616 results about "Flame detection" patented technology

The invention discloses a forest fire hidden danger identification method and device for a power transmission line based on a multi-modal large model, and the method comprises the steps: constructing a forest fire hidden danger identification model for a smog or flame-containing image outputted by a conventional target detection model, and enabling the model to guide a dialogue set through the combination with forest fire hidden danger identification, thereby achieving the recognition of the forest fire hidden danger. The dialogue context and the image visual features are fused, so that the deep fusion of the text semantics and the firework image features is realized; and finally, inputting the deeply fused features into a large language model, so that a forest fire hidden danger output dialogue of the power transmission line can be obtained by utilizing the deep semantic understanding capability of the large language model, and then a forest fire hidden danger recognition result of the power transmission line is obtained. Therefore, according to the invention, the judgment of whether the smoke and fire can damage the power transmission line is realized, the technology upgrade from smoke and fire existence detection to equipment threat research and judgment is completed, and on the basis, the problems of resource waste and untimely crisis response caused by reporting all smoke and fire information in the traditional technology can be avoided.

The invention discloses a virtual fire extinguishing training system and method based on deep learning flame recognition. The virtual fire extinguishing training system comprises a flame image acquisition module, a deep learning flame recognition module, a fire scene generation module, a fire extinguishing equipment simulation module and a scene customization module. And the flame image acquisition module is used for acquiring flame images and video data in different environments in real time through a high-definition camera, an infrared thermal imager, a panoramic camera and a sensor network. According to the invention, feature extraction and model training are carried out on mass flame images by using a deep learning flame recognition algorithm, accurate recognition of multi-dimensional information such as flame form, color, dynamic features and the like is realized, the flame detection capability in a complex environment is remarkably improved, and meanwhile, the system combines a physical engine and a scene generation module, so that the system is more accurate and reliable. The three-dimensional fire scene can be dynamically generated according to the real-time recognition result, the effects of fire spreading, smoke diffusion and the like are achieved, the constraint of a preset scene is eliminated, and training is more flexible and authentic.

The invention provides a combustible gas leakage protection system, which is characterized in that a control module of the combustible gas leakage protection system judges whether the concentration of combustible gas is lower than a first concentration threshold value or not when receiving concentration information detected by a concentration detection module, and generates a first control instruction when the concentration of the combustible gas is not lower than the first concentration threshold value. The control module further generates a second control instruction when the combustible gas concentration is not lower than a second concentration threshold value and the flame detection module detects that no open fire exists, and generates a third control instruction when the combustible gas concentration is not lower than the second concentration threshold value and the open fire exists. And the protection module executes different protection actions according to different control instructions of the control module. According to the application, the response scheme is divided into three grades, and corresponding protection strategies are implemented according to the detection signals of different grades, so that the response sensitivity and accuracy when the combustible gas leaks can be remarkably improved.

The invention discloses an infrared visible light flame detection system and method, and belongs to the technical field of computer vision detection. In order to solve the problems that sensor data is very redundant and has time delay, and the current flame detection technology is difficult to effectively integrate information from different modals and accurately position the flame position, the flame position is detected by analyzing infrared and visible light image data. The method is suitable for fire detection in large factories, warehouses and outdoor open large spaces such as forest parks, the target model can detect flames at high speed and high precision in a complex environment, the flames can be detected more accurately by combining thermal imaging data captured by an infrared camera and detail images provided by a visible light camera, and the target model can be applied to fire detection. Even under the influence of negative factors such as light, temperature and shielding, the stability and accuracy of the system can be kept, the accuracy and efficiency of flame detection are improved, and the method is especially suitable for severe environment conditions and occasions where a traditional method is difficult to apply.

