Flame detection method, system, device and storage medium

Abstract

The flame detection method, system, device and storage medium provided by this application mention that the method can determine whether a flame has appeared by detecting the scene with ultraviolet light, and can determine the situation and location of the flame in advance, and can detect it before it develops into a large-scale situation. Multiple detections are performed on the location of the flame, and the results of multiple detections are combined to determine whether the flame has evolved into a flame. Each technology detects the fire from a different angle, complementing and verifying each other, reducing the misjudgment and missed judgment that may be caused by a single technology, greatly improving the accuracy and reliability of fire detection, and quickly responding to the occurrence of flames, allowing users to take corresponding measures for the flame more quickly. Multiple detection methods are used to detect flames, comprehensively considering the multi-dimensional characteristics of the flame such as the position, spectrum, and temperature, to more comprehensively understand the true situation of the flame and avoid misjudgments that may be caused by single feature judgment.

Description

Technical Field

[0001] The present application relates to the field of fire detection technology, and in particular to flame detection methods, systems, devices, and storage media. Background Art

[0002] How to detect fire in time and send out alarm signals has become a research focus.

[0003] Currently, fire detection mainly includes smoke sensors, image-based smoke and fire detection algorithms, and thermal imaging high-temperature detection. Smoke sensors primarily detect the presence of smoke in the environment and issue an alarm. Existing smoke sensors rely on the generation of smoke, while electrical fires, chemical fires, or dry fires may produce less or less noticeable smoke. Image-based smoke and fire detection algorithms use AI models to detect the presence of flame features in images to determine the presence of flames. Image-based smoke and fire detection algorithms are easily interfered with by similar flame features. Thermal imaging high-temperature detection determines the presence of a fire by detecting whether there are high-temperature areas exceeding a preset temperature in the scene. Thermal imaging high-temperature detection is easily interfered with by high temperatures caused by flames, such as sunlight reflection.

[0004] Existing fire detection technologies, such as smoke sensors, have problems such as long fire detection time and delayed response when smoke is generated in some situations where there is little or no obvious smoke. Summary of the Invention

[0005] The technical problem to be solved by this application is that the existing fire detection technology is susceptible to misjudgment in complex environments due to the influence of multiple factors, and the detection process has problems such as long time consumption and delayed response.

[0006] In order to solve the above problems, in order to solve the above technical problems or at least partially solve the above technical problems, the present application provides a flame detection method, system, device and storage medium.

[0007] In a first aspect, the present invention discloses a flame detection method, comprising:

[0008] Performing real-time ultraviolet detection on the environment to obtain real-time ultraviolet pulse data within a preset time period. When the real-time ultraviolet pulse data is greater than a preset ultraviolet pulse threshold data, it is determined that there is a flame in the current environment; the real-time ultraviolet pulse data within the preset time period includes the cumulative number of pulses and the standard deviation within the preset time period, and the preset ultraviolet pulse threshold data is calculated based on the cumulative number of pulses and the standard deviation within the preset time period;

[0009] The flame area is acquired based on the real-time video image of the scene, and multiple detection processes are performed on the flame area. The multiple detection and processing results are combined with the preset threshold to make a judgment. When the result exceeds the preset threshold, it is judged that the flame has evolved into a flame. The multiple detection processes include flame position positioning processing, brightness detection processing, and thermal imaging temperature detection processing of the flame area.

[0010] Perform brightness detection and thermal imaging temperature detection on the flame area to obtain the brightness information and temperature information of the flame area;

[0011] The flame area brightness information and flame area temperature information are integrated with time, sorted according to the time domain, and flame brightness data and flame temperature data are obtained;

[0012] Calculate the corresponding inverse numbers for the flame brightness data and flame temperature data within a preset period of time, and combine them into a brightness inverse series and a temperature inverse series;

[0013] The flame changes are judged based on the brightness inverse series and the temperature inverse series.

[0014] Preferably, the method of acquiring the flame area based on the on-site real-time video image, performing multiple detection processes on the flame area, combining the multiple detection process results with a preset threshold to make a judgment, and judging that the flame has evolved into a flame when the result exceeds the preset threshold, specifically includes the following steps:

[0015] Acquire real-time video images from the scene, perform fire detection and processing frame by frame based on the real-time video images, and obtain the flame position characteristics and continuous images with the same flame;

[0016] The flame position is obtained based on the flame position characteristics, and the brightness of the light emitted by the flame is detected to obtain the brightness information of the flame area. When the brightness of the flame area exceeds the preset threshold, it is determined that the flame has evolved into a flame;

[0017] Thermal imaging detection and processing are performed on the area where the flame is located to obtain temperature information of the flame area. When the temperature of the flame area exceeds the preset threshold, it is determined that the flame has evolved into a flame.

[0018] Preferably, the method of acquiring a live real-time video image and performing fire detection processing frame by frame based on the live real-time video image to obtain a flame position feature and continuous images with the same flame specifically includes the following steps:

[0019] Acquire live video images, perform fire and smoke detection on each frame of the live video images, mark images with flames, and obtain images with flame features and the location features of the flames in the images;

[0020] The flame position difference is obtained based on the images with flame features that are consecutive in time sequence. If the flame position difference is less than a preset threshold and appears consecutively in a preset number of images, it is determined that the flames in the corresponding preset number of images with flame features that are consecutive in time sequence are the same flame.

[0021] Preferably, the method of obtaining the flame position according to the flame position feature, performing brightness detection based on the brightness of the light emitted by the flame to obtain the brightness information of the flame area, and determining that the flame has evolved into a flame when the brightness of the flame area exceeds a preset threshold, specifically includes the following steps:

[0022] Based on the flame position characteristics and the corresponding real-time video image, the flame area on site is obtained, and the brightness detection processing and average value calculation processing are performed on the flame area to obtain the average brightness of the first flame area;

[0023] After adding the filter, the flame area is subjected to brightness detection processing to obtain the average brightness of the second flame area; the filter filters out light with a wavelength greater than 750nm;

[0024] The infrared ratio is calculated by combining the average brightness of the first flame area and the average brightness of the second flame area. When the infrared ratio exceeds the preset threshold, it is determined that the flame has evolved into a flame.

