Boiler burner safety time testing method based on CCD camera

Abstract

The invention relates to the field of boiler burner safety, in particular to a boiler burner safety time testing method based on a CCD camera, and the method comprises the steps: collecting a real-time image of a boiler burner based on the CCD camera; establishing an image target feature recognition model based on an image processing algorithm according to the real-time image of the boiler burner to obtain a feature recognition result of the real-time image of the boiler burner; the feature recognition result of the real-time image of the boiler burner is used for carrying out safety test processing to obtain the safety time test result of the boiler burner, the ignition safety time and flameout safety time of the boiler burner are tested through the CCD camera, and the safety time of the boiler burner can be tested in a high-precision and high-reliability mode. The problems that an existing testing method is complex in wiring, prone to interference and limited in application scene and has certain potential safety hazards are solved.

Description

Technical Field

[0001] This invention relates to the field of boiler burner safety, and more specifically to a boiler burner safety time test method based on a CCD camera. Background Technology

[0002] The burner is a crucial component of a boiler, and its safe operating time is a vital safety indicator during boiler operation. With the implementation of the "Boiler Safety Technical Regulations" TSG11-2020 and the "Technical Conditions for Liquid and Gaseous Fuel Burners for Boilers" GB / T36699-2018, the requirement for precise testing of boiler burner safe operating time values ​​has been established. In particular, the ignition safe operating time and flameout safe operating time are strictly controlled to within 3 seconds and 1 second respectively. The ignition safe operating time refers to the time interval between the opening and closing of the ignition fuel solenoid valve when ignition fails, no flame forms, or the flame detection device cannot detect a flame during the ignition phase. The flameout safe operating time refers to the time interval between the start of the main fuel solenoid valve closing when the flame is lost or the flame detection device cannot detect a flame during normal burner operation.

[0003] Traditional burner safety time testing methods mainly rely on manual timing. This involves manually removing the flame detection device from the burner during the ignition and normal operation phases to simulate ignition failure and flame extinction. A high-precision timing device is used to determine the time interval from the start of the action to the burner stopping combustion, which are then used as the ignition safety time and the flameout safety time, respectively. However, manual timing is often subject to subjective errors, and the standard-required ignition safety time and flameout safety time intervals are too short, resulting in significant deviations in actual field applications.

[0004] Currently, the ignition safety time and flameout safety time of burners are usually tested on-site using burner safety performance testing instruments. These instruments integrate the ignition safety time testing module and the flameout safety time testing unit into the same device. On-site wiring is required from the user's boiler burner control cabinet to obtain signals from the ignition fuel solenoid valve, main fuel solenoid valve, and flame detection device, which are then connected to the testing instrument. The testing instrument disconnects the circuit of the flame detection device to simulate ignition failure and flame extinction. The ignition safety time and flameout safety time are determined by detecting the time interval between the gain and loss of voltage signals at the corresponding wiring points of the ignition fuel solenoid valve and main fuel solenoid valve through an embedded timing system.

[0005] Compared to traditional testing methods, this method reduces the error caused by subjective human factors and improves accuracy. However, it often requires professionals or burner manufacturers to perform wiring to obtain signals, which is complex and increases labor costs. Since different test units have different voltage levels, line interference can easily occur, affecting test accuracy. In addition, sometimes it is necessary to damage the wiring in the user's burner control cabinet to connect the tester, which poses a risk of false triggering. Furthermore, since the current signal of the flame detection device cut off during the simulated flameout test is low, the corresponding interface of the tester cannot receive the current from active electrical equipment. The corresponding interface of the tester connected in parallel with the ignition fuel solenoid valve and the main fuel solenoid valve usually receives a voltage of 220VAC. Therefore, there is often a risk of electrical equipment being damaged due to wiring errors.

[0006] Therefore, there is an urgent need for a method that can test the safe time of boiler burners with high precision and high reliability without the need for wiring. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides a method for testing the safe time of boiler burners based on a CCD camera. By acquiring real-time images through a CCD camera and establishing a recognition model, the ignition safe time and flameout safe time of boiler burners can be tested quickly and accurately.

[0008] To achieve the above objectives, the present invention provides a method for testing the safe time of a boiler burner based on a CCD camera, comprising: S1. Acquire real-time images of the boiler burner using a CCD camera; S2. Based on the real-time image of the boiler burner, establish an image target feature recognition model using an image processing algorithm to obtain the feature recognition results of the real-time image of the boiler burner; S3. The boiler burner safety time test result is obtained by using the feature recognition results of the real-time image of the boiler burner for safety test processing.

