Intelligent identification method for abnormal temperature of coal fire based on multi-source time-series thermal infrared satellite images

Abstract

The application discloses a coal fire abnormal temperature intelligent identification method based on multi-source time sequence thermal infrared satellite images, selects multi-source thermal infrared and visible light data of various satellites in recent period, reduces the influence of cloud layer on telemetry results, adopts a specific correction model in combination with digital elevation data to perform terrain correction on the thermal infrared data of the satellite, weakens the influence of the terrain effect in the thermal infrared satellite image on temperature values in the identification of the coalfield fire area, then trains a YOLOv5 model to detect the coalfield area and the non-coalfield area, simultaneously fuses visible light and thermal infrared satellite image data, weakens the influence of human activities on the identification of the coalfield fire area in the thermal infrared satellite image, finally extracts time sequence change characteristics of the temperature data of the satellite image for at least one year, and finally effectively improves the identification efficiency and accuracy of the thermal anomaly of the thermal infrared satellite image through a U-Net-CBAM model with a fusion attention mechanism, so that precise identification and positioning of the abnormal temperature area of the coalfield in a complex terrain are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of coalfield fire area identification, in particular to a coal fire abnormal temperature intelligent identification method based on multi-source time-series thermal infrared satellite images. BACKGROUND

[0002] Coal fire combustion will produce and release CO, C2H2, CH4 and other toxic and harmful gases and emit a large amount of heat into the atmosphere, while forming a large area of burned-out area and causing ground subsidence, which seriously endangers the health of local residents and the ecological environment. Coal fires are mainly distributed in some complex terrain and rarely visited areas. Conventional underground coal fire detection technology is limited by detection equipment and personnel costs, and it is difficult to find burning coalfield fire areas in time and effectively manage them. Remote sensing detection method can use the high temperature, smoke and dust and other associated characteristics of coal fire to remotely monitor the entire range of coalfield in unpopulated areas, so as to achieve the purpose of early detection and early management of coal fire.

[0003] The existing remote sensing detection technology mainly uses thermal infrared remote sensing to detect coalfield fire areas, but this detection technology is easily affected by factors such as cloud cover, terrain, specific heat capacity of surface deposits, human activities and seasonal changes, which makes it prone to misjudgment of non-fire areas when extracting thermal anomalies. Especially in complex terrain areas, the misjudgment probability is higher due to the influence of terrain.

[0004] Therefore, how to provide a new coal fire abnormal temperature intelligent identification method can effectively reduce the influence of complex terrain on identification, so as to realize accurate identification and positioning of abnormal temperature areas of coalfield in complex terrain, which is the research direction of the present application. SUMMARY

[0005] In view of the problems existing in the prior art, the present application provides a coal fire abnormal temperature intelligent identification method based on multi-source time-series thermal infrared satellite images, which combines visible light images, thermal infrared satellite images, terrain correction, ground feature mask and time-series temperature change characteristics, can effectively reduce the influence of complex terrain on identification, so as to realize accurate identification and positioning of abnormal temperature areas of coalfield in complex terrain.

[0006] In order to achieve the above purpose, the technical scheme adopted by the present application is as follows: a coal fire abnormal temperature intelligent identification method based on multi-source time-series thermal infrared satellite images, comprising the following steps:

[0007] Step 1: Determine the geographical location information of the coalfield area to be monitored;

[0008] Step 2: According to the geographical location of the coalfield area, obtain the multi-source thermal infrared satellite image data and visible light image data of the area within a certain period of time before the satellite;

[0009] Step three, temperature inversion is performed on the obtained multi-source thermal infrared satellite image data to obtain the surface inversion temperature at different positions in the coalfield area, and a thermal infrared temperature image is formed.

[0010] Step four, according to the geographical position of the coalfield area, corresponding digital elevation data is obtained, then the thermal infrared temperature image is terrain corrected in combination with the digital elevation data, and according to a preset target detection label, a corresponding label is identified from the visible light image.

[0011] Step five, the corrected thermal infrared temperature image of the coalfield area and the coalfield area identified from the visible light image are fused to obtain a corrected thermal infrared temperature image.

[0012] Step six, the temperature time sequence change characteristics of the corrected thermal infrared temperature image are extracted.