[0025] Preferably, the method of performing thermal imaging detection on the flame area to obtain temperature information of the flame area and determining that the flame has evolved into a flame when the temperature of the flame area exceeds a preset threshold specifically includes the following steps:

[0026] Based on the real-time video images of the scene, the flame area is detected and processed in real time to obtain the visible light flame area information, including the coordinate information and area range of the visible light flame area; the flame area detection process is processed by the target detection algorithm;

[0027] For the on-site real-time use of infrared thermal imaging cameras to obtain real-time thermal imaging images, the flame thermal imaging area is located in the real-time thermal imaging image according to the coordinate information of the visible light flame area within the regional range, and the flame thermal imaging corresponding area is obtained;

[0028] Traverse the corresponding area of ​​the thermal imaging, calculate the maximum grayscale value in the area, and obtain the temperature value of the flame area based on the obtained thermal imaging grayscale value and thermal imaging information; the thermal imaging information includes ambient temperature, preset distance, and detection target emissivity;

[0029] When the temperature value of the flame area is greater than the preset temperature threshold, it is determined that the flame has evolved into a flame.

[0030] Preferably, when a flame is detected in the environment, an open fire warning is issued;

[0031] The real-time video image calculates the number of flame areas in the image according to the time domain. When the number of flame areas increases, it is judged that the flames in the environment have evolved into flames and the area is getting larger, and a fire warning is issued.

[0032] Preferably, judging the flame change according to the brightness inverse sequence and the temperature inverse sequence comprises the following steps:

[0033] Observe the increase and decrease of the values ​​in the brightness reverse sequence. If the reverse sequence number gradually increases, it means that the fluctuation of the flame brightness data has increased, which may mean that the flame is burning unstably or is affected by other factors. If the reverse sequence number gradually decreases, it may mean that the flame brightness is tending to be stable.

[0034] For the temperature reverse sequence, if the reverse sequence number increases, the fluctuation of the flame temperature becomes larger and the combustion state becomes unstable; if the reverse sequence number decreases, it means that the flame temperature gradually stabilizes;

[0035] If the brightness inverse series and the temperature inverse series increase at the same time, it means that the flame is in a state of drastic change, the combustion is unstable and there is a tendency for the fire to spread; if both decrease at the same time, it means that the flame is gradually stabilizing.

[0036] In a second aspect, the present invention discloses a flame detection system, which includes the above-mentioned flame detection method.

[0037] In a third aspect, the present invention discloses a computer device comprising a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0038] Memory for storing computer programs;

[0039] The processor is configured to implement the steps of the above method when executing the program stored in the memory.

[0040] In a fourth aspect, the present invention discloses a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.

[0041] The above technical solution provided by this application has the following advantages compared with the existing technology:

[0042] The flame detection method, system, device and storage medium provided in the present application mention that by detecting the scene with ultraviolet light, it is possible to determine whether a flame has appeared, and the situation and location of the flame can be determined in advance, and it can be detected before it develops into a large-scale situation. Multiple tests are performed on the location of the flame, and the results of multiple tests are combined to determine whether the flame has evolved into a flame. Each technology detects the fire from a different angle, complements and verifies each other, reduces the misjudgment and missed judgment that may be caused by a single technology, greatly improves the accuracy and reliability of fire detection, and quickly responds to the generation of flames, allowing users to take corresponding measures to the flame more quickly.

[0043] Furthermore, multiple detection methods are used to detect flames, taking into account the multi-dimensional characteristics of the flame, such as position, spectrum, temperature, etc., to gain a more comprehensive understanding of the true situation of the flame and avoid misjudgment that may be caused by judging by a single feature.

[0044] Furthermore, in addition to determining whether there is a flame at the scene, it can also provide information on the flame's location, temperature, and spectral characteristics. In some scenarios where fire needs to be judged, it can provide firefighters and related personnel with more detailed and comprehensive fire conditions, helping them to formulate more effective fire-fighting strategies and rescue plans. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0047] Figure 1 A flame detection method provided in this application Figure 1 ;

[0048] Figure 2 A flame detection method provided in this application Figure 2 ;

[0049] Figure 3 This is a schematic diagram of a specific flow chart of step S2 of a flame detection method provided in this application. DETAILED DESCRIPTION

[0050] To make the objectives, technical solutions, and advantages of this application more clear, the technical solutions of this application will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are part of the embodiments of this application, not all of them. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this application.

[0051] First, see Figure 1-3 The present invention discloses a flame detection method, which comprises:

[0052] Step S1: Performing real-time ultraviolet detection on the environment to obtain real-time ultraviolet pulse data within a preset time period. When the real-time ultraviolet pulse data is greater than a preset ultraviolet pulse threshold data, it is determined that there is a flame in the current environment; the real-time ultraviolet pulse data within the preset time period includes the cumulative number of pulses and the standard deviation within the preset time period, and the preset ultraviolet pulse threshold data is calculated based on the cumulative number of pulses and the standard deviation within the preset time period;

[0053] Step S2: The flame area is acquired based on the real-time video image of the scene, and multiple detection processes are performed on the flame area. The multiple detection and processing results are combined with the preset threshold to make a judgment. When the result exceeds the preset threshold, it is judged that the flame has evolved into a flame; the multiple detection processes include ultraviolet detection, flame position positioning, brightness detection, and thermal imaging temperature detection of the flame area.

