the ground;

[0093] S3. According to the size relationship between H1 and H2, the coordinate parameter relationship of the camera mounted on the drone is calculated as follows;

[0094] If H2 is greater than H1, the following coordinate parameter relationship is used:

[0095]

[0096] In the formula, x, y, z are the coordinate parameters of the drone camera; the ground height is defined as 0, H0 is the height of the camera from the ground; Δh is the height of the camera from the working surface;

[0097] If H2 is less than H1, adjust the camera's focal length to ensure that the detection object occupies 1% of the screen, and use the following coordinate parameter relationship:

[0098]

[0099] In the formula, x, y, z are the coordinate parameters of the drone camera; Δh is the height of the camera from the working surface; f' is the adjusted focal length; the related ω' is the field of view angle after the focal length is adjusted; the ground height is defined as 0, and H1' is the height of the camera from the ground;

[0100] S4. Calculate the coordinate position range of the drone to obtain valid data, as follows:

[0101]

[0102] In the formula, x, y, z are the coordinate parameters of the drone camera; X and Y are the coordinates of the hot work point; d is the error range of the drone hovering accuracy, d = 0.5m;

[0103] S5. When the three-dimensional coordinates of the drone are within the effective range, the images captured by the camera are transmitted back for analysis and processing; otherwise, the data captured by the camera is not processed;

[0104] S6. Calculate the drone's endurance, set a low battery alarm, and perform multiple drone relays:

[0105] Cmin =P 飞行 *t+P 悬停 *t

[0106] In the formula, C min The battery power limit of the drone is low; 飞行 is the flight power of the UAV; P 悬停 is the hovering power of the UAV; t is the flight time required for the UAV to fly from the hangar to the current position coordinate point above the operation point;

[0107] Step 4: Video image detection of hot work safety supervision;

[0108] It includes setting the category of the object to be detected in the hot work operation, and quickly analyzing and detecting the features of the returned valid data according to the preset image intelligent detection model. The input is the valid data sent back by the camera, and the output is an image with the coordinate information of the detection frame. The output result includes the category of the detected object, the pixel coordinate information of the center point of the detection frame (x i ,y i )

[0109] Step 5: Determine the actual safety situation of hot work and issue risk warnings.

[0110] The step one comprises:

[0111] A1. Through the risk operation announcements uploaded by each unit every day, the operation permit number is bound to the on-site operation ticket;

[0112] A2. After successful binding, the safety supervision system automatically extracts the "operation permit number" of the operation permit, and then compares it with the "today's operation content" in the risk operation announcement under the operation area to extract the key information of the hot work, including "hot work time", "hot work location", "hot work level" and "hot work type";

[0113] A3. On-site personnel use the distributed handheld explosion-proof tablet computer to scan the QR code to bind the drone equipment number with the operation permit on the operation permit management platform.

[0114] The second step comprises:

[0115] B1. Filter out the hot work tasks before the current date according to the extracted hot work time information;

[0116] B2. According to the extracted hot work location information, set the hot work point spatial coordinates to (X, Y, Z);

[0117] B3. According to the level and type of hot work, and referring to the standard management specifications, the hot work impact range radius R is set to 15 meters, which is consistent with the combustible leakage and exposure range and the gas detection range. The standard management specifications are the China National Petroleum Corporation's corporate standard "Hot Work Safety Management Specification" (Q / SY 1241-2009) and "Southwest Oil and Gas Field Company Hot Work Safety Management Specification" (Southwest Division Quality

[2016] No. 31), which is consistent with the combustible leakage and exposure range and the gas detection range;

[0118] B4. The circular area with the spatial coordinates (X, Y, Z) of the hot work point as the center and R as the radius is the hot work safety supervision area.

[0119] In step S2, the detection objects include open flames, flammable gas bottles and flammable liquid bottles.