[0013] Step seven, the temperature time sequence change characteristics obtained in step six are used to train an abnormal temperature identification model, and after completion, the model is used to identify the abnormal temperature area in the coalfield area.

[0014] Further, the geographical position information in step one includes the central position longitude and latitude coordinates of the coalfield area and the boundary range longitude and latitude coordinates of the coalfield area.

[0015] Further, the multi-source thermal infrared and visible light image data used in step two includes Landsat satellite thermal infrared and visible light image data, SDGSAT-1 satellite thermal infrared and visible light image data, and ZY1F satellite thermal infrared and visible light image data; and the data acquisition time period is at least one year of thermal infrared data acquired from the current time.

[0016] Further, the thermal infrared temperature inversion function is used to perform temperature inversion on the multi-source thermal infrared satellite image data, specifically:

[0017] The Landsat satellite thermal infrared temperature inversion function is:

[0018]

[0019] Where T L-LST is the Landsat satellite thermal infrared surface inversion temperature, T L-DN is the detection value of the Landsat satellite thermal infrared data.

[0020] The SDGSAT-1 satellite three-band (i.e. 8~10.5, 10.3~11.3, 11.5~12.5 μm) thermal infrared temperature inversion function is:

[0021]

[0022]

[0023]

[0024]

[0025]

[0026] where d ra is the radiance, d DN is the reflectivity value of the SDGSAT-1 satellite thermal infrared data, a GAIN is the gain value of the radiometric calibration, b BIAS is the offset value of the radiometric calibration, c BG is the cold space background correction coefficient; T BT is the brightness temperature, k1 and k2 are the radiometric calibration constants; h represents the Planck constant, c is the speed of light, k is the Boltzmann constant, and λ ef is the effective wavelength of the sensor; T SDGSAT-1-LST is the SDGSAT-1 satellite thermal infrared surface retrieval temperature, T1, T2, and T3 are the brightness temperatures of the three thermal infrared bands, and a1 and b1 are empirical coefficients.

[0027] The ZY1F satellite thermal infrared temperature retrieval function is:

[0028]

[0029]

[0030] where E em is the surface emissivity; T BT is the brightness temperature; T ZY1F-LST is the ZY1F satellite thermal infrared surface retrieval temperature; and λ Wl is the center wavelength of the thermal infrared band.

[0031] Further, the specific process of the topographic correction in step four is:

[0032] The thermal infrared temperature image is topographically corrected by using digital elevation data in combination with the C topographic correction model and the SCS+C topographic correction model:

[0033] The C topographic correction model is:

[0034]

[0035] where L cor is the reflectivity of the thermal infrared temperature image after topographic correction; L obs is the reflectivity of the original thermal infrared temperature image; θ i and θ zrespectively the solar zenith angle and the incidence angle determined according to digital elevation data; C is an empirical coefficient;

[0036] The SCS+C terrain correction model:

[0037]

[0038] where θ v is the solar reflection angle determined according to digital elevation data.

[0039] Further, the target detection label in the fourth step includes a coalfield area and a non-coalfield area such as a town, water area, etc., a preset target detection label is input into a YOLOv5 target detection model for training, and after completion, the coalfield area and the non-coalfield area are identified from the visible light image.

[0040] Further, in the fifth step, the coalfield area identified from the visible light image is fused with the thermal infrared temperature image by using a spatial matching method to obtain a corrected thermal infrared temperature image. The specific process is as follows: the coalfield area identified from the visible light image is corresponded to the corresponding area of the thermal infrared temperature image, and the area is extracted to be the corrected thermal infrared temperature image (thermal infrared temperature image of the coalfield area).