The invention discloses a forest fire fighting robot path planning method and system based on ground-air cooperation, and the method comprises the steps: carrying out the air scanning of a disaster area through employing an unmanned plane to carry a laser radar and infrared and visible light cameras, and obtaining the space information and fire intensity of a fire point in real time; the unmanned aerial vehicle is in communication connection with the wind power fire extinguishing robot, shares the three-dimensional point cloud model and performs environment proofreading and positioning; an improved A * algorithm is used for path planning, and a path planning mathematical model with the minimum disaster degree as the target is constructed; through fusion of a global satellite navigation system and an inertial navigation system, the working path of the wind power fire extinguishing robot is dynamically updated and planned; after a target fire point is reached, the wind power fire extinguishing robot identifies a fire source, calculates a safe fire extinguishing distance by using a flame detection model result, and dynamically adjusts an air duct angle according to a flame state to perform fire extinguishing operation. Through cooperative work between the unmanned aerial vehicle and the fire extinguishing robot, the efficiency and the intelligent degree of fire extinguishing operation are improved.

The invention discloses a method for deploying a flame detection spiking neural network on an FPGA (Field Programmable Gate Array). According to the method, codes of the flame detection spiking neural network are written by using a Viis HLS tool, an IP (Internet Protocol) core file is generated and burnt on the FPGA, a related digital circuit is generated, the processing speed of the flame detection spiking neural network is accelerated, and the operation power consumption is reduced. And in the design process, the technical means such as ATAN arc tangent activation function calculation and short-circuit logic calculation strategies are realized by utilizing a lookup table, so that unnecessary calculation and power consumption can be reduced, the use amount of a DSP (Digital Signal Processor) is remarkably reduced, the hardware calculation operation process is simplified, and the pulse neural network can run on an FPGA (Field Programmable Gate Array) with relatively low power consumption. Therefore, the method provided by the invention provides a set of solution for implementing effective deployment on the low-power-consumption edge equipment based on the FPGA for a flame detection application scene.

The invention relates to the technical field of hydrogenation, and discloses a hydrogenation system and method suitable for small solid hydrogen storage, and the hydrogenation system comprises an emergency cut-off system, a scram button, an emergency switch-off system and a control system, the emergency cut-off valve is electrically connected with the emergency stop button and the PLC control system and is used for cutting off a hydrogen passage when an abnormality occurs; and the PLC control system is electrically connected with the emergency cut-off system, is electrically connected with the hydrogen leakage and flame detection system and the process equipment system, and is used for receiving signals and controlling the operation of each module. According to the invention, hydrogenation is carried out on a large batch of small solid hydrogen storage tanks in the form of a hydrogenation tank changing machine. The machine is connected with the solid hydrogen storage tank in a quick-plug connector mode, when hydrogenation is needed, the empty tank only needs to be inserted into equipment, plug-and-play is achieved, and the empty tank is pulled out after being filled up. The problem of filling efficiency of a large batch of solid hydrogen storage tanks can be solved, the service efficiency of a hydrogen refueling station is improved, the requirements for the working conditions and the number of operators are not high, human intervention is little, and the automation degree is higher.

The invention discloses a gas-phase combustion model construction method based on a dynamic self-adaptive grid, which comprises the following steps: establishing a steady-state flame control equation set according to the property of target fuel, and obtaining laminar flame velocities and laminar flame thicknesses under different conditions; establishing an unsteady-state reactive Navier-Stokes equation set, performing spatial filtering on each variable in the equation set, and constructing a sub-lattice scale turbulence model; coupling a thickened flame model and a disturbance factor model in the filtered equation set; limiting the encrypted area of the self-adaptive grid near a flame surface through a flame detection function; and solving the equation set to obtain a variable of a next time step, updating a disturbance factor and the number of layers of the adaptive grid, and re-dividing the adaptive grid of the layered block structure to realize dynamic tracking of the flame position and real-time adjustment of the grid resolution. According to the method, the disturbance factor and the grid encryption layer number can be adjusted in real time based on the laminar flame speed and the flame thickness in the current state, and the combustion modeling precision and the calculation efficiency are improved.