[0120] The step five comprises:

[0121] C1. Preset the safe distance between different inspection objects and the hot spot, as well as the safe distance between different inspection objects. For example, when using gas welding (cutting) for hot work, the distance between oxygen cylinders and acetylene cylinders shall not be less than 5 meters, and the distance between the two and the hot work site shall not be less than 10 meters; combustible materials within 5 meters around the hot work area should be removed or isolated with flame retardant materials; no other combustibles are allowed to leak or be exposed within a radius of 15 meters;

[0122] C2. Determine the distance between different types of detection objects and the hot spot, as well as the distance between different detection objects;

[0123] C3. Issue danger warnings for unsafe behaviors of people and unsafe conditions of objects, and feedback to the central control supervision and management personnel. Through drone calls and on-site intervention, unsafe behaviors of people and unsafe conditions of objects can be stopped in time. If there is an alarm for open smoke or open flames, the mounted fire extinguishing module can be used to timely avoid the possibility of small sparks turning into fire accidents.

[0124] The step C2 comprises:

[0125] H1, calculate the horizontal coverage length L1 of the camera

[0126]

[0127] H2, calculate the length L2 that the camera can cover vertically

[0128]

[0129] The rectangle with L1 as the length and L2 as the width is the monitoring area of ​​the camera;

[0130] H3, based on the camera's horizontal coverage length L1, vertical coverage length L2 and image size (w j 、h j ), get the actual size of the image;

[0131] H4, use drones to determine the distance between different detection objects and the fire point;

[0132] H5. Determine the difference between the actual distance of different detection objects from the ignition point and the preset safety distance, and the difference between the actual distance between different detection objects and the preset safety distance: if the distance is greater than the preset safety distance, the detection object is safely placed and no danger warning is issued; if the distance is greater than the preset safety distance, the detection object is unsafely placed and a danger warning is issued;

[0133] H6. Based on the image intelligent detection algorithm, determine whether there are any unsafe behaviors of people or unsafe conditions of objects at the scene; if there are no unsafe behaviors of people or unsafe conditions of objects, no danger alarm will be issued; if there are one or more unsafe behaviors of people or unsafe conditions of objects, danger alarms will be issued according to the type.

[0134] The step H3 comprises:

[0135]

[0136] The step H4 comprises:

[0137] S1. Use the tilt sensor on the drone to keep the drone camera perpendicular to the ground.

[0138] S2. According to the pixel position of the center of the hot spot detection frame and the pixel position information of the center of other detection object frames, the distance D1 of different detection objects from the hot spot is determined:

[0139]

[0140] S3. Determine the distance D2 between different detection objects based on the pixel position information of the center point of the detection frame between the detection objects:

[0141]

[0142] The unsafe behaviors mentioned include open smoke, open flame, not wearing work clothes, not wearing safety helmets, vehicle intrusion, personnel intrusion, single or double person inspection, smoking, making phone calls and falling.

[0143] The unsafe conditions of the objects include stepping on gas cylinders, carrying gas cylinders on shoulders, no fire extinguishers at the hot work site, and bottles and cans placed upside down at the hot work site.

[0144] By utilizing drone video surveillance, the problems of low accuracy in detecting hot work operations, many blind spots in monitoring, and poor results in on-site environments can be effectively solved. By measuring whether the actual distance of the detection object from the hot work point reaches the preset safety distance, adding a shouting module and a fire extinguishing module, loading common image intelligent detection algorithms such as open smoke, open flame, not wearing work clothes, not wearing safety helmets, vehicle intrusion, personnel intrusion, single or double inspections, smoking, making phone calls, and falling, as well as special image intelligent detection algorithms related to hot work operations such as stepping on gas cylinders, carrying gas cylinders on shoulders, no fire extinguishers at the hot work site, and bottles and cans placed upside down at the hot work site, the on-site drone shouting can promptly stop unsafe behaviors of people and eliminate unsafe conditions of objects.