[0054] Specifically, in step S1, the UV detection component records and analyzes data when a flame begins to burn, generating a large amount of UV light in the air and the number of pulses changes. When specific conditions are met, it can detect the presence of sparks or open flames (UV), providing early warning of fires. This alarm is issued before a fire develops into a visible, large-scale flame, buying valuable time for evacuation and activation of fire-fighting equipment, helping to control the spread of the fire and minimize fire damage.

[0055] Specifically, in step S2, flame position detection can determine the flame position and whether the flames in multiple connected images are the same flame, thereby determining whether the flame has evolved into a flame. Infrared light detection determines whether the flame has evolved into a flame based on the brightness value, and thermal imaging temperature detection obtains temperature changes to determine whether the flame has evolved into a flame. It combines multiple technologies such as flame position detection, infrared light detection, and thermal imaging temperature detection. Each technology detects fire from different angles, complements and verifies each other, reduces misjudgments and missed judgments that may be caused by a single technology, and greatly improves the accuracy and reliability of fire detection.

[0056] It can be understood that by using ultraviolet light to detect the presence of flames at the scene, it is possible to determine the presence and location of flames in advance, and to detect them before they develop into a large-scale situation. Multiple tests are performed on the location of the flames, and the results of multiple tests are combined to determine whether the flames have evolved into flames. Each technology detects fires from different angles, complementing and verifying each other, reducing the misjudgments and missed judgments that may occur with a single technology, greatly improving the accuracy and reliability of fire detection, and quickly responding to the occurrence of flames, allowing users to take appropriate measures to deal with the flames more quickly. At different stages of a fire, different technologies can play their unique advantages. Ultraviolet light detection can capture signals in a timely manner when ultraviolet rays are generated by flames in the early stages of a fire; flame location positioning identifies flames by analyzing the shape and position changes of the flames in the image; infrared light detection and thermal imaging temperature detection further confirm the flames from the aspects of brightness and temperature characteristics, enabling the entire system to more comprehensively and accurately judge the fire situation, providing a reliable basis for timely firefighting measures.

[0057] Furthermore, multiple detection methods are used to detect flames, taking into account the multi-dimensional characteristics of the flame, such as position, spectrum, temperature, etc., to gain a more comprehensive understanding of the true situation of the flame and avoid misjudgment that may be caused by judging by a single feature.

[0058] Furthermore, in addition to determining whether there is a flame at the scene, it can also provide information on the flame's location, temperature, and spectral characteristics. In some scenarios where fire needs to be judged, it can provide firefighters and related personnel with more detailed and comprehensive fire conditions, helping them to formulate more effective fire-fighting strategies and rescue plans.

[0059] As an example, after step S1, when a flame is detected in the environment, an open flame warning is issued. Specifically, the ultraviolet detection component can record and analyze data when a flame begins to burn, a large amount of ultraviolet light is generated in the air, and the number of pulses changes. When specific conditions are met, a warning of the presence of an electric spark or open flame can be determined in advance, providing an early warning of a fire. This allows the alarm to be issued before the fire develops into a clearly visible large-scale flame, buying valuable time for personnel evacuation and the activation of fire-fighting equipment, helping to control the spread of the fire and reduce fire damage. In some scenarios, such as indoors and laboratories, small flames may appear due to electrical short circuits or improper operation. Initially, they can be considered controlled. However, if left unchecked, these small flames can escalate into flames, leading to fires. Because the flames are small in the early stages, some recognition methods are difficult to detect. They need to evolve into larger flames before they can be recognized. Therefore, it is necessary to identify the flames as they appear, alerting the user to take timely measures to extinguish them and prevent them from escalating into flames.

[0060] As an example, after step S2, the real-time video image is calculated based on the time domain to determine the number of flame regions within the image. When the number of flame regions increases, it is determined that the flames in the environment are evolving into flames, and a fire warning is issued. Specifically, a flame may not evolve directly from a small flame, but may evolve from multiple flames. Therefore, the number of flame regions can be obtained in the real-time video image, and the distance between the flame regions can be detected. Based on the changes in distance and number, it is determined whether the flames have evolved into flames. When the distance between the flame regions decreases, it can be considered that the flames are converging and have the potential to evolve into flames. As the number of flame regions increases, the likelihood of evolving into flames also increases. In some scenarios, such as household appliance fires or uncontrolled fireworks displays, detecting the number and distance of flame regions can better determine the possibility of flame generation. A warning signal can be issued before a flame appears, prompting the user to take appropriate measures to extinguish the flames in advance.

[0061] Step S2 includes:

[0062] Step S3: Perform brightness detection and thermal imaging temperature detection on the flame area to obtain flame area brightness information and flame area temperature information;

[0063] Step S4: Integrate the flame area brightness information and flame area temperature information with time, sort them according to the time domain, and obtain flame brightness data and flame temperature data;

[0064] Step S5: Calculate the corresponding inverse numbers for the flame brightness data and flame temperature data within a preset period of time, and combine them into a brightness inverse number series and a temperature inverse number series;

[0065] Step S6: judging the flame change according to the brightness inverse sequence and the temperature inverse sequence.

[0066] Specifically, in step S3, the brightness information and temperature information of the flame area are obtained according to the above step S1, wherein the brightness detection uses a visible light imaging device (such as a high-definition camera) to obtain an image of the flame area, and the thermal imaging temperature detection uses a thermal imaging device (such as an infrared thermal imager) to scan the flame area.

[0067] In step S4, each detected flame area brightness and temperature information is timestamped. Timestamps can be accurate to seconds or smaller, recording the specific time of each detection. Two data lists are created: one for storing the flame brightness information and its corresponding timestamp, and the other for storing the flame temperature information and its corresponding timestamp. These two data lists are sorted by timestamp, using common sorting algorithms (such as bubble sort or quick sort) to arrange the data in chronological order, resulting in chronologically ordered flame brightness and temperature data.