The invention relates to the field of chemical incineration, in particular to a waste gas heat value self-adaptive combined combustor which comprises an ignition device, a flame detector, a sight glass, a hearth, a waste gas pipeline, a waste gas combustion-supporting pipeline, a main gas pipeline and a main gas combustion-supporting pipeline. Different types of waste gas can enter the hearth through the waste gas pipeline, the waste gas can effectively flow to the hearth under the action of the waste gas combustion-supporting pipeline, main gas can be better and fully combusted through the arrangement of the main gas pipeline and the main gas combustion-supporting pipeline, meanwhile, the ignition device, the flame detector and the sight glass are matched, and the combustion efficiency is improved. The high-concentration waste gas combustion-supporting device has the advantages that in the chemical incineration process, when the heat value of the waste gas is low, the waste gas can be sprayed into the hearth, when the heat value of the waste gas is high, combustion-supporting air and the high-concentration waste gas can be rapidly mixed, and meanwhile the effect of stabilizing waste gas flames can be achieved.

The invention relates to the technical field of dust combustion, and provides a dust explosion combustion experiment device in order to solve the problems that an existing dust explosion combustion experiment device is insufficient in parameter control precision, and an experiment multi-focus single variable lacks a dust removing mechanism and a safety protection measure, and the dust explosion combustion experiment device comprises an experiment table which is provided with a power chamber and an equipment chamber; one side of the experiment table is hinged with an access door, two sides of the top of the experiment table are provided with conveying main pipes extending to the equipment chamber, a power supply assembly is arranged in the electric power chamber; the protective shell assembly comprises an outer shell and an inner shell; a plurality of exhaust pipes are arranged at the top of the inner shell; the dust generation assembly comprises a plurality of dust storage tanks and an air compressor; the ignition control assembly comprises a pulse igniter, a flame detection feedback circuit and an ignition head; the environment adjusting assembly comprises a box body, an oxygen tank and a humidifier, wherein the oxygen tank and the humidifier are arranged in the box body. The device has the advantages that the dust concentration is accurately regulated and controlled, the environmental parameters are dynamically coupled and controlled, dust residues are rapidly removed, and the experiment safety is high.

The invention discloses a safety and disinfection integrated system based on AI analysis linkage control electromagnetic valves, and relates to the technical field of intelligent fire fighting, and the safety and disinfection integrated system comprises a flame detection camera, a network input and output module, a safety and disinfection integrated machine, an execution module and an acousto-optic alarm bell. According to the security and disinfection integrated system for linkage control of the electromagnetic valve based on AI analysis, the advanced AI visual recognition technology and an electromagnetic valve linkage control mechanism are deeply fused, the efficiency and the precision degree of fire safety management are greatly improved, flame characteristics are precisely recognized by means of a deep learning model, and the safety and disinfection effect is improved. The limitation that traditional fire behavior detection is easily influenced by environmental factors is broken through, the accuracy and reliability of fire behavior recognition are greatly enhanced, the system supports multi-modal data fusion, the confidence coefficient of fire behavior judgment is further improved, adjacent electromagnetic valves can be automatically started according to the position of a fire behavior in the linkage control link, an effective fire extinguishing isolation belt is constructed, and the fire extinguishing efficiency is improved. And fire spreading is effectively restrained.

The invention relates to the field of hearth flame detection, and particularly discloses a hearth flame detection device and method, and the method comprises the following steps: a, capturing a flame image in a hearth through a high-resolution infrared thermal imager, and transmitting the flame image to an image processing unit; step b, executing the following operations in the image processing unit to evaluate and optimize the image quality: removing noise in the image by applying an adaptive median filtering algorithm, wherein the algorithm effect is quantified by calculating a flame image noise removal effect evaluation value Enoise (l); the flame state parameters are monitored in real time, once the parameters are lower than the preset threshold value, the safety pre-warning system is triggered immediately, the real-time monitoring and pre-warning mechanism helps to find and handle potential problems such as instability or extinguishment of flames in time, and therefore safety accidents are effectively prevented.