[0009] Preferably, the real-time image acquisition of the boiler burner based on the CCD camera includes: S1-1. Establish the position of the image acquisition camera based on the CCD camera; S1-2. Based on the position of the image acquisition camera, the images of the boiler burner observation window, the main fuel solenoid valve, and the ignition fuel solenoid valve are acquired as real-time images of the boiler burner. The image acquisition camera positions include the boiler burner observation window position, the main fuel solenoid valve position, and the ignition fuel solenoid valve position. The angle between the CCD camera lens of the main fuel solenoid valve position and the normal of the corresponding indicator light position is no greater than 30 degrees.

[0010] Furthermore, based on the location of the image acquisition camera, images of the boiler burner observation window, the main fuel solenoid valve, and the ignition fuel solenoid valve are acquired as real-time images of the boiler burner, including: S1-2-1. When the boiler burner is in the ignition stage, directly acquire the image of the boiler burner observation window and the image of the main fuel solenoid valve. After adjusting the CCD camera to high frame rate mode, acquire the image of the ignition fuel solenoid valve. Use the image of the boiler burner observation window, the image of the main fuel solenoid valve and the image of the ignition fuel solenoid valve as the real-time image of the boiler burner. S1-2-2. When the boiler burner is in normal operation, the position of the ignition fuel solenoid valve is directly acquired. After adjusting the CCD camera to low frame rate mode, the images of the boiler burner observation window and the main fuel solenoid valve are acquired respectively. The images of the boiler burner observation window, the main fuel solenoid valve, and the ignition fuel solenoid valve are used as real-time images of the boiler burner. S1-2-3. When the flame of the boiler burner is abnormal, after adjusting the CCD camera to high frame rate mode, the images of the boiler burner observation window, the main fuel solenoid valve, and the ignition fuel solenoid valve are collected as real-time images of the boiler burner. The flame anomaly is defined as a 50% decrease in flame brightness between adjacent frames.

[0011] Furthermore, based on the real-time images of the boiler burner, an image target feature recognition model is established using image processing algorithms to obtain the feature recognition results of the real-time images of the boiler burner, including: An image target feature recognition model is established based on image processing algorithms using the real-time images of the boiler burner. The feature recognition results of the real-time image of the boiler burner are obtained based on the image target feature recognition model.

[0012] Furthermore, establishing an image target feature recognition model based on image processing algorithms using the real-time images of the boiler burner includes: A flame image state recognition model is established based on image processing algorithms using the real-time images of the boiler burner. A model for identifying indicator lights was established using the real-time images of the boiler burner. The flame image state recognition model and the indicator light recognition model are used as image target feature recognition models.

[0013] Furthermore, establishing a flame image state recognition model based on image processing algorithms using the real-time images of the boiler burner includes: Historical images of furnace flame states were used as a training set. Using the training set as input, and the corresponding flame color, flame brightness, flame area, flame area growth rate, flame outline, and flame flicker frequency as output, a flame image state recognition model is established based on deep learning.

[0014] Furthermore, establishing an indicator light recognition model using the real-time images of the boiler burner includes: The position of indicator lights in real-time images of boiler burners is obtained based on color features; A training set is established using the positions of the indicator lights; A deep learning-based indicator light recognition model was established using the training set.

[0015] Furthermore, the feature recognition results of the real-time image of the boiler burner obtained according to the image target feature recognition model include: The image target feature recognition result is obtained by inputting the real-time image of the boiler burner observation window into the flame image state recognition model of the image target feature recognition model. The main fuel solenoid valve image and the ignition fuel solenoid valve image of the real-time image of the boiler burner are respectively input into the indicator light recognition model of the image target feature recognition model to obtain the indicator light recognition result; The image target feature recognition results and indicator light recognition results are used as the feature recognition results of the real-time image of the boiler burner.

[0016] Furthermore, the boiler burner safety time test results are obtained by using the feature recognition results of the real-time images of the boiler burner for safety testing, including: S3-1. The flameout safety test result of the boiler burner is obtained by performing flameout safety test processing using the feature recognition result of the real-time image of the boiler burner. S3-1-1. Obtain the current time as the start time t of the flameout safety test; S3-1-2. Based on the start time t of the flameout safety test, obtain the feature recognition results of the boiler burner real-time images of 5 consecutive frames and the corresponding image target feature recognition results. S3-1-3. Determine whether the flame area of ​​the image target feature recognition result corresponding to the feature recognition result of the boiler burner real-time image of the adjacent 5 consecutive frames is less than the flame area threshold. If yes, execute S3-1-4. Otherwise, update the position of the adjacent 5 consecutive frames and return to S3-1-2. S3-1-4. Determine whether the flame brightness value of the image target feature recognition result corresponding to the feature recognition result of the boiler burner real-time image of the adjacent 5 consecutive frames is less than the flame brightness value threshold. If so, obtain the current time as the boiler burner flame extinguishing time t1. Otherwise, update the position of the adjacent 5 consecutive frames and return to S3-1-2. S3-1-5. When the boiler burner flame goes out, the main fuel solenoid valve indicator light extinguishing time t2 is obtained based on the feature recognition result of the real-time image of the boiler burner and the corresponding indicator light recognition result. S3-1-6. The difference between the extinguishing time t2 of the main fuel solenoid valve indicator light and the extinguishing time t1 of the boiler burner flame is used as the result of the boiler burner flameout safety test. Wherein, the flame area threshold is 5% of the flame area of ​​the real-time image of the boiler burner at the start time t of the flameout safety test, corresponding to the image target feature recognition result; and the flame brightness threshold is 10% of the brightness value of the real-time image of the boiler burner at the start time t of the flameout safety test, corresponding to the image target feature recognition result. S3-2. The ignition safety test results of the boiler burner are obtained by performing ignition safety test processing using the feature recognition results of the real-time image of the boiler burner. S3-3. Use the flameout safety test results and ignition safety test results of the boiler burner as the boiler burner safety time test results.