[0041] Further, the abnormal temperature identification model in the seventh step is a U-Net-CBAM model fused with an attention mechanism, and the U-Net model architecture is as follows:

[0042] ① Down-sampling;

[0043] Convolution layer:

[0044]

[0045] where the input temperature time series change feature map , H is the height, W is the width, and C is the number of channels; the convolution kernel , k is the kernel size, D is the output channel number; the bias ; the convolution operation is represented by *; σ a is the ReLU activation function, ; the output feature size is ; the maximum pooling layer: ;

[0046] ② Up-sampling;

[0047] Transposed convolution: ;

[0048] where the input temperature time series change feature map ; the transposed convolution kernel ; the output feature map size is expanded to 2h × 2w; the bias value is b;

[0049] skip connection: ;

[0050] concatenate the feature map of the first layer of the encoder with the feature map of the corresponding layer of the decoder;

[0051] ③, joint loss function;

[0052] weighted cross-entropy loss:

[0053]

[0054] Dice loss:

[0055]

[0056] joint loss function: ;

[0057] where w is the weight of the thermal anomaly pixel, is the true label, L Dice is the predicted value, and a is the hyperparameter adjustment weight;

[0058] ④, back propagation and optimization; ;

[0059] where θ is the gradient descent optimization parameter, η is the learning rate, is the gradient of the loss function to the parameter.

[0060] Further, the CBAM module is embedded between the encoder, the decoder and the skip connection layer of the U-Net model architecture, and the CBAM module includes a channel attention module and a spatial attention module;

[0061] wherein the channel attention module:

[0062]

[0063] wherein represents the input temperature time series variation feature map; and respectively, the output dimension is ; MLP is a multilayer perceptron, including two fully connected layers with ReLU activation function in between; σ a is a Sigmoid activation function, which normalizes the output;

[0064] spatial attention module:

[0065]

[0066] wherein The results after average pooling and maximum pooling are spliced into a two-channel feature map in the channel dimension; Convolution operation using 7x7 convolution kernel; sigma a Sigmoid function, output spatial attention map .

[0067] Further, it further comprises the eighth step of: according to the abnormal temperature area identified in step seven, collecting temperature data of the abnormal temperature area on site and forming a temperature three-dimensional distribution map of the area, verifying the abnormal temperature area on site.

[0068] Compared with the prior art, the present application adopts the combination of visible light image, thermal infrared satellite image, terrain correction, ground object mask and time series temperature change characteristics, and has the following advantages:

[0069] 1. For the problem that the coalfield fire area in remote sensing image is affected by cloud layer and terrain, and it is difficult to realize accurate detection of coalfield fire area temperature anomaly; in the present application, according to the coalfield geographic location information, the multi-source thermal infrared and visible light image data of various satellites in recent years are selected, the influence of cloud layer on the telemetry result is reduced. In addition, C correction, SCS+C correction model is used, combined with digital elevation data, the terrain correction is carried out on the satellite thermal infrared data, the influence of terrain effect in thermal infrared satellite image on the temperature value in the identification of coalfield fire area is weakened, and the cleanliness of the input thermal infrared data of the coalfield anomaly identification model is improved.

[0070] 2. For the problem of misjudgment of non-mining area temperature anomaly area caused by human activities, the YOLOv5 model is trained based on visible light satellite image data to detect coalfield area and non-coalfield area; at the same time, the influence of human activities on the identification of coalfield fire area in thermal infrared satellite image is weakened by further fusing visible light and thermal infrared satellite image data, and the cleanliness of the input thermal infrared data of the coalfield fire area thermal anomaly identification model is improved.

[0071] 3. For the problem of low thermal anomaly recognition efficiency and recognition lag of coalfield fire area in remote sensing image, the thermal infrared satellite image after terrain correction and target recognition is used as the source data, the time series change characteristics of the satellite image temperature data for at least 1 year are extracted, the U-Net-CBAM model with attention mechanism is fused, combined with the above processing of infrared data cleanliness, the recognition efficiency and accuracy of thermal infrared satellite image thermal anomaly are finally effectively improved, and the accurate identification and positioning of the abnormal temperature area of the coalfield in the complex terrain whole region are realized. BRIEF DESCRIPTION OF DRAWINGS

[0072] Figure 1 It is the coalfield temperature anomaly area identification and verification flowchart of the embodiment of the present application;

[0073] Figure 2is a temperature inversion result map of a satellite thermal infrared satellite image of a coalfield mining area in an embodiment of the present application;

[0074] Figure 3 is a C, SCS+C terrain correction model training flowchart in an embodiment of the present application;

[0075] Figure 4 is a terrain correction comparison chart based on Landsat satellite thermal infrared satellite image data in an embodiment of the present application;