The invention discloses an integrated pulverized coal burner. A burner head comprises a cooling water jacket, a cooling water partition plate, a fuel gas nozzle, an ignition oxygen nozzle and a pulverized coal nozzle. The burner body comprises a cooling water outer pipe, a cooling water partition pipe, a cooling water inner pipe, a pulverized coal sleeve, an ignition oxygen sleeve, a fuel gas sleeve, an ignition gun mounting pipe and a flame detection pipe; a fuel gas channel is formed between the ignition gun mounting pipe and the fuel gas sleeve, an ignition oxygen channel is formed between the fuel gas sleeve and the ignition oxygen sleeve, a pulverized coal channel is formed between the ignition oxygen sleeve and the pulverized coal sleeve, and a process oxygen channel is formed between the pulverized coal sleeve and the cooling water inner pipe. A cooling water outlet cavity is formed between the cooling water outer pipe and the cooling water partition pipe, and a cooling water inlet cavity is formed between the cooling water inner pipe and the cooling water partition pipe. Various key functions such as flame detection, an ignition gun, start-up fuel gas, start-up oxygen, process pulverized coal and process oxygen are integrated into single equipment, and the device has the advantages that the cost is saved, and the efficiency is improved.

The invention relates to the technical field of computer vision, and discloses a flame recognition and fire extinguishing system for indoor fire fighting, comprising: a binocular infrared imaging module for acquiring a thermal imaging image; the monocular visible light imaging module is used for acquiring a visible light image; the infrared stereoscopic vision module is used for carrying out processing and feature extraction on the thermal imaging images, carrying out feature aggregation to obtain an aggregated feature map, carrying out parallax optimization on the aggregated feature map, and carrying out calculation according to a parallax optimization result to obtain depth information of the thermal imaging images at different visual angles; the multi-mode stereoscopic vision fusion module is used for fusing the thermal imaging image with the depth information and the visible light image; the flame identification and fire behavior prediction module is used for performing flame detection according to the fused image, predicting a bounding box, confidence and a category label of the flame in the fused image, and predicting a future diffusion position of the flame, a diffusion scale of the flame and a diffusion trend of the flame; and the self-adaptive water gun spraying control module controls the water gun to execute spraying operation.

The invention relates to the technical field of image processing, in particular to a flame detection method based on a multi-modal fusion image, which comprises the following steps: acquiring a visible light near-infrared thermal infrared image and matching positions, extracting multi-channel difference to generate a feature set, combining adjacent blocks to judge a direction and a brightness trend, and identifying a temperature rise gathering area. And combining the three types of images to confirm flame characteristics and position a detection result. According to the invention, by collecting visible light, near-infrared and thermal infrared images, cross-modal accurate alignment is realized by using interpolation, and the fusion effect is optimized. The flame area is accurately extracted according to the pixel deviation characteristics, and false alarm and missing alarm are reduced. The positioning precision is improved by combining block boundary tracking and temperature change analysis, and a flame target is effectively recognized in a complex environment. The multi-mode complementation enhances the adaptability of the system, and ensures that the system has high stability and robustness in a dynamic interference environment.

The invention relates to a coal mine flame pseudo fire interference suppression method based on time sequence consistency analysis, and belongs to the technical field of coal mine safety monitoring. According to the method, aiming at the problem of high false alarm rate of a flame detection system caused by false fire source interference such as electric welding sparks and mechanical friction in a mine environment, the flame identification accuracy is improved through time sequence consistency analysis. According to the technical scheme, the method comprises the following steps: acquiring an infrared spectrum flame saliency probability graph of continuous image frames in an underground coal mine; extracting a flame candidate region by adopting Otsu adaptive threshold segmentation in combination with morphological optimization and hole filling; and performing time sequence consistency analysis on the continuous frame candidate regions, and calculating the spatial overlapping degree, the morphological stability and the infrared thermal intensity change trend of adjacent frames. According to the method, the false fire false alarm rate is remarkably reduced, the robustness to complex backgrounds and dynamic illumination is enhanced, the method can be independently integrated to an existing infrared flame detection system or used in cooperation with a deep learning model, and the reliability of mine fire monitoring is improved.