[0017] Furthermore, the ignition safety test results of the boiler burner are obtained by using the feature recognition results of the real-time image of the boiler burner for ignition safety testing, including: The indicator light on time t3 and the indicator light off time t4 of the ignition fuel solenoid valve are obtained using the feature recognition results of the real-time image of the boiler burner. The difference between the indicator light extinguishing time t4 and the indicator light igniting time t3 is used as the ignition safety test result of the boiler burner.

[0018] Compared with the closest existing technology, the present invention has the following advantages: Testing the ignition and shutdown safety times of boiler burners using a CCD camera offers high precision and reliability. This method overcomes the limitations of existing testing methods, such as complex wiring, susceptibility to interference, and restricted applicability, which pose certain safety risks. By simultaneously capturing the flame and valve indicator light status using a CCD camera, image processing technology identifies the time points when the flame extinguishes and the indicator lights illuminate and extinguish, calculating the time difference between these specific time points to obtain the safety time. This eliminates the need for any physical wiring, avoiding circuit interference and complex installation, while simultaneously improving testing accuracy and reliability. Attached Figure Description

[0019] Figure 1 This is a flowchart of a boiler burner safety time test method based on a CCD camera provided by the present invention; Figure 2 This is a flowchart of the flame recognition algorithm for a boiler burner safety time test method based on a CCD camera provided by the present invention. Detailed Implementation

[0020] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example This invention provides a method for testing the safe time of boiler burners based on a CCD camera, such as... Figure 1 As shown, it includes: S1. Acquire real-time images of the boiler burner using a CCD camera; S2. Based on the real-time image of the boiler burner, establish an image target feature recognition model using an image processing algorithm to obtain the feature recognition results of the real-time image of the boiler burner; S3. The boiler burner safety time test result is obtained by using the feature recognition results of the real-time image of the boiler burner for safety test processing.

[0023] S1 specifically includes: S1-1. Establish the position of the image acquisition camera based on the CCD camera; S1-2. Based on the position of the image acquisition camera, the images of the boiler burner observation window, the main fuel solenoid valve, and the ignition fuel solenoid valve are acquired as real-time images of the boiler burner. The image acquisition camera positions include the boiler burner observation window position, the main fuel solenoid valve position, and the ignition fuel solenoid valve position. The angle between the CCD camera lens of the main fuel solenoid valve position and the normal of the corresponding indicator light position is no greater than 30 degrees.

[0024] S1-2 specifically includes: S1-2-1. When the boiler burner is in the ignition stage, directly acquire the image of the boiler burner observation window and the image of the main fuel solenoid valve. After adjusting the CCD camera to high frame rate mode, acquire the image of the ignition fuel solenoid valve. Use the image of the boiler burner observation window, the image of the main fuel solenoid valve and the image of the ignition fuel solenoid valve as the real-time image of the boiler burner. S1-2-2. When the boiler burner is in normal operation, the position of the ignition fuel solenoid valve is directly acquired. After adjusting the CCD camera to low frame rate mode, the images of the boiler burner observation window and the main fuel solenoid valve are acquired respectively. The images of the boiler burner observation window, the main fuel solenoid valve, and the ignition fuel solenoid valve are used as real-time images of the boiler burner. S1-2-3. When the flame of the boiler burner is abnormal, after adjusting the CCD camera to high frame rate mode, the images of the boiler burner observation window, the main fuel solenoid valve, and the ignition fuel solenoid valve are collected as real-time images of the boiler burner. The flame anomaly is defined as a 50% decrease in flame brightness between adjacent frames.

[0025] S2 specifically includes: S2-1. Using the real-time images of the boiler burner, establish an image target feature recognition model based on image processing algorithms; S2-2. Obtain the feature recognition results of the real-time image of the boiler burner based on the image target feature recognition model.