[0068] In step S5, a preset time period is determined, such as the past 10 minutes or another time period set according to actual needs. Data within the preset time period is selected from the time-sorted flame brightness data. For this set of brightness data, its inverse sequence number is calculated. A double loop can be used to traverse the brightness data sequence, counting the number of inverse pairs to obtain the inverse sequence number of the brightness data within the time period. Using the same method, data within the preset time period is selected from the time-sorted flame temperature data, and its inverse sequence number is calculated. The brightness inverse sequence numbers obtained from each calculation are arranged in chronological order to form a brightness inverse sequence. The temperature inverse sequence numbers obtained from each calculation are also arranged in chronological order to form a temperature inverse sequence.

[0069] In step S6, the changing trend of the brightness inverse series is analyzed. The increase or decrease in the values ​​in the brightness inverse series is observed. If the inverse series gradually increases, it indicates that the flame brightness data is fluctuating more, which may indicate that the flame is burning unstably or is affected by other factors. If the inverse series gradually decreases, it may indicate that the flame brightness is becoming stable. Furthermore, the changing trend of the temperature inverse series is analyzed. Similarly, for the temperature inverse series, an increase in the inverse series may indicate that the flame temperature is fluctuating more and the combustion state is unstable; if the inverse series decreases, it may indicate that the flame temperature is gradually stabilizing. Finally, a comprehensive judgment is made based on the brightness and temperature inverse series. For example, if the brightness and temperature inverse series increase simultaneously, it may indicate that the flame is in a state of rapid fluctuation, combustion is unstable, and there may be a tendency for the fire to spread. If both decrease simultaneously, it may indicate that the flame is gradually stabilizing, that is, combustion is maintaining stability. Based on the changes in the inverse series, different judgment rules can be set to determine the specific changes in the flame, such as whether there are signs of re-ignition or whether the fire is under control.

[0070] In scenarios such as home kitchens and laboratories where combustion experiments are being conducted, the flame needs to continue to burn and remain within a controllable range. It is also necessary to prevent the flame from growing larger and larger. Once it reaches a certain peak value, its state needs to be suppressed from continuing to change. By detecting the brightness and temperature inverse series, the flame change can be judged to ensure that the flame remains within a controllable range. When the flame is about to get out of control, an alarm, such as sound or light, can be issued in a timely manner to remind the user to stop the current operation in time, reduce the flame state, and avoid fire.

[0071] In addition, in some cases where it is necessary to keep the flame burning continuously and avoid the flame from going out, the reverse sequence can be used to judge the evolution of the current flame to avoid the flame from going out midway, such as garbage incineration and steel smelting. In such cases, if the flame goes out, it may cause insufficient combustion temperature, resulting in incomplete combustion or the production of additional substances. Therefore, it is necessary to ensure that the flame continues to burn. Then, by detecting the brightness and temperature of the flame area, the reverse sequence is obtained to judge the evolution of the flame. When it starts to get smaller, a corresponding signal can be issued to remind the user to make timely adjustments to keep the flame burning at the corresponding temperature.

[0072] In addition, according to the changes in the flame, the flame extinction situation can be identified to prevent the extinguished flame from reigniting. For example, in garbage disposal site combustion treatment, forest fires, etc., it is necessary to detect whether the flame is reigniting. Through the brightness, temperature and the corresponding inverse series, according to the monotonicity of the inverse series, it is judged whether the brightness and temperature meet the conditions for reignition. When the conditions are met, the user can be notified through a signal to take corresponding measures to extinguish the flame, reduce the temperature, and avoid reignition.

[0073] Step S1 specifically includes the following steps:

[0074] Step S11: Every 1 second, use a photoelectric sensor based on ultraviolet (UV) detection to count the number of pulses of ultraviolet rays in the air per unit time. When the number of pulses N in the i-th second is i Greater than the preset threshold N t Start recording data when

[0075] Step S12: When N i ≥N t When the number of pulses N is continuously recorded i , N i+1 ,...,N i+n , the statistics exceed the threshold N t The number of times is a, the recording time is n (n ≥ 2), the cumulative number of pulses N and its standard deviation σ are calculated, and the real-time ultraviolet pulse data within the preset time period is obtained;

[0076] Step S13: When n≥2 and a>0.8n, real-time judgment is performed. If u*n>N is satisfied at the same time t+1 , σ>0.2u, then it is judged that there is an electric spark or open flame (UV) warning, otherwise it is judged as an interference source, where N t+1 and 0.2u is the preset UV pulse threshold data.

[0077] The standard deviation is σ and the calculation method is as follows:

[0078] , ;

[0079] u is the arithmetic mean of n pulses.

[0080] Specifically, if the number of pulses detected by the sensor per second exceeds a preset threshold, the UV pulse data is recorded, along with the duration. The cumulative number of UV pulses within the preset time period and their corresponding standard deviation are then obtained. Ultraviolet radiation from sparks / open flames is typically high-intensity, long-lasting, unstable, and flickering. The cumulative number of pulses, N, can be significantly greater than typical interference, with dramatic fluctuations and a large standard deviation. By determining whether the cumulative number of pulses and its standard deviation within this time period exceed the preset threshold, the presence of a flame can be determined, providing early detection. If (a > 0.8n), the percentage of signals exceeding the threshold exceeds 80% over n consecutive seconds of monitoring, indicating that the UV anomaly is continuous, consistent with the physical characteristics of flames / sparks (continuous combustion / bursting). Conversely, if a is very small (for example, a < 0.8n), the exceeding-threshold signals are sporadic and intermittent, more likely interference (such as momentary electromagnetic pulses or sunlight flickering) than actual flames. Furthermore, industrial environments are plagued by numerous interference sources (such as electromagnetic radiation from motors, flickering LED lights, and ambient stray light). These interference sources may occasionally trigger the UV sensor to exceed the threshold, but they will not consistently exceed it. Furthermore, based on engineering experience, using an 80% threshold ratio ensures that real flames are rarely missed (a persistent flame will inevitably result in a high threshold ratio). Furthermore, the 80% threshold ratio filters out most interference. Setting it too low increases false alarms, while setting it too high increases false alarms.