[0076] Among them, (a) is the original temperature inversion chart, (b) is the C terrain correction effect chart, (c) is the SCS+C terrain correction effect chart;

[0077] Figure 5 is a thermal infrared satellite image feature mask flowchart based on the YOLOv5 model in an embodiment of the present application;

[0078] Figure 6 is a time sequence change feature chart of the average temperature of the thermal infrared satellite image of a certain mining area after processing in an embodiment of the present application;

[0079] Figure 7 is a U-Net model structure chart in an embodiment of the present application;

[0080] Figure 8 is a CBAM module structure chart in an embodiment of the present application;

[0081] Figure 9 is a comparison chart of known abnormal temperature areas of a coalfield and abnormal temperature identification results of an embodiment of the present application;

[0082] Figure 10 is an abnormal temperature distribution chart of a part of the coalfield in the Xinjiang Uygur Autonomous Region identified by the embodiment of the present application;

[0083] Figure 11 is an abnormal temperature area temperature distribution chart obtained by using the unmanned aerial vehicle on-site verification method;

[0084] Figure 12 is an abnormal temperature area temperature distribution chart obtained by using the handheld thermal infrared instrument on-site verification method. DETAILED DESCRIPTION

[0085] The present application will be further described below.

[0086] As shown in Figure 1 , the present embodiment comprises the following steps:

[0087] Step one, determine the geographical position information of the required monitoring coalfield area, including the central position latitude and longitude coordinates of the coalfield area and the boundary range latitude and longitude coordinates of the coalfield area.

[0088] Step two, according to the geographical location of the coalfield region, satellite multi-source thermal infrared satellite image data and visible light image data of the region in the previous period are acquired; the multi-source thermal infrared and visible light image data used include Landsat satellite thermal infrared and visible light image data, SDGSAT-1 satellite thermal infrared and visible light image data and ZY1F satellite thermal infrared and visible light image data; and the data acquisition time period is at least one year of thermal infrared data from the current time.

[0089] Step three, using a thermal infrared temperature inversion function, the multi-source thermal infrared satellite image data is temperature-inverted to form a thermal infrared temperature image as shown in Figure 2 , specifically:

[0090] The Landsat satellite thermal infrared temperature inversion function is:

[0091]

[0092] Wherein T L-LST is the Landsat satellite thermal infrared surface inversion temperature, T L-DN is the detection value of the Landsat satellite thermal infrared data;

[0093] The SDGSAT-1 satellite three-band (i.e. 8~10.5, 10.3~11.3, 11.5~12.5 μm) thermal infrared temperature inversion function is:

[0094]

[0095]

[0096]

[0097]

[0098]

[0099] Wherein d ra is the radiation brightness, d DN is the reflectivity value of the SDGSAT-1 satellite thermal infrared data, a GAIN is the gain value of radiation calibration, b BIAS is the offset value of radiation calibration, c BG is the cold space background correction coefficient; T BT is the brightness temperature, k1 and k2 are radiation calibration constants; h represents the Planck constant, c is the speed of light, k is the Boltzmann constant, λ ef is the effective wavelength of the sensor; T SDGSAT-1-LSTT1, T2, T3 are brightness temperatures of three thermal infrared bands, a1, b1 are empirical coefficients;

[0100] The thermal infrared temperature inversion function of ZY1F satellite is:

[0101]

[0102]

[0103] wherein E em is the surface emissivity, [0.9, 0.99]; T BT is the brightness temperature; T ZY1F-LST is the thermal infrared surface inversion temperature of ZY1F satellite; λ Wl is the center wavelength of the thermal infrared band.

[0104] Step four, according to the geographical position of the coalfield region, the corresponding digital elevation data is obtained, the digital elevation data is resampled to match the satellite image resolution; then the thermal infrared temperature image is terrain corrected combined with the digital elevation data as shown in Figure 3 , the specific process is: using the digital elevation data combined with the C terrain correction model and the SCS+C terrain correction model to perform terrain correction on the thermal infrared temperature image:

[0105] The C terrain correction model:

[0106]

[0107] wherein L cor is the reflectivity of the thermal infrared temperature image after terrain correction; L obs is the reflectivity of the original thermal infrared temperature image; θ i and θ z are the solar zenith angle and the incidence angle determined according to the digital elevation data, respectively; C is an empirical coefficient;

[0108] The SCS+C terrain correction model:

[0109]

[0110] wherein θ v is the solar reflection angle determined according to the digital elevation data. As an example, Figure 4 is a terrain correction comparison chart based on Landsat satellite thermal infrared satellite image data.