[0145] By using the fire-fighting module mounted on the on-site drone, the possibility of small sparks turning into fire accidents can be avoided in a timely manner, the safety supervision measures can be enriched, the safety supervision content can be expanded, and the safety supervision effect can be optimized. It has very good practical application effects and promotion and application significance.

[0146] The purpose of setting the length of the gas cylinder or fire extinguisher on site in the formula is to ensure that the screen occupancy rate of the detection object is required to be 1%, thereby ensuring the recognition accuracy of the corresponding intelligent detection algorithms such as stepping on the gas cylinder, carrying the gas cylinder on the shoulder, no fire extinguisher at the hot work site, and bottles and cans placed upside down at the hot work site.

[0147] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various changes and variations. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for safety supervision of hot work using drone video monitoring, characterized in that: The following steps are involved: Step 1: Extract on-site hot work information and bind the drone; Step 2: Set up a safety supervision area for hot work; Step 3: Obtain effective video data, including the placement status of on-site gas cylinders and fire extinguishers and the operation behavior of on-site workers in carrying gas cylinders, including the following steps: S1. Calculate the camera's field of view Horizontal field of view: ω 水平 =2*arctan(w / 2f) Vertical field of view: ω 垂直 =2*arctan(h / 2f) Where f is the focal length of the camera, w is the width of the sensor size, and h is the height of the sensor size; S2. Calculate the flight altitude ranges H1 and H2 of the drone based on the hot work safety supervision area, the field of view angle range of the drone video monitoring, and the requirement that the detection object screen accounts for 1%; In the formula, L is the length of the longest side of the gas cylinder or fire extinguisher on site, ω0 is the maximum value of the horizontal field of view and the vertical field of view of the camera; the ground height is defined as 0, and Z is the height of the hot work point from the ground; S3. According to the size relationship between H1 and H2, the coordinate parameter relationship of the camera mounted on the drone is calculated as follows; If H2 is greater than H1, the following coordinate parameter relationship is used: In the formula, x, y, z are the coordinate parameters of the drone camera; the ground height is defined as 0, H0 is the height of the camera from the ground; Δh is the height of the camera from the working surface; If H2 is less than H1, adjust the camera's focal length to ensure that the detection object occupies 1% of the screen, and use the following coordinate parameter relationship: In the formula, x, y, z are the coordinate parameters of the drone camera; Δh is the height of the camera from the working surface; f' is the adjusted focal length; the related ω' is the field of view angle after the focal length is adjusted; the ground height is defined as 0, and H1' is the height of the camera from the ground; S4. Calculate the coordinate position range of the drone to obtain valid data, as follows: In the formula, x, y, z are the coordinate parameters of the drone camera; X and Y are the coordinates of the hot work point; d is the error range of the drone hovering accuracy, d = 0.5m; S5. When the three-dimensional coordinates of the drone are within the effective range, the images captured by the camera are transmitted back for analysis and processing; Otherwise, the data captured by the camera will not be processed; S6. Calculate the drone's endurance, set a low battery alarm, and perform multiple drone relays: C min =P 飞行 *t+P 悬停 *t In the formula, C min The battery power limit of the drone is low; 飞行 P is the flight power of the UAV; 悬停 is the hovering power of the UAV; t is the flight time required for the UAV to fly from the hangar to the current position coordinate point above the operation point; Step 4: Video image detection of hot work safety supervision; Step 5: Determine the actual safety situation of hot work and issue risk warnings.

2. A method for safety supervision of hot work using drone video monitoring as claimed in claim 1, characterized in that: The step one comprises: A1. Through the risk operation announcements uploaded by each unit every day, the operation permit number is bound to the on-site operation ticket; A2. After successful binding, the safety supervision system automatically extracts the "operation permit number" of the operation permit, and then compares it with the "today's operation content" in the risk operation announcement under the operation area to extract the key information of the hot work, including "hot work time", "hot work location", "hot work level" and "hot work type"; A3. On-site personnel use the distributed handheld explosion-proof tablet computer to scan the QR code to bind the drone equipment number with the operation permit on the operation permit management platform.