[0081] Understandably, in the early stages of a fire, the burning flames generate a large amount of ultraviolet light in the air, and the number of pulses changes dynamically. By monitoring and analyzing ultraviolet pulse data, fire-related ultraviolet signals can be detected in the early stages of a fire, even before flames and smoke have significantly spread, providing an important basis for early warning of fires. Using ultraviolet light to detect flames has a fast response speed and can capture changes in ultraviolet light signals in the early stages of a fire, enabling ultra-early warning. It is highly sensitive to ultraviolet light signals generated by sparks or open flames, effectively detecting potential fire hazards. As an independent detection method, it can complement other detection technologies and improve the reliability of fire detection systems.

[0082] As an embodiment, in addition to calculating the standard deviation, the variance value of the accumulated pulse number may also be calculated.

[0083] Step S2 specifically includes the following steps:

[0084] Step S21: Acquire a real-time video image of the scene, perform smoke and fire detection processing frame by frame based on the real-time video image of the scene, and obtain the flame position characteristics and continuous images with the same flame;

[0085] Step S22: Obtaining the flame position based on the flame position feature, performing brightness detection based on the brightness of the light emitted by the flame to obtain the brightness information of the flame area, and determining that the flame has evolved into a flame when the brightness of the flame area exceeds a preset threshold;

[0086] Step S23: Perform thermal imaging detection processing on the area where the flame is located to obtain temperature information of the flame area. When the temperature of the flame area exceeds a preset threshold, it is determined that the flame has evolved into a flame.

[0087] Specifically, in step S21, smoke and fire detection is performed on the real-time video image to obtain the flame position feature. When the flame position feature appears in the image in a continuous time sequence, it can be considered that a flame has appeared at that location. Based on the flame position feature, the specific location of the flame is obtained. In step S22, after the specific location of the flame is obtained, the flame is located and its brightness value is detected. Based on the change in the brightness value, it is determined whether the flame area is a real flame. Based on the change in the brightness value, it is verified whether the flame obtained in the previous step has evolved into a flame, thereby improving the accuracy and reliability of the detection. In step S23, thermal imaging detection is performed on the area where the flame is located to obtain a thermal imaging image. The maximum temperature value of the flame area is detected from the thermal imaging image. The maximum temperature value is compared with the preset threshold to determine whether the flame has evolved into a flame.

[0088] Step S21 specifically includes the following steps:

[0089] Step S211: Acquire live video images, perform smoke and fire detection on each frame of the live video images, mark images with flames, and obtain images with flame features and location features of the flames in the images;

[0090] Step S212: Obtaining flame position differences based on time-sequentially consecutive images with flame features. If the flame position difference is less than a preset threshold and appears consecutively in a preset number of images, determine that the flames in the preset number of time-sequentially consecutive images with flame features are the same flame.

[0091] It can be understood that the flame position features are obtained in the real-time video images at the scene to locate the position of the flame. Among them, the AI ​​model is used to detect and process the real-time video images at the scene. The AI ​​model is trained with a large amount of fire image data to learn the characteristics of the flame, including color, shape, dynamic changes, etc. The system collects images in real time through image sensors, and the trained AI model performs smoke and fire detection on the images. The image sensor is set on the wall or other objects at the scene. When a flame is detected, the flame frame position information is recorded. One frame of position information is recorded as (xa0, ya0, xb0, yb0). (xa0, ya0) is the coordinate of the upper left corner of the flame, and (xb0, yb0) is the coordinate of the lower right corner of the flame. The flame position information (xa0, ya0) is detected for the subsequent i-th consecutive time. i, ya i , xb i , yb i ) Make a judgment. According to the overlapping ratio relationship r calculated from the positions of the flame frames before and after, the calculation formula of r is as follows:

[0092] ;

[0093] When Tfl < r < Tfh (the default value of Tfl is 0.2, and the default value of Tfh is 0.9), it is judged as the same flame. When the above conditions are met for more than 6 consecutive frames of detected flames, it is judged that a flame has been detected.

[0094] Among them, the formula r is the ratio of the overlapping area of two frames of flame rectangles to the area of the first frame of flame, which is used to judge whether the flame is continuous. r≈1: The two frames of flames almost completely overlap, and it is judged that the same flame is moving / continuing. r≈0: The two frames of flames hardly overlap, and it is judged as a new flame or interference. The specific operation is as follows: When the flame is first detected (the 0th frame), record (xa0, ya0, xb0, yb0). For each subsequent detected flame (the ith frame), calculate r using the formula. If r is large, such as r > 0.5, r can be judged using a custom threshold, and it can be considered as the same flame and continue to track; if r is very small (close to 0), it can be considered as a new flame and re-initialize the tracking. Step S22 specifically includes the following steps:

[0095] Step S221: Based on the flame position characteristics and the corresponding real-time video image, obtain the on-site flame area, perform brightness detection processing and average value calculation processing on the flame area, and obtain the average brightness of the first flame area;

[0096] Step S222: After adding a filter, perform brightness detection processing on the flame area again to obtain the average brightness of the second flame area; the filter filters out light with a wavelength greater than 750nm;

[0097] Step S223: Combine the average brightness of the first flame area and the average brightness of the second flame area to calculate the infrared occupancy ratio. When the infrared occupancy ratio exceeds the preset threshold, it is judged that the flame has evolved into a fire.