[0026] S2-1 specifically includes: S2-1-1. Establish a flame image state recognition model based on image processing algorithms using the real-time images of the boiler burner. S2-1-2. Establish an indicator light recognition model using the real-time images of the boiler burner; S2-1-3. The flame image state recognition model and the indicator light recognition model are used as image target feature recognition models.

[0027] S2-1-1 specifically includes: S2-1-1-1. Obtain historical furnace flame state images as a training set; S2-1-1-2. Using the training set as input, and the corresponding flame color, flame brightness, flame area, flame area growth rate, flame outline and flame flicker frequency as output, a flame image state recognition model is established based on deep learning for training.

[0028] S2-1-2 specifically includes: S2-1-2-1. Obtain the indicator light positions of the boiler burner in real-time images based on color features; S2-1-2-2, Establish a training set using the positions of the indicator lights; S2-1-2-3. Utilize the training set to establish an indicator light recognition model based on deep learning.

[0029] S2-2 specifically includes: S2-2-1. Using the real-time image of the boiler burner observation window, input the flame image state recognition model of the image target feature recognition model to obtain the image target feature recognition result; S2-2-2: Using the main fuel solenoid valve image and the ignition fuel solenoid valve image of the real-time image of the boiler burner, respectively input into the indicator light recognition model of the image target feature recognition model to obtain the indicator light recognition result; S2-2-3. Use the image target feature recognition result and the indicator light recognition result as the feature recognition result of the real-time image of the boiler burner.

[0030] S3 specifically includes: S3-1. The flameout safety test result of the boiler burner is obtained by performing flameout safety test processing using the feature recognition result of the real-time image of the boiler burner. S3-2. The ignition safety test results of the boiler burner are obtained by performing ignition safety test processing using the feature recognition results of the real-time image of the boiler burner. S3-3. Use the flameout safety test results and ignition safety test results of the boiler burner as the boiler burner safety time test results.

[0031] S3-1 specifically includes: S3-1-1. Obtain the current time as the start time t of the flameout safety test; S3-1-2. Based on the start time t of the flameout safety test, obtain the feature recognition results of the boiler burner real-time images of 5 consecutive frames and the corresponding image target feature recognition results. S3-1-3. Determine whether the flame area of ​​the image target feature recognition result corresponding to the feature recognition result of the boiler burner real-time image of the adjacent 5 consecutive frames is less than the flame area threshold. If yes, execute S3-1-4. Otherwise, update the position of the adjacent 5 consecutive frames and return to S3-1-2. S3-1-4. Determine whether the flame brightness value of the image target feature recognition result corresponding to the feature recognition result of the boiler burner real-time image of the adjacent 5 consecutive frames is less than the flame brightness value threshold. If so, obtain the current time as the boiler burner flame extinguishing time t1. Otherwise, update the position of the adjacent 5 consecutive frames and return to S3-1-2. S3-1-5. When the boiler burner flame goes out, the main fuel solenoid valve indicator light extinguishing time t2 is obtained based on the feature recognition result of the real-time image of the boiler burner and the corresponding indicator light recognition result. S3-1-6. The difference between the extinguishing time t2 of the main fuel solenoid valve indicator light and the extinguishing time t1 of the boiler burner flame is used as the result of the boiler burner flameout safety test. Wherein, the flame area threshold is 5% of the flame area of ​​the real-time image of the boiler burner at the start time t of the flameout safety test, corresponding to the image target feature recognition result; and the flame brightness threshold is 10% of the brightness value of the real-time image of the boiler burner at the start time t of the flameout safety test, corresponding to the image target feature recognition result.

[0032] S3-2 specifically includes: S3-2-1. Using the feature recognition results of the real-time image of the boiler burner, obtain the indicator light on time t3 and the indicator light off time t4 of the ignition fuel solenoid valve respectively. S3-2-2, The difference between the indicator light extinguishing time t4 and the indicator light igniting time t3 is used as the ignition safety test result of the boiler burner.