The invention discloses a waste lithium battery electrified recovery system, and relates to the field of waste lithium battery recovery, the waste lithium battery electrified recovery system comprises a flame detection system, a temperature monitoring system, an oxygen content monitoring system and a primary double-shaft shredding machine, and further comprises a nitrogen charging system used for charging nitrogen into the primary double-shaft shredding machine; the explosion venting assembly is arranged on one side of the first-stage double-shaft shredder shell and can discharge impact force gas generated by battery explosion to the outside of the box body; the explosion venting assembly comprises an explosion venting cylinder which is arranged on one side of the first-stage double-shaft shredding machine shell; when a flame detection system detects flames, the parallel four-connecting-rod mechanism is controlled to enable the limiting relieving piece to relieve limiting of the limiting block and the sliding limiting piece, under the guiding effect of the guide column and the elastic force effect of the free spring, the explosion venting frame moves towards the side away from the explosion venting barrel, and therefore the explosion venting frame is opened, and active explosion venting is achieved.

This invention discloses a method and system for detecting wall radiation information through flame penetration. The method determines the radiation wavelength range, and within this range, establishes a wall radiation intensity detection model and a flame radiation intensity detection model, showing the distribution of wall radiation intensity and flame radiation intensity with radiation wavelength. It also obtains the radiation wavelength λ when the wall radiation intensity equals the flame radiation intensity. eq The peak wavelength λ when the wall radiation intensity is at its maximum. max When the radiation wavelength λ eq Less than the peak wavelength λ max At that time, the wavelength for detecting wall radiation information is the peak wavelength λ. max When the radiation wavelength λ eq Greater than the peak wavelength λ max At that time, the wavelength for detecting wall radiation information is the radiation wavelength λ. eq The wall radiation information is detected according to the detection wavelength determined above. The present invention finds a suitable band range for wall radiation detection by accurately quantifying wall radiation and flame radiation, and can obtain wall radiation information by penetrating the flame.

The invention discloses an infrared and ultraviolet flame detector with multiple safety redundancy and an anti-explosion control method, and relates to the technical field of flame detection, and the anti-explosion control method comprises the steps: S1, obtaining the multi-source fire data of a target object through a signal obtaining module; s2, analyzing the multi-source fire data through a dynamic coupling adjustment module; s3, performing time domain association verification on the first monitoring data through a time domain continuity verification module, and generating second monitoring data; and S4, the grading execution module sends an alarm signal according to the second monitoring data and adopts grading execution processing. Through ultraviolet polarization degree and dielectric constant dual-threshold triggering, interference of the metal reflective layer is stripped, so that the smoldering detection rate of a covered scene is increased from less than or equal to 60% to more than or equal to 95%; and moreover, the smoldering authenticity is subjected to cross validation through the bidirectional verification module, so that the problem of missing report of a traditional flame detector caused by a covering effect of'reflective interference 'on'smoldering monitoring' is avoided.

The invention discloses an infrared flame detection method and system based on deep learning, and the method comprises the following steps: obtaining infrared flame image sets in different scenes, and marking the infrared flame image sets; performing Mosaic data enhancement on the marked infrared flame image set to obtain an enhanced image set; improving the YOLOv8 model, and training the improved model based on the enhanced image set to obtain a flame detection model; and detecting and positioning flame by using the flame detection model to obtain an ignition position. On the basis of improving the recognition accuracy and the feature extraction performance, the infrared flame image target can be rapidly and accurately detected and positioned, and finally, after the target flame is detected, the fire water monitor device is used for extinguishing the fire in the target area.

The utility model relates to the technical field of household alarm, and discloses a gas cooker alarm device which comprises a flame detection mechanism arranged at a gas cooker end and an alarm mechanism arranged at a door end, the flame detection mechanism is used for detecting whether a gas cooker works or not, and the alarm mechanism is used for detecting whether the gas cooker works or not. The alarm mechanism comprises a going-out detection unit, an alarm unit, a transmission unit and a main control unit, the going-out detection unit, the alarm unit and the transmission unit are all electrically connected with the main control unit, the going-out detection unit is used for detecting whether a user goes out or not, and the flame detection mechanism is in communication connection with the main control unit through the transmission unit. And under the condition that the flame detection mechanism detects that the gas cooker is working and the going-out detection unit detects that a user goes out, the main control unit controls the alarm unit to give an alarm. According to the utility model, the problem that in the prior art, when a user goes out, the alarm device is difficult to trigger the alarm because the alarm condition is not met, and after the alarm is triggered, the user is not at home, so that the loss which is difficult to redeem is caused can be solved.