[0098] It can be understood that when the flame area is detected, the average brightness of the current flame area is counted as Y, the average brightness of the full screen is Y1, the infrared detection device switches the filter from infrared cut-off (750nm) to full-pass, and the average brightness of the current flame area is counted again as , the average brightness of the full screen is . The calculation formula of the infrared occupancy ratio r1 of the flame area is as follows:

[0099] ;

[0100] The infrared proportion r1 of the flame area is calculated based on the infrared pixels in the image. When r1>0.3, it is determined to be a flame. The characteristics of the infrared part of the flame radiation spectrum are used to further verify the flame detected by the AI ​​image algorithm. By analyzing the brightness changes of the flame under different filter states, the infrared proportion is calculated to determine whether the area is truly a flame, thereby improving the accuracy and reliability of fire detection. The flame is detected from the perspective of spectral characteristics, which complements other detection methods and increases the basis for judgment. It is more sensitive to the infrared radiation characteristics of the flame and can detect some flames that are not obvious under visible light. It is not interfered by factors such as smoke that affect the propagation of visible light, and can also detect flames well in smoky environments.

[0101] Step S23 specifically includes the following steps:

[0102] Step S231: Based on the real-time video image of the scene, real-time detection and processing of the flame area are performed to obtain information about the visible light flame area, including coordinate information and area range of the visible light flame area; the flame area detection and processing are performed using a target detection algorithm;

[0103] Step S232: Using an infrared thermal imaging camera to obtain a real-time thermal imaging image on site, locating the flame thermal imaging area in the real-time thermal imaging image based on the coordinate information of the visible light flame area within the area range, and obtaining the flame thermal imaging corresponding area;

[0104] Step S233: Traverse the corresponding area of ​​the thermal imaging, count the maximum grayscale value in the area, and obtain the temperature value of the flame area based on the obtained thermal imaging grayscale value and thermal imaging information; the thermal imaging information includes the ambient temperature, the preset distance, and the emissivity of the detection target;

[0105] Step S234: When the temperature value of the flame area is greater than the preset temperature threshold, it is determined that the flame has evolved into flame.

[0106] Specifically, three points corresponding to the same position are selected in each of the images obtained by visible light and thermal imaging, and the visible light is used as a punctuation point for detection and judgment to meet certain requirements. The mapping relationship between visible light coordinates and thermal imaging coordinates is calculated through affine transformation. The flame area obtained previously is mapped to the thermal imaging image to calculate the corresponding area. The maximum grayscale value is traversed through the thermal imaging area, and the temperature information is calculated by combining the current ambient temperature, preset distance, detection target emissivity and other parameters. The calculated temperature information is compared with the preset temperature threshold T. If the calculated temperature is greater than the preset temperature threshold T, it is considered that the flame in the area has evolved into a flame. The specific operation is as follows: three points corresponding to the same position are selected in each of the visible light and thermal imaging for calibration. Assuming that the visible light coordinate points are (x1, y1), (x2, y2), (x3, y3), the visible light is used as a punctuation point for detection and judgment to meet the following requirements:

[0107] x1≠x2, x2≠x3, x1≠x3;

[0108] 0.3<|(y2-y1) / (x2-x1)|<3;

[0109] 0.3<|(y3-y1) / (x3-x1)|<3;

[0110] 0.3<|(y3-y2) / (x3-x2)|<3;

[0111] (x2-x1)(x3-x1)+(y2-y1)(y3-y1)≥0;

[0112] (x1-x2)(x3-x2)+(y1-y2)(y3-y2)≥0;

[0113] (x1-x3)(x2-x3)+(y1-y3)(y2-y3)≥0;

[0114] First, the x-axes of the three visible light coordinates must not completely overlap, ensuring that the three points are not collinear. If they are collinear, the three points can only define a straight line, making it impossible to calibrate the plane mapping relationship. This prevents the affine transformation matrix from being uniquely solved due to the three points being collinear. Then, the absolute value of the slope of the line connecting any two of the three points must be limited to between 0.3 and 3, avoiding extreme cases where the slope is close to horizontal (slope ≈ 0) or close to vertical (slope tends to infinity). If the slope is too close to 0 (horizontal) or infinity (vertical), slight camera distortion can amplify coordinate errors, affecting calibration accuracy. Limiting the slope range ensures a more balanced distribution of the three points and improves the stability of subsequent affine transformations. Finally, using the vector dot product, the dot product of the vectors forming the three points is calculated. By limiting the dot product to be greater than or equal to 0, the angle θ between any two vectors is kept ≤ 90°, ensuring that the triangle formed by the three points is convex and free of concave points, ensuring a reasonable distribution. Avoiding close collinearity or concave distribution of the three points, such as a V-shaped concave pattern, ensures the spatial coverage required by the affine transformation. Specifically, we first eliminate collinear points by ensuring x1≠x2≠x3, then apply slope constraints to eliminate points with extreme angles to minimize distortion sensitivity. Finally, we verify convexity to eliminate points with irrational distribution, ensuring that the three points can support the plane mapping. Only three points that simultaneously meet these conditions can serve as reliable calibration feature points for subsequent affine transformation calculations of coordinate mapping relationships, ensuring more accurate alignment of visible light and thermal imaging coordinates and laying the foundation for dual-light fusion in flame detection.

[0115] Furthermore, by combining visible light and thermal imaging technology, detection is carried out from two aspects: the temperature characteristics and visual characteristics of the flame. The temperature information of the flame is obtained through thermal imaging, and the specific position and range of the flame is determined in combination with the visible light image. The existence and situation of the flame are judged more comprehensively, providing accurate temperature and position basis for fire detection. This method can not only provide intuitive visual information of the flame, but also accurately measure the flame temperature; the temperature detection of the flame is relatively accurate, and can effectively distinguish between high-temperature objects and flames. Through dual-light positioning and mapping relationship calculation, the visible light and thermal imaging information can be accurately matched, thereby improving the accuracy and reliability of detection.