[0111] At the same time, according to the preset target detection label, it includes a coalfield region and a non-coalfield region such as a town, a water area and the like, as Figure 5As shown, the preset target detection label is input into the YOLOv5 target detection model for training, and after completion, the coalfield area and non-coalfield area are recognized from the visible light image.

[0112] Step five, the coalfield area corrected from the thermal infrared temperature image and the coalfield area recognized from the visible light image are fused by using a spatial matching method, and the specific process is as follows: the coalfield area recognized from the visible light image is corresponded to the corresponding area of the thermal infrared temperature image, and the area is extracted as the corrected thermal infrared temperature image (thermal infrared temperature image of the coalfield area).

[0113] Step six, the temperature time series change characteristics of the corrected thermal infrared temperature image for at least one year are extracted, as shown in Figure 6

[0114] Step seven, the temperature time series change characteristics obtained in step six are used to train the abnormal temperature recognition model, and after completion, the model is used to recognize the abnormal temperature area in the coalfield area, and the specific process is as follows: the abnormal temperature recognition model is a U-Net-CBAM model fused with attention mechanism, as shown in Figure 7 The U-Net model architecture is as follows:

[0115] ① Down-sampling;

[0116] Convolution layer:

[0117]

[0118] Wherein the input temperature time series change feature map , H is the height, W is the width, and C is the number of channels; the convolution kernel , k is the kernel size, D is the output channel number; the bias ; * represents convolution operation; σ a is the ReLU activation function, ; the output feature size is ; the maximum pooling layer: ; the spatial dimension is compressed by down-sampling, and the pooling window is 2x2 and the step is 2;

[0119] ② Up-sampling;

[0120] Transposed convolution: ;

[0121] Wherein the input temperature time series change feature map ; the transposed convolution kernel ; the output feature map size is expanded to 2h x 2w; the bias value is b;

[0122] Skip connection: ;

[0123] ​The feature map of the first layer of the encoder is spliced with the feature map of the corresponding layer of the decoder;

[0124] ③, joint loss function;

[0125] Weighted cross-entropy loss:

[0126]

[0127] Dice loss:

[0128]

[0129] Joint loss function: ;

[0130] where w is the weight of the thermal anomaly pixel, is the true label, L Dice is the predicted value, and a is the hyperparameter adjustment weight;

[0131] ④, back propagation and optimization; ;

[0132] where θ is the gradient descent optimization parameter, η is the learning rate, is the gradient of the loss function to the parameter.

[0133] As Figure 8 shown, the CBAM module is embedded between the encoder, decoder and skip connection layer of the U-Net model architecture, and the CBAM module includes a channel attention module and a spatial attention module;

[0134] wherein the channel attention module:

[0135]

[0136] wherein represents the input temperature time series variation feature map; and are global average pooling and global maximum pooling in the spatial dimension, respectively, and the output dimension is ; MLP is a multilayer perceptron, including two fully connected layers with ReLU activation functions in between; σ a is a Sigmoid activation function, which normalizes the output ([0, 1]);

[0137] Spatial attention module:

[0138]

[0139] wherein represents the result of splicing the average pooling and maximum pooling in the channel dimension into a two-channel feature map; For convolution operations using 7×7 kernels; σ a The sigmoid function outputs a spatial attention map. .

[0140] Step 8: Based on the abnormal temperature areas identified in Step 7, such as... Figure 9 The figure shows a comparison with known coalfield anomalous temperature areas. As can be seen from the figure, the method of the present invention can accurately locate known anomalous temperature areas.

[0141] Next, the method of this invention is used to locate abnormal temperature areas in coalfields in the Xinjiang Uygur Autonomous Region, such as... Figure 10 As shown; In order to ensure the positioning accuracy of the present invention, temperature data of any abnormal temperature area is collected on-site for verification. Specifically, based on the exploration range and terrain undulation conditions of the coalfield, the on-site temperature image data collection methods are divided into two types: drones and handheld thermal infrared meters.