The invention discloses a liquid rocket engine combustion tail flame detection device and method, the device comprises an air inlet pipe, an exhaust pipe, a gas chamber packaging pipe and an intelligent controller, a plurality of detection devices such as a temperature sensor and a spectrum analyzer are arranged in the gas chamber packaging pipe along the airflow direction, and the detection devices are connected with the intelligent controller to form a data transmission channel; the detection method comprises the steps that the device is initialized, and all devices are calibrated and debugged; after the engine is started, tail flame parameters are detected in real time and transmitted to the intelligent controller; the controller performs preprocessing and deep analysis on the data; comparing an analysis result with a preset normal parameter range to judge the working state of the tail flame; and if normal, the data are backed up periodically, and the long-term trend of the performance of the tail flame can also be predicted. According to the method, the tail flame parameters can be comprehensively and accurately detected in real time, intelligent analysis and abnormal early warning are achieved, and the safety and reliability of an engine are improved.

The invention relates to a gas equipment detection circuit, gas equipment, a gas equipment detection method and device, computer equipment, a storage medium and a computer program product. The gas equipment detection circuit comprises a fire detection circuit arranged in a combustion chamber of gas equipment and comprising a magnetic sensitive component; the first end of the magnetic sensitive component is connected with a power supply and the second end is grounded for detecting a magnetic field generated by flame in the combustion chamber; the amplification circuit comprises a positive-phase input end, a negative-phase input end and an amplification signal output end, and is used for amplifying voltage signals input into the positive-phase input end and the negative-phase input end; the positive phase input end is connected with the first end of the magnetic sensitive component, the negative phase input end is grounded, and the amplified signal output end is used for outputting an amplified voltage signal; and the control circuit is connected with the amplified signal output end and is used for feeding back a flame detection result according to the received amplified voltage signal. By adopting the gas equipment detection circuit, the state of flame in the combustion chamber can be accurately detected.

The embodiment of the invention provides a flame detection method and device based on image recognition, electronic equipment and a medium, and relates to the technical field of computer vision. The method comprises the steps of obtaining a target video; performing video frame extraction on the target video to obtain at least two target video frames; according to the color feature of each target video frame and a preset target flame color rule, performing flame color region detection on each target video frame to obtain a flame color region; performing moving area detection on any two adjacent target video frames to obtain a moving area of each target video frame; wherein the adjacent target video frames are used for representing the adjacent video frames in the time sequence; and performing overlapped region screening on the flame color region and the moving region of each target video frame to obtain a target flame region. According to the embodiment of the invention, the accuracy and reliability of flame detection can be improved.

The utility model discloses an oxyhydrogen burner for SOEC, which comprises a cylindrical steam generating device, and the outer wall of the steam generating device is connected with a pure water inlet pipeline and a steam outlet pipeline; a cylinder is sleeved with the steam generating device, the bottom of the cylinder is close to the bottom of the steam generating device, the lowest end of the cylinder is connected with an outlet pipeline, and the outlet pipeline protrudes out of the lower end of the steam generating device. The outer wall of the cylinder body is connected with a flame detection opening; a hydrogen inlet pipeline and an oxygen inlet pipeline are inserted above the outer wall of the barrel; a flame holder, a catalytic grating and a premixing device are arranged in the barrel. Fuel and air can be evenly mixed, combustion is sufficient, and the heat utilization rate is increased.