[0033] In this embodiment, the specific process of the boiler burner safety time test method based on CCD camera is as follows: 1) Camera Calibration, Setup, and Acquisition: Three industrial CCD cameras are positioned at suitable locations and angles within the boiler burner operating environment. One camera is placed at the burner observation window to clearly and accurately capture the moment the burner flame extinguishes. The other two cameras are positioned near the main fuel solenoid valve and the ignition fuel solenoid valve, respectively, to clearly capture the instantaneous movement of the indicator lights reflecting the actions of these two valves, avoiding any obstructions. For the observation window camera, a high-temperature resistant quartz glass protective cover is installed in front of the lens to effectively resist the harsh environment such as high temperatures at the boiler site. A regular cleaning mechanism is designed, with the lens surface cleaned every 4 hours using compressed air to ensure image quality. The cameras at the main fuel solenoid valve and the ignition fuel solenoid valve are equipped with dustproof housings, and the lens angle is precisely adjusted to an angle ≤30° with the normal of the indicator light to avoid glare interference caused by direct sunlight. Camera parameters, such as exposure time and gain, were adjusted to adapt to the characteristics of flame and indicator light brightness. Each camera was individually calibrated to ensure image quality and color accuracy. A standard time signal source and a high-precision time synchronizer were used to synchronize the three cameras to ensure timestamp consistency. A multi-mode acquisition strategy was adopted to address the combustion characteristics at different stages. During the ignition stage, the camera at the ignition fuel solenoid valve started high-frame-rate mode acquisition for 10 seconds to ensure the capture of key features at the moment of ignition. During normal operation, the camera at the observation port and the camera located at the main fuel solenoid valve acquired data in low-frame-rate mode. Once an abnormal flame was detected, the system immediately switched to high-frame-rate mode acquisition, with a sudden drop in flame brightness as the switching condition. When a decrease in flame brightness of more than 50% was detected for two consecutive frames, high-frame-rate mode acquisition was triggered.

[0034] 2) Flameout safety time test: After starting the boiler burner and bringing it into normal combustion mode, the burner is shut down to simulate a fault-induced flameout by adjusting the air-fuel ratio, air pressure switch, and gas pressure switch. A CCD camera placed at the observation window captures images in real time at a high frame rate to obtain image data of the flame during normal combustion and image data of the moment the flame is extinguished. The system automatically records the timestamp of each frame and uses a flame image recognition algorithm to determine whether the flame is extinguished. The timestamp of the moment of extinguishment is recorded as the flame extinguishing time point as t1. Once the flame is determined to be extinguished, a signal command is sent through the multi-camera synchronous signal acquisition program to activate the camera at the main fuel solenoid valve. This camera captures the working status image of the main fuel solenoid valve indicator light. During normal burner operation, the main fuel solenoid valve is always energized and in a normally open state, so the indicator light is always on. After the flame is extinguished, it closes due to power loss, and the indicator light should also turn off. The CCD camera captures the image data of the main fuel solenoid valve indicator light from the on state to the off state instantaneously, and records the timestamp as t2. Subtracting the flame extinguishing time from this timestamp yields the flameout safety time as t2-t1. According to the "Boiler Safety Technical Regulations" TSG11-2020 and the "Technical Conditions for Liquid and Gas Fuel Burners for Boilers" GB / T36699-2018, the flameout safety time must not exceed 1 second. Therefore, the system sets the flameout safety time threshold to ≤1 second. If this threshold is exceeded, the system issues a warning that the flameout safety time has exceeded the limit.

[0035] 3) Ignition Safety Time Test: In this stage, first manually remove the flame detection device probe or disconnect its circuit, then restart the boiler burner. The burner will run the ignition program normally for ignition. However, since the flame detector is disconnected, simulating ignition failure, the ignition fuel solenoid valve is generally in a normally closed state, so the indicator light is also in a constantly dim state. During the execution of the ignition program, the ignition fuel solenoid valve will be energized and open, and the indicator light will also brighten until the ignition program reaches the point where the flame detector cannot detect a flame and is considered to have failed ignition. At this point, the ignition fuel solenoid valve closes, and the indicator light goes out. The CCD camera at the ignition fuel solenoid valve captures the image of the moment the indicator light lights up. The system records the data and image data of the moment the indicator light goes out, and records the timestamps as t3 and t4. The duration of the ignition fuel solenoid valve light being on is the difference between the two timestamps, t4-t3, which gives the ignition safety time. According to the "Boiler Safety Technical Regulations" TSG11-2020 and the "Technical Conditions for Liquid and Gas Fuel Burners for Boilers" GB / T36699-2018, the ignition safety time must not exceed 3 seconds. Therefore, the system sets the flameout safety time threshold to ≤3 seconds. If this threshold is exceeded, the system will issue an ignition safety time exceeding the limit warning.

[0036] In this embodiment, the boiler burner safety time test method based on a CCD camera, the actual application system process includes: 1) CCD camera: An industrial-grade CCD camera is selected, which has the characteristics of low noise, high sensitivity, and high resolution, and the frame rate is no less than 200 frames / second. It can clearly image in complex lighting environments, adapt to environmental factors such as high temperature and dust on site, and can realize the function of simultaneous signal acquisition of multiple cameras.

[0037] 2) Image Acquisition Card: Connects to the CCD camera, responsible for acquiring image data captured by the CCD camera and transmitting the data to the data processing unit. A gigabit network image acquisition card compatible with the CCD camera should be selected, supporting high-speed data transmission. The data transmission rate should be no less than 200MB / s to ensure real-time transmission of high frame rate image data.