[0116] In a second aspect, the present invention discloses a flame detection system, which includes the above-mentioned flame detection method.

[0117] Specifically, the system implements the flame detection method disclosed in the first aspect. The method uses ultraviolet light to detect the scene to determine whether there is a flame. It can determine the situation and location of the flame in advance and can detect it before it develops into a large-scale situation. Multiple tests are performed on the flame location, and the results of multiple tests are combined to determine whether the flame has evolved into a flame. Each technology detects the fire from a different angle, complements and verifies each other, reduces the misjudgment and missed judgment that may be caused by a single technology, greatly improves the accuracy and reliability of fire detection, and quickly responds to the generation of flames, allowing users to take corresponding measures to the flame more quickly.

[0118] Furthermore, multiple detection methods are used to detect flames, taking into account the multi-dimensional characteristics of the flame, such as position, spectrum, temperature, etc., to gain a more comprehensive understanding of the true situation of the flame and avoid misjudgment that may be caused by judging by a single feature.

[0119] Furthermore, in addition to determining whether there is a flame at the scene, it can also provide information on the flame's location, temperature, and spectral characteristics. In some scenarios where fire needs to be judged, it can provide firefighters and related personnel with more detailed and comprehensive fire conditions, helping them to formulate more effective fire-fighting strategies and rescue plans.

[0120] In a third aspect, the present invention discloses a computer device comprising a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0121] Memory for storing computer programs;

[0122] The processor is configured to implement the steps of the above method when executing the program stored in the memory.

[0123] Specifically, a corresponding computer program is stored in the memory of the computing device, and the processor executes the computer program in the memory to implement the method disclosed in the first aspect. The method uses ultraviolet light to detect the scene to determine whether there is a flame. The situation and location of the flame can be determined in advance, and it can be detected before it develops into a large-scale situation. Multiple detections are performed on the location of the flame, and the results of multiple detections are combined to determine whether the flame has evolved into a flame. Each technology detects the fire from a different angle, complements and verifies each other, reduces the misjudgment and missed judgment that may be caused by a single technology, greatly improves the accuracy and reliability of fire detection, and quickly responds to the generation of flames, allowing users to take corresponding measures to the flame more quickly.

[0124] Furthermore, multiple detection methods are used to detect flames, taking into account the multi-dimensional characteristics of the flame, such as position, spectrum, temperature, etc., to gain a more comprehensive understanding of the true situation of the flame and avoid misjudgment that may be caused by judging by a single feature.

[0125] Furthermore, in addition to determining whether there is a flame at the scene, it can also provide information on the flame's location, temperature, and spectral characteristics. In some scenarios where fire needs to be judged, it can provide firefighters and related personnel with more detailed and comprehensive fire conditions, helping them to formulate more effective fire-fighting strategies and rescue plans.

[0126] In a fourth aspect, the present invention discloses a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.

[0127] Specifically, a computer program is stored in a computer-readable storage medium, and when the computer program is executed by a processor, the method disclosed in the first aspect can be implemented. The method uses ultraviolet detection to detect the scene to determine whether there is a flame. The situation and location of the flame can be determined in advance, and it can be detected before it develops into a large-scale situation. Multiple detections are performed on the location of the flame, and the results of multiple detections are combined to determine whether the flame has evolved into a flame. Each technology detects the fire from a different angle, complements and verifies each other, reduces the misjudgment and missed judgment that may be caused by a single technology, greatly improves the accuracy and reliability of fire detection, and quickly responds to the generation of flames, allowing users to take corresponding measures to the flames more quickly.

[0128] Furthermore, multiple detection methods are used to detect flames, taking into account the multi-dimensional characteristics of the flame, such as position, spectrum, temperature, etc., to gain a more comprehensive understanding of the true situation of the flame and avoid misjudgment that may be caused by judging by a single feature.

[0129] Furthermore, in addition to determining whether there is a flame at the scene, it can also provide information on the flame's location, temperature, and spectral characteristics. In some scenarios where fire needs to be judged, it can provide firefighters and related personnel with more detailed and comprehensive fire conditions, helping them to formulate more effective fire-fighting strategies and rescue plans.

[0130] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0131] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0132] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature identified as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0133] In the present invention, unless otherwise expressly specified or limited, terms such as "mounted," "connected," "connect," and "fixed" should be understood broadly. For example, they may refer to connection, detachable connection, or integration; mechanical connection or electrical connection; direct connection or indirect connection through an intermediate medium; and internal communication between two components or interaction between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0134] In the present invention, unless otherwise expressly specified or limited, a first feature being "above" or "below" a second feature may include the first and second features being in direct contact, or may include the first and second features being in contact not directly but through another feature between them. Furthermore, a first feature being "above," "above," and "above" a second feature may include the first feature being directly above or obliquely above the second feature, or may simply mean that the first feature is higher in level than the second feature. A first feature being "below," "below," and "below" a second feature may include the first feature being directly below or obliquely below the second feature, or may simply mean that the first feature is lower in level than the second feature.

[0135] In the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms should not be understood as necessarily referring to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification.

[0136] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, to the extent such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to encompass such changes and modifications.