A flame detection system is disclosed. The flame detection system includes: an image capture unit configured to capture an image of a field of view (FOV); a flame detector configured to capture a first set of spectral band data and a second set of spectral band data; and a processor configured to classify the one or more images and the first set of spectral band data, assign a weight to each of the second set of spectral band data, receive a flame count for each of the second set of spectral band data, determine a weighted count for each of the flame counts, and determine a weighted count for each of the flame counts. A number of weighted counts exceeding its respective threshold is determined, and a flame is determined to be present within the FOV when the number is greater than or equal to a number threshold, and a flame is determined to be absent within the FOV when the number is less than the number threshold.

The invention discloses a flame detection probe cooling system, method and device, electronic equipment and a storage medium, and relates to the technical field of flame detection probe cooling, and the main technical scheme comprises a cooling module, a control module and a temperature monitoring module; the cooling module is connected with the control module, the control module is connected with the temperature monitoring module, and the temperature monitoring module is placed around a flame detection probe; the cooling module is used for generating cooling air so as to cool the flame detection probe by using the cooling air; the temperature monitoring module is used for monitoring the temperature of the flame detection probe and transmitting the temperature to the control module; and the control module adjusts the air volume of the cooling air generated by the cooling module based on the temperature. Compared with the prior art, according to the embodiment of the invention, the air volume of the cooling air generated by the cooling module in the flame detection probe cooling system is adjusted according to the temperature change of the flame detection probe, so that the cooling cost of the flame detection probe is reduced.

The invention discloses a flame detection method based on a visual large model. The method comprises the following steps: extracting a multi-scale context semantic enhancement feature map through a feature extractor in combination with a context aggregation module; the fire source positioning accuracy is improved by fusing multi-scale information through cross-layer feature two-way convolution transmission by using a two-way feature interaction module; sensing geometric features of a fire source through a dynamic adaptive point representation module, and screening samples to suppress interference in combination with density clustering and spatial constraint; a multi-task loss function containing classification, regression (generalized IoU and center distance IoU loss), multi-scale supervision and spatial constraint is constructed, and a model is trained in combination with an AdamW optimizer and a dynamic attenuation learning rate strategy. According to the method, small-scale fire source features can be effectively captured, complex background false detection is reduced, the detection precision on multiple data sets is superior to that of a traditional method, and the method is suitable for real-time monitoring and fire early warning of scenes such as hydraulic power plants.

The invention discloses a dual-band flame detection method and system based on visible light and near-infrared images. The dual-band flame detection method comprises the following steps: collecting a dual-band flame image, carrying out joint calibration on a visible light camera and infrared imaging equipment, and obtaining internal and external parameters of the visible light camera and the infrared imaging equipment; performing horizontal alignment on the collected visible light image and the infrared image by using distortion correction and epipolar correction; marking the visible light image and the near-infrared image after horizontal alignment, and constructing a dual-band flame data set; constructing a dual-band flame detection model based on visible light and near-infrared images; and training a dual-band flame detection model by using the dual-band flame data set, and monitoring flame information in the scene in real time by the trained dual-band flame detection model. Compared with an existing flame detection model, the detection precision is higher, and the anti-interference capability is higher.

The invention belongs to the technical field of boiler maintenance, and particularly relates to a boiler four-pipe foreign matter blockage early warning method and system. The method comprises the steps of obtaining boiler four-tube foreign matter blockage historical data; collecting boiler four-tube real-time state data of a reheater, an induced draft fan and a flame detection cooling fan; a fault mode-influence-risk mapping relation is established by adopting FMEA based on historical data, and a risk assessment framework is formed; generating an initial risk level; training a machine learning algorithm model by adopting historical data and generating a blockage probability; and dynamically adjusting the initial risk level according to the blockage probability, generating a final risk level and outputting a corresponding maintenance strategy. The system comprises a data acquisition module, an equipment monitoring module, a risk division module, a risk assessment module, a jam prediction module and a decision module. According to the scheme, the FMEA and the machine learning algorithm are fused, early warning, accurate risk assessment and dynamic maintenance decision making of boiler foreign matter blockage are achieved, and boiler operation safety and maintenance efficiency are improved.