[0038] 3) Data processing unit: It adopts a high-performance computer, equipped with a multi-core processor and large-capacity memory, receives image data transmitted by the image acquisition card, identifies the flame extinguishing and indicator light working status through image recognition algorithms, calculates the flameout safety time and ignition safety time, and sets the safety time threshold to realize the over-limit warning function.

[0039] 4) Storage Unit: Connected to the data processing unit, this unit stores image data, timestamp data, and calculated security time data. The storage unit also includes a data backup module that periodically backs up the stored data to a remote server to prevent data loss. Solid-state drives (SSDs) are used as the storage medium to ensure data storage stability and fast read / write speeds.

[0040] 5) Display unit: Connected to the data processing unit, it is used to display the flame image, indicator light status, and calculated flameout safety time and ignition safety time in real time, so that operators can intuitively understand the test situation.

[0041] System workflow: 1) System Initialization: Start the data processing unit and CCD camera, and perform a system self-test. Set the CCD camera parameters, such as frame rate, exposure time, and gain, to ensure clear imaging of the flame and indicator lights.

[0042] 2) Image Acquisition and Processing: A CCD camera acquires real-time image data of the boiler burner flame and the solenoid valve indicator lights, which is then transmitted to the data processing unit via an image acquisition card. The data processing unit processes the images in real time, identifying the status of the flame and indicator lights. The system automatically records the timestamps of flame extinguishing and indicator light operation status.

[0043] 3) Safety time calculation: When the system detects that the flame is extinguished, record the timestamp t1. When the system detects that the main fuel solenoid valve working indicator light goes out from the moment it turns on, record the timestamp t2. Calculate the flameout safety time as t2-t1 and compare it with the flameout safety time threshold set by the system to determine whether an over-limit warning has been issued. During the ignition stage, the timestamps for the moment the ignition fuel solenoid valve working indicator light turns on and the moment it goes out are t3 and t4. Calculate the ignition safety time as t4-t3 and compare it with the ignition safety time threshold set by the system to determine whether an over-limit warning has been issued.

[0044] 4) Data storage and display: The system stores image data, timestamp data, and calculation results in the storage unit. The display unit displays the flame image, indicator light status, and test results in real time for operators to view.

[0045] 5) System Shutdown: After the test, the system automatically saves all data and shuts down the CCD camera and data processing unit. Testers can query and analyze the test results using the data management software.

[0046] On-site testing and compatibility: 1) On-site safety pre-treatment: A comprehensive on-site safety inspection and pre-treatment must be carried out before testing to ensure the safety and reliability of the testing process. Use a professional grounding resistance tester to measure the grounding resistance of the boiler burner control cabinet and record the measurement results. If the grounding resistance does not meet the requirements, rectification must be carried out before testing can proceed. Place obvious warning signs near the CCD camera to prevent unauthorized personnel from approaching. Arrange dedicated personnel to be on duty during the testing process to ensure safety. Insulate and wrap exposed wires near the solenoid valve to prevent accidental contact between camera cables and equipment wires that may cause interference. Check whether the insulation layer of all cables is intact. If there is any damage, it should be replaced in time. Arrange the cable routing reasonably to avoid contact with high-temperature equipment or moving parts.

[0047] 2) Emergency Interruption Mechanism: To address abnormal situations during the testing process, a real-time safety monitoring and emergency interruption mechanism is adopted. Real-time safety monitoring mainly targets abnormal flame patterns captured by the camera at the observation port (such as a sudden increase in flame area caused by deflagration) and the status of the solenoid valve indicator light exceeding the safe time threshold. When an abnormal situation is detected, the test is immediately terminated, an audible and visual alarm is triggered, and on-site personnel are notified to cut off the fuel supply, ensuring that the system is in a safe state. The abnormal time point and related image data are recorded to provide a basis for subsequent analysis.

[0048] 3) Data Secondary Verification Mechanism: After the test is completed, a secondary verification of the data is required to ensure the accuracy and reliability of the test data. The test results of the CCD camera are automatically compared with the action log time of the PLC built into the burner control cabinet (if any). When the difference between the two exceeds 50ms, the manual review process is initiated. The manual review requires playing back the images of key time periods frame by frame to confirm the time points of flame state changes, comparing the solenoid valve action time with the flame state change time, checking whether the timestamp records are accurate, and whether there are any omissions or errors. Questionable data points are marked, and retesting is required if necessary.