3. A method for safety supervision of hot work using drone video monitoring as claimed in claim 1, characterized in that: The second step comprises: B1. Filter out the hot work tasks before the current date according to the extracted hot work time information; B2. According to the extracted hot work location information, set the hot work point spatial coordinates to (X, Y, Z); B3. According to the level and type of hot work, and referring to the standard management specifications, and consistent with the combustible leakage and exposure range and gas detection range, the hot work impact range radius R is set to 15 meters; B4. The circular area with the spatial coordinates (X, Y, Z) of the hot work point as the center and R as the radius is the hot work safety supervision area.

4. A method for safety supervision of hot work using drone video monitoring as claimed in claim 1, characterized in that: In step S2, the detection objects include open flames, flammable gas bottles and flammable liquid bottles.

5. The method for safety supervision of hot work using drone video monitoring as claimed in claim 1, characterized in that: The step five comprises: C1. Preset the safe distance between different detection objects and the starting point of the fire, as well as the safe distance between different detection objects; C2. Determine the distance between different types of detection objects and the hot spot, as well as the distance between different detection objects; C3. Issue danger warnings when unsafe behaviors of people and unsafe conditions of objects are detected, and feedback is given to the central control supervision and management personnel. Through drone calls and on-site intervention, unsafe behaviors of people and unsafe conditions of objects can be stopped in time.

6. A method for safety supervision of hot work using drone video monitoring as claimed in claim 5, characterized in that: The step C2 comprises: H1, calculate the horizontal coverage length L1 of the camera or H2, calculate the length L2 that the camera can cover vertically or The rectangle with L1 as the length and L2 as the width is the monitoring area of ​​the camera; H3, according to the horizontal coverage length L1 of the camera, the vertical coverage length L2 and the image size (wj, hj), the actual size of the image is obtained; H4, use drones to determine the distance between different detection objects and the fire point; H5. Determine the difference between the actual distance of different detection objects from the ignition point and the preset safety distance, and the difference between the actual distance between different detection objects and the preset safety distance: if the distance is greater than the preset safety distance, the detection object is safely placed and no danger warning is issued; if the distance is greater than the preset safety distance, the detection object is unsafely placed and a danger warning is issued; H6. Based on the image intelligent detection algorithm, determine whether there are any unsafe behaviors of people or unsafe conditions of objects at the scene; if there are no unsafe behaviors of people or unsafe conditions of objects, no danger alarm will be issued; if there are one or more unsafe behaviors of people or unsafe conditions of objects, danger alarms will be issued according to the type.

7. A method for safety supervision of hot work using drone video monitoring as claimed in claim 6, characterized in that: The step H3 comprises:

8. A method for safety supervision of hot work using drone video monitoring as claimed in claim 6, characterized in that: The step H4 comprises: S1. Use the tilt sensor on the drone to keep the drone camera perpendicular to the ground. S2. According to the pixel position of the center of the hot spot detection frame and the pixel position information of the center of other detection object frames, the distance D1 of different detection objects from the hot spot is determined: S3. Determine the distance D2 between different detection objects based on the pixel position information of the center point of the detection frame between the detection objects:

9. A method for safety supervision of hot work using drone video monitoring as claimed in claim 6, characterized in that: The unsafe behaviors mentioned include open smoke, open flame, not wearing work clothes, not wearing safety helmets, vehicle intrusion, personnel intrusion, single or double person inspection, smoking, making phone calls and falling.

10. A method for safety supervision of hot work using drone video monitoring as claimed in claim 6, characterized in that: The unsafe conditions of the objects include stepping on gas cylinders, carrying gas cylinders on shoulders, no fire extinguishers at the hot work site, and bottles and cans placed upside down at the hot work site.

Images