[0142] The on-site verification method using drones involves: in coalfield fire areas with relatively gentle terrain and a large monitoring range, drones acquire thermal infrared satellite imagery data of the coalfield. During backend processing, the EXIF ​​and RGB information from the original imagery data are merged into the converted thermal infrared temperature imagery data, increasing the number of feature points in the drone's thermal infrared temperature imagery, and stitching together a temperature distribution map of the abnormal temperature area; for example... Figure 11 This is a temperature distribution map of the abnormal temperature area verified by drone.

[0143] The on-site verification method for handheld thermal infrared meters is as follows: In coalfield fire areas with significant terrain undulations, a handheld thermal infrared meter is used to remotely acquire temperature distribution data of abnormal temperature areas at the detection slope location. For example... Figure 12 This is a temperature distribution map of the abnormal temperature area verified using a handheld thermal infrared meter.

[0144] like Figures 9 to 12 Multiple verifications have shown that the method of the present invention can effectively reduce the impact of complex terrain on identification, thereby achieving accurate identification and location of abnormal temperature areas in coalfields with complex terrain.

[0145] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for intelligent identification of coal fire anomaly temperature based on multi-source time-series thermal infrared satellite imagery, characterized in that, Includes the following steps: Step 1: Determine the geographical location of the coalfield area to be monitored; Step 2: Based on the geographical location of the coalfield area, acquire multi-source thermal infrared satellite imagery and visible light imagery data of the area over a previous period using satellites; Step 3: Perform temperature inversion on the acquired multi-source thermal infrared satellite image data to obtain the surface inversion temperature at different locations in the coalfield area, forming a thermal infrared temperature image; Step 4: Obtain the corresponding digital elevation data based on the geographical location of the coalfield area. Then, perform terrain correction on the thermal infrared temperature image based on the digital elevation data. At the same time, input the preset target detection labels, which include coalfield areas and non-coalfield areas, into the YOLOv5 target detection model for training. After completion, the coalfield area and non-coalfield area can be identified from the visible light image. Step 5: Fuse the corrected thermal infrared temperature image of the coalfield area with the coalfield area identified from the visible light image to obtain the corrected thermal infrared temperature image. Step 6: Extract the time-series temperature variation features of the corrected thermal infrared temperature image; Step 7: Use the temperature time series change features obtained in Step 6 to train the abnormal temperature identification model. After completion, use the model to identify abnormal temperature areas in the coalfield area.

2. The intelligent identification method for coal fire anomaly temperature based on multi-source time-series thermal infrared satellite imagery according to claim 1, characterized in that, The geographical location information in step one includes the latitude and longitude coordinates of the center location of the coalfield area and the latitude and longitude coordinates of the boundary range of the coalfield area.

3. The intelligent identification method for coal fire anomaly temperature based on multi-source time-series thermal infrared satellite imagery according to claim 1, characterized in that, The multi-source thermal infrared and visible light image data used in step two include Landsat satellite thermal infrared and visible light image data, SDGSAT-1 satellite thermal infrared and visible light image data, and ZY1F satellite thermal infrared and visible light image data; and the data acquisition period is at least one year of thermal infrared data acquired from the current time backward.

4. The intelligent identification method for coal fire anomaly temperature based on multi-source time-series thermal infrared satellite imagery according to claim 3, characterized in that, Step three employs a thermal infrared temperature inversion function to perform temperature inversion on multi-source thermal infrared satellite image data, specifically as follows: The Landsat satellite thermal infrared temperature inversion function is: Where T L-LST For the Landsat satellite thermal infrared surface inversion temperature, T L-DN These are the detected values ​​from Landsat satellite thermal infrared data; The SDGSAT-1 satellite's three-band thermal infrared temperature inversion function is as follows: Where d ra For radiance, d DN a represents the reflectance value of thermal infrared data from the SDGSAT-1 satellite. GAIN b is the gain value for radiation calibration. BIAS c is the offset value for radiometric calibration. BG T is the cold air background correction factor; BT Where k is the brightness temperature, k1 and k2 are radiation calibration constants; h represents Planck's constant, c is the speed of light, k is Boltzmann's constant, and λ is the light intensity temperature. ef T is the effective wavelength of the sensor. SDGSAT-1-LST T1 represents the surface temperature retrieved by the SDGSAT-1 satellite using thermal infrared radiation. T1, T2, and T3 are the brightness temperatures of the three thermal infrared bands, and a1 and b1 are empirical coefficients. The thermal infrared temperature inversion function for the ZY1F satellite is: Where E em For surface emissivity; T BT For brightness temperature; T ZY1F-LST For the thermal infrared surface temperature retrieved by the ZY1F satellite; λ Wl It is the center wavelength of the thermal infrared band.