[0049] In this embodiment, the identification algorithm for the boiler burner safety time test method based on a CCD camera is as follows: 1) such as Figure 2 As shown, the flame recognition algorithm is as follows: A deep learning-based flame detection model can accurately identify the presence or absence of flames, reducing human judgment errors and improving detection sensitivity. The process begins by using a massive database of expert-diagnosed furnace flame state images as an expert library. Support Vector Machines (SVMs) or neural network models are trained through self-learning to extract flame feature data, resulting in a flame image recognition algorithm. Combustion images of the furnace flame are acquired using a CCD camera, and the acquired image data undergoes filtering, edge contour processing, and brightness signal processing. The RGB images are converted to the HSV color space to facilitate the extraction of flame color features. The self-learning image recognition algorithm extracts flame feature data such as furnace flame color, flame brightness, flame area, flame area growth rate, flame outline, and flicker frequency to update the image library. Multiple self-learning training sessions are conducted to continuously optimize the flame image recognition algorithm, improving the accuracy of judgment and the precision of control. The optimized flame recognition algorithm is then used to classify the flame state and determine the presence or absence of a flame.

[0050] To accurately determine flame extinction, a dual-condition + time window enhanced judgment logic is adopted. First, a primary extinction judgment is performed. When the flame area is detected to be less than 5% of the initial area for 5 consecutive frames, and the brightness value is lower than 10% of the initial brightness, the primary judgment is that the flame is extinguished. Then, the flame status is verified again within a 10ms time window. If the above two extinction conditions are still met, the flame is confirmed to be extinguished, and the timestamp t1 is recorded. The dual-condition + time window mechanism can effectively avoid misjudgment caused by flame jitter or instantaneous occlusion.

[0051] To address the impact of potential reflections and other interference factors on flame status recognition during boiler operation, a region masking anti-interference processing scheme is adopted. The flame region is pre-divided, and threshold segmentation is performed in the HSV color space based on the flame's color characteristics (e.g., red, yellow) to extract and label possible flame regions, generating a region mask. Morphological processing is then applied only to feature variations within the region of interest to eliminate small noise areas and connect adjacent flame regions. For interference signals in non-flame areas, the characteristic patterns of various interference sources are collected and analyzed to establish an interference feature library. An interference detection algorithm is designed to identify interference regions in the image in real time, marking these regions so they can be ignored during flame recognition.

[0052] 2) Indicator Light Recognition Algorithm: Based on color and shape features, this indicator light recognition model accurately identifies the on / off state of indicator lights. First, the indicator light's location in the image is determined using color features. Typically, indicator lights are red or green, clearly distinguishable from their surroundings. Template matching or deep learning-based object detection algorithms (such as YOLO) can be used to precisely locate the indicator light. Then, the indicator light's state is identified by performing color analysis on the located indicator light area to determine whether the light is on or off.

[0053] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0054] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0055] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0056] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0057] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for testing the safe time of a boiler burner based on a CCD camera, characterized in that, include: S1. Acquire real-time images of the boiler burner using a CCD camera; S2. Based on the real-time image of the boiler burner, establish an image target feature recognition model using an image processing algorithm to obtain the feature recognition results of the real-time image of the boiler burner; S2-1. Using the real-time images of the boiler burner, establish an image target feature recognition model based on image processing algorithms; S2-1-1. Establish a flame image state recognition model based on image processing algorithms using the real-time images of the boiler burner. S2-1-2. Establish an indicator light recognition model using the real-time images of the boiler burner; S2-1-3. The flame image state recognition model and the indicator light recognition model are used as image target feature recognition models; S2-2. Obtain the feature recognition results of the real-time image of the boiler burner based on the image target feature recognition model; S3. The boiler burner safety time test result is obtained by using the feature recognition results of the real-time image of the boiler burner for safety test processing.

2. The boiler burner safety time test method based on a CCD camera as described in claim 1, characterized in that, The real-time images of the boiler burner acquired using a CCD camera include: S1-1. Establish the position of the image acquisition camera based on the CCD camera; S1-2. Based on the position of the image acquisition camera, the images of the boiler burner observation window, the main fuel solenoid valve, and the ignition fuel solenoid valve are acquired as real-time images of the boiler burner. The image acquisition camera positions include the boiler burner observation window position, the main fuel solenoid valve position, and the ignition fuel solenoid valve position. The angle between the CCD camera lens of the main fuel solenoid valve position and the normal of the corresponding indicator light position is no greater than 30 degrees.

3. The boiler burner safety time test method based on a CCD camera as described in claim 2, characterized in that, Based on the location of the image acquisition camera, images of the boiler burner observation window, the main fuel solenoid valve, and the ignition fuel solenoid valve are acquired as real-time images of the boiler burner, including: S1-2-1. When the boiler burner is in the ignition stage, directly acquire the image of the boiler burner observation window and the image of the main fuel solenoid valve. After adjusting the CCD camera to high frame rate mode, acquire the image of the ignition fuel solenoid valve. Use the image of the boiler burner observation window, the image of the main fuel solenoid valve and the image of the ignition fuel solenoid valve as the real-time image of the boiler burner. S1-2-2. When the boiler burner is in normal operation, the position of the ignition fuel solenoid valve is directly acquired. After adjusting the CCD camera to low frame rate mode, the images of the boiler burner observation window and the main fuel solenoid valve are acquired respectively. The images of the boiler burner observation window, the main fuel solenoid valve, and the ignition fuel solenoid valve are used as real-time images of the boiler burner. S1-2-3. When the flame of the boiler burner is abnormal, after adjusting the CCD camera to high frame rate mode, the images of the boiler burner observation window, the main fuel solenoid valve, and the ignition fuel solenoid valve are collected as real-time images of the boiler burner. The flame anomaly is defined as a 50% decrease in flame brightness between adjacent frames.