[0137] The above description is a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and such modifications or substitutions are intended to be within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A flame detection method, characterized in that: include: Perform real-time ultraviolet detection on the environment and obtain real-time ultraviolet pulse data within a preset time period. When the real-time ultraviolet pulse data is greater than the preset ultraviolet pulse threshold data, it is determined that there is a flame in the current environment; The real-time ultraviolet pulse data within the preset time period includes the cumulative number of pulses and the standard deviation within the preset time period, and the preset ultraviolet pulse threshold data is calculated based on the cumulative number of pulses and the standard deviation within the preset time period; The flame area is acquired based on the real-time video image of the scene, and multiple detection processes are performed on the flame area. The multiple detection and processing results are combined with the preset threshold to make a judgment. When the result exceeds the preset threshold, it is judged that the flame has evolved into a flame. The multiple detection processes include flame position positioning processing, brightness detection processing, and thermal imaging temperature detection processing of the flame area. Perform brightness detection and thermal imaging temperature detection on the flame area to obtain the brightness information and temperature information of the flame area; The flame area brightness information and flame area temperature information are integrated with time, sorted according to the time domain, and flame brightness data and flame temperature data are obtained; Calculate the corresponding inverse numbers for the flame brightness data and flame temperature data within a preset period of time, and combine them into a brightness inverse sequence and a temperature inverse sequence; The flame changes are judged based on the brightness inverse series and the temperature inverse series.

2. The flame detection method according to claim 1, characterized in that: The method of obtaining a flame area based on a real-time video image of the scene, performing multiple detection processes on the flame area, combining the multiple detection process results with a preset threshold value to make a judgment, and judging that the flame has evolved into a flame when the result exceeds the preset threshold value specifically includes the following steps: Acquire real-time video images from the scene, perform fire detection and processing frame by frame based on the real-time video images, and obtain the flame position characteristics and continuous images with the same flame; The flame position is obtained based on the flame position characteristics, and the brightness of the light emitted by the flame is detected to obtain the brightness information of the flame area. When the brightness of the flame area exceeds the preset threshold, it is determined that the flame has evolved into a flame; Thermal imaging detection and processing are performed on the area where the flame is located to obtain temperature information of the flame area. When the temperature of the flame area exceeds the preset threshold, it is determined that the flame has evolved into a flame.

3. The flame detection method according to claim 2, characterized in that: The method of acquiring a live real-time video image and performing smoke and fire detection processing frame by frame based on the live real-time video image to obtain a flame position feature and continuous images with the same flame specifically includes the following steps: Acquire live video images, perform fire and smoke detection on each frame of the live video images, mark images with flames, and obtain images with flame features and the location features of the flames in the images; The flame position difference is obtained based on the images with flame features that are consecutive in time sequence. If the flame position difference is less than a preset threshold and appears consecutively in a preset number of images, it is determined that the flames in the corresponding preset number of images with flame features that are consecutive in time sequence are the same flame.

4. The flame detection method according to claim 2, characterized in that: The method of obtaining the flame position according to the flame position feature, performing brightness detection based on the brightness of the light emitted by the flame to obtain the brightness information of the flame area, and determining that the flame has evolved into a flame when the brightness of the flame area exceeds a preset threshold, specifically includes the following steps: Based on the flame position characteristics and the corresponding real-time video image, the flame area on site is obtained, and the brightness detection processing and average value calculation processing are performed on the flame area to obtain the average brightness of the first flame area; After adding the filter, the flame area is subjected to brightness detection processing to obtain the average brightness of the second flame area; the filter filters out light with a wavelength greater than 750nm; The infrared ratio is calculated by combining the average brightness of the first flame area and the average brightness of the second flame area. When the infrared ratio exceeds the preset threshold, it is determined that the flame has evolved into a flame.

5. The flame detection method according to claim 2, characterized in that: The method of performing thermal imaging detection on the flame area to obtain temperature information of the flame area and determining that the flame has evolved into a flame when the temperature of the flame area exceeds a preset threshold specifically includes the following steps: Based on the real-time video images of the scene, the flame area is detected and processed in real time to obtain the visible light flame area information, including the coordinate information and area range of the visible light flame area; the flame area detection process is processed by the target detection algorithm; For the on-site real-time use of infrared thermal imaging cameras to obtain real-time thermal imaging images, the flame thermal imaging area is located in the real-time thermal imaging image according to the coordinate information of the visible light flame area within the regional range, and the flame thermal imaging corresponding area is obtained; Traverse the corresponding area of ​​the thermal imaging, calculate the maximum grayscale value in the area, and obtain the temperature value of the flame area based on the obtained thermal imaging grayscale value and thermal imaging information; the thermal imaging information includes ambient temperature, preset distance, and detection target emissivity; When the temperature value of the flame area is greater than the preset temperature threshold, it is determined that the flame has evolved into a flame.

6. The flame detection method according to claim 1, characterized in that: When flames are detected in the environment, an open fire warning is issued; The real-time video image calculates the number of flame areas in the image according to the time domain. When the number of flame areas increases, it is judged that the flames in the environment have evolved into flames and the area is getting larger, and a fire warning is issued.

7. The flame detection method according to claim 1, characterized in that: The flame change is judged according to the brightness inverse sequence and the temperature inverse sequence, which specifically includes the following steps: Observe the increase and decrease of the values ​​in the brightness reverse sequence. If the reverse sequence number gradually increases, it means that the fluctuation of the flame brightness data has increased, which may mean that the flame is burning unstably or is affected by other factors. If the reverse sequence number gradually decreases, it may mean that the flame brightness is tending to be stable. For the temperature reverse sequence, if the reverse sequence number increases, the fluctuation of the flame temperature becomes larger and the combustion state becomes unstable; if the reverse sequence number decreases, it means that the flame temperature gradually stabilizes; If the brightness inverse series and the temperature inverse series increase at the same time, it means that the flame is in a state of drastic change, the combustion is unstable and there is a tendency for the fire to spread; if both decrease at the same time, it means that the flame is gradually stabilizing.

8. A flame detection system, characterized in that: The flame detection method includes any one of claims 1 to 7.

9. A computer device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus; Memory for storing computer programs; A processor, configured to implement the steps of the method according to any one of claims 1 to 7 when executing a program stored in a memory.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images