5. The intelligent identification method for coal fire anomaly temperature based on multi-source time-series thermal infrared satellite imagery according to claim 1, characterized in that, The specific process of terrain correction in step four is as follows: Topographic correction of thermal infrared temperature images was performed using digital elevation data combined with C-type and SCS+C-type topographic correction models. The C-type terrain correction model: Where L cor The reflectance of the topographically corrected thermal infrared temperature image; L obs The reflectance of the original thermal infrared temperature image; θ i and θ z These are the solar zenith angle and the angle of incidence determined based on digital elevation data, respectively; C is an empirical coefficient. The SCS+C terrain correction model: Where θ v The solar reflection angle is determined based on digital elevation data.

6. The intelligent identification method for coal fire anomaly temperature based on multi-source time-series thermal infrared satellite imagery according to claim 1, characterized in that, In step five, a spatial matching method is used to fuse the coalfield area identified by the visible light image with the thermal infrared temperature image to obtain a corrected thermal infrared temperature image.

7. The intelligent identification method for coal fire anomaly temperature based on multi-source time-series thermal infrared satellite imagery according to claim 1, characterized in that, The abnormal temperature identification model in step seven is the U-Net-CBAM model that integrates an attention mechanism. The U-Net model architecture is as follows: ① Downsampling; Convolutional layers: The input temperature time-series variation feature map Where H is the height, W is the width, and C is the number of channels; convolution kernel Where k is the core size, D is the number of output channels; bias ; * indicates convolution operation; σ a It is the ReLU activation function. ; Output feature size is Max pooling layer: ; ② Upsampling; Transposed convolution: ; The input temperature time-series variation feature map Transposed convolution kernel ; The output feature map size is increased to 2h × 2w; The bias value is b; Jump links: ; The feature map of the encoder's first layer is concatenated with the feature map of the corresponding layer of the decoder; ③ Joint loss function; Weighted cross-entropy loss: Dice loss: Joint loss function: ; Where w is the weight of the thermally anomalous pixel. For real labels, L Dice The predicted value is α, where α is the hyperparameter adjustment weight. ④ Backpropagation and optimization; ; Where θ is the gradient descent optimization parameter, and η is the learning rate. This represents the gradient of the loss function with respect to the parameters.

8. The intelligent identification method for coal fire anomaly temperature based on multi-source time-series thermal infrared satellite imagery according to claim 7, characterized in that, The CBAM module is embedded between the encoder, decoder, and skip connection layers of the U-Net model architecture. The CBAM module includes a channel attention module and a spatial attention module. Among them, the channel attention module: in This represents a time-series temperature variation feature map of the input. and These represent global average pooling and global max pooling in the spatial dimension, respectively, with an output dimension of... MLP stands for Multilayer Perceptron, which contains two fully connected layers with a ReLU activation function in between; σ a The Sigmoid activation function normalizes the output. Spatial attention module: in This means concatenating the results of average pooling and max pooling into a two-channel feature map along the channel dimension; For convolution operations using 7×7 kernels; σ a The sigmoid function outputs a spatial attention map. .

9. The intelligent identification method for coal fire anomaly temperature based on multi-source time-series thermal infrared satellite imagery according to claim 1, characterized in that, It also includes step eight: based on the abnormal temperature area identified in step seven, temperature data of the abnormal temperature area is collected on-site and a three-dimensional temperature distribution map of the area is generated to verify the abnormal temperature area on-site.

Images