4. The boiler burner safety time test method based on a CCD camera as described in claim 1, characterized in that, The flame image state recognition model established using the real-time images of the boiler burner and based on image processing algorithms includes: Historical images of furnace flame states were used as a training set. Using the training set as input, and the corresponding flame color, flame brightness, flame area, flame area growth rate, flame outline, and flame flicker frequency as output, a flame image state recognition model is established based on deep learning.

5. The boiler burner safety time test method based on a CCD camera as described in claim 4, characterized in that, The model for identifying indicator lights using real-time images of the boiler burner includes: The position of indicator lights in real-time images of boiler burners is obtained based on color features; A training set is established using the positions of the indicator lights; A deep learning-based indicator light recognition model was established using the training set.

6. The boiler burner safety time test method based on a CCD camera as described in claim 1, characterized in that, The feature recognition results of the real-time image of the boiler burner obtained according to the image target feature recognition model include: The image target feature recognition result is obtained by inputting the real-time image of the boiler burner observation window into the flame image state recognition model of the image target feature recognition model. The main fuel solenoid valve image and the ignition fuel solenoid valve image of the real-time image of the boiler burner are respectively input into the indicator light recognition model of the image target feature recognition model to obtain the indicator light recognition result; The image target feature recognition results and indicator light recognition results are used as the feature recognition results of the real-time image of the boiler burner.

7. The boiler burner safety time test method based on a CCD camera as described in claim 1, characterized in that, The safety test results of the boiler burner are obtained by using the feature recognition results of the real-time image of the boiler burner for safety testing, including: S3-1. The flameout safety test result of the boiler burner is obtained by performing flameout safety test processing using the feature recognition result of the real-time image of the boiler burner. S3-1-1. Obtain the current time as the start time t of the flameout safety test; S3-1-2. Based on the start time t of the flameout safety test, obtain the feature recognition results of the boiler burner real-time images of 5 consecutive frames and the corresponding image target feature recognition results. S3-1-3. Determine whether the flame area of ​​the image target feature recognition result corresponding to the feature recognition result of the boiler burner real-time image of the adjacent 5 consecutive frames is less than the flame area threshold. If yes, execute S3-1-4. Otherwise, update the position of the adjacent 5 consecutive frames and return to S3-1-2. S3-1-4. Determine whether the flame brightness value of the image target feature recognition result corresponding to the feature recognition result of the boiler burner real-time image of the adjacent 5 consecutive frames is less than the flame brightness value threshold. If so, obtain the current time as the boiler burner flame extinguishing time t1. Otherwise, update the position of the adjacent 5 consecutive frames and return to S3-1-2. S3-1-5. When the boiler burner flame goes out, the main fuel solenoid valve indicator light extinguishing time t2 is obtained based on the feature recognition result of the real-time image of the boiler burner and the corresponding indicator light recognition result. S3-1-6. The difference between the extinguishing time t2 of the main fuel solenoid valve indicator light and the extinguishing time t1 of the boiler burner flame is used as the result of the boiler burner flameout safety test. Wherein, the flame area threshold is 5% of the flame area of ​​the real-time image of the boiler burner at the start time t of the flameout safety test, corresponding to the image target feature recognition result; and the flame brightness threshold is 10% of the brightness value of the real-time image of the boiler burner at the start time t of the flameout safety test, corresponding to the image target feature recognition result. S3-2. The ignition safety test results of the boiler burner are obtained by performing ignition safety test processing using the feature recognition results of the real-time image of the boiler burner. S3-3. Use the flameout safety test results and ignition safety test results of the boiler burner as the boiler burner safety time test results.

8. The boiler burner safety time test method based on a CCD camera as described in claim 7, characterized in that, The ignition safety test results of the boiler burner are obtained by performing ignition safety test processing using the feature recognition results of the real-time image of the boiler burner, including: The indicator light on time t3 and the indicator light off time t4 of the ignition fuel solenoid valve are obtained using the feature recognition results of the real-time image of the boiler burner. The difference between the indicator light extinguishing time t4 and the indicator light igniting time t3 is used as the ignition safety test result of the boiler burner.

Images