Forest Fire Detection and Location Method and Device Based on Mid-Wave Infrared Image

Abstract

This invention provides a method and apparatus for forest fire detection and location based on mid-wave infrared (MBIR) images. The method includes: processing the acquired MBIR image using a full-image filtering algorithm to obtain a coarse location of the fire target in the MBIR image; determining a window image of the fire target based on the coarse location; processing the window image of the fire target using a window filtering algorithm to obtain a fine location of the fire target in the MBIR image; performing geometric correction on the geometric distortion of the MBIR image; and obtaining the spatial three-dimensional location result of the fire target based on the fine location and the geometrically corrected MBIR image. This solution not only has low computational complexity, high efficiency, and high accuracy, but also meets the requirements for on-orbit real-time processing.

Description

Technical Field

[0001] This invention relates to the field of image detection technology, and in particular to a method and apparatus for detecting and locating forest fires based on mid-wave infrared images. Background Technology

[0002] Traditional onboard forest fire detection equipment lacks real-time on-orbit data processing capabilities, only performing onboard image capture and data downlink functions, leaving forest fire detection and location to ground equipment. The forest fire detection sensor is the first onboard device capable of real-time on-orbit forest fire detection, employing a decision-level fusion approach based on multispectral images to achieve forest fire detection and location. How to achieve more accurate forest fire detection and location when using this decision-level fusion approach is a pressing issue that needs to be addressed. Summary of the Invention

[0003] This invention provides a method and apparatus for forest fire detection and location based on mid-wave infrared images, which can achieve high-precision forest fire detection and location based on mid-wave infrared images and can meet the requirements of on-orbit real-time processing.

[0004] In a first aspect, embodiments of the present invention provide a forest fire detection and location method based on mid-wave infrared images, comprising:

[0005] The acquired mid-wave infrared image is processed by a full-image filtering algorithm to obtain the coarse location of the fire point target in the mid-wave infrared image;

[0006] Based on the coarse localization of the fire point target, the window image of the fire point target is determined;

[0007] A window filtering algorithm is applied to the window image of the fire point target to obtain the precise location of the fire point target in the mid-wave infrared image;

[0008] Geometric correction is performed on the geometric distortion of the mid-wave infrared image. Based on the precise positioning of the fire target and the geometrically corrected mid-wave infrared image, the spatial three-dimensional positioning result of the fire target is obtained.

[0009] In one possible implementation, the process of performing a full-image filtering algorithm on the acquired mid-wave infrared image includes:

[0010] Based on the row filter, the acquired mid-wave infrared image is convolved according to the row pixels to obtain the background gray level estimate of the corresponding pixel position;

[0011] Based on the pixel grayscale value of each pixel in the mid-wave infrared image and the estimated background grayscale value of the corresponding pixel position, determine whether each pixel is a fire point target;

[0012] The pixels identified as fire points are clustered to obtain a coarse localization of the fire points in the mid-wave infrared image.

[0013] In one possible implementation, determining whether each pixel is a fire point target includes:

[0014] Based on the response rate calibration results of the gray-scale difference and temperature difference of each pixel on the ground, the relative gray-scale threshold between the normal temperature background and the fire point target is determined.

[0015] For each pixel, the following steps are performed: determine whether the pixel grayscale value of the pixel is greater than the sum of the estimated background grayscale value of the corresponding pixel position and the relative grayscale threshold. If so, the pixel is determined to be a fire target.

[0016] In one possible implementation, determining the window image of the fire point target based on the coarse localization of the fire point target includes:

[0017] Extract the minimum envelope rectangle from the coarsely located fire point target, and expand the minimum envelope rectangle by a set number of pixels to obtain the window image of the fire point target.

[0018] In one possible implementation, the step of performing window filtering algorithm processing on the window image of the fire point target to obtain the precise localization of the fire point target in the mid-wave infrared image includes:

[0019] Using the border pixels in the window image of the fire target as the background, the background grayscale of the window image of the fire target is estimated to obtain the background grayscale estimate of the window image.

[0020] Based on the grayscale value of each pixel in the window image and the estimated grayscale value of the background of the window image, determine whether each pixel in the window image is a fire target;

[0021] The pixels identified as fire points are clustered to obtain the final fire point targets;

[0022] The centroid method is used to obtain the precise location of the fire point target in the mid-wave infrared image.

[0023] In one possible implementation, before performing background grayscale estimation on the window image of the fire target, the method further includes: removing outliers from the border pixels that serve as the background, and using the remaining border pixels as the background, in order to perform background grayscale estimation on the window image of the fire target.

[0024] Secondly, embodiments of the present invention also provide a forest fire detection and location device based on mid-wave infrared images, comprising:

[0025] The coarse localization unit is used to process the acquired mid-wave infrared image using a full-image filtering algorithm to obtain the coarse localization of the fire point target in the mid-wave infrared image.

[0026] A window image determination unit is used to determine a window image of the fire point target based on the coarse localization of the fire point target.

[0027] The fine positioning unit is used to perform window filtering algorithm processing on the window image of the fire point target to obtain the fine positioning of the fire point target in the mid-wave infrared image;

[0028] The result determination unit is used to perform geometric correction on the geometric distortion of the mid-wave infrared image, and obtain the spatial three-dimensional positioning result of the fire point target based on the precise positioning of the fire point target and the geometrically corrected mid-wave infrared image.

[0029] Thirdly, embodiments of the present invention also provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the method described in any embodiment of this specification.

[0030] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform the methods described in any embodiment of this specification.

[0031] This invention provides a method and apparatus for forest fire detection and location based on mid-wave infrared images. A full-image filtering algorithm can quickly and roughly locate fire targets in mid-wave infrared images. Then, by applying a window filtering algorithm to the coarsely located fire targets, a precise location can be obtained. Since the precise location process is performed on a window of the fire target image, there is no need to process pixel areas outside the fire target, greatly reducing the computational load. Furthermore, it can perform high-precision processing on small window areas, thus achieving precise fire target location. After geometric correction, the spatial three-dimensional location result of the fire target is obtained. This solution not only has low computational load, high efficiency, and high accuracy, but also meets the requirements for real-time on-orbit processing. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1This is a flowchart of a forest fire detection and location method based on mid-wave infrared images provided by an embodiment of the present invention;

[0034] Figure 2 This is a hardware architecture diagram of an electronic device provided in an embodiment of the present invention;

[0035] Figure 3 This is a structural diagram of a forest fire detection and positioning device based on mid-wave infrared images provided in an embodiment of the present invention. Detailed Implementation

[0036] 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 some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0037] As mentioned earlier, forest fire detection and localization based on multispectral images using a decision-level fusion approach requires target detection and information extraction for each spectral image separately, followed by spatial localization and spectral band matching of multispectral targets from time-division imaging to successfully obtain the final forest fire detection result. According to Wien's displacement law, the infrared radiation peak of forest fires is located in the mid-wave spectrum, while the infrared radiation peak of normal temperature environments is generally located in the long-wave spectrum. For forest fire target detection under normal temperature conditions, the forest fire detection sensor can use the mid-wave infrared image as a reference to predict the location of suspicious targets in other spectral infrared images, and then perform subsequent decision-level information fusion. Therefore, it is necessary to study a forest fire detection and localization method based on mid-wave infrared images to improve the accuracy of real-time on-orbit forest fire detection. Currently, however, there is no method for real-time on-orbit forest fire detection based on mid-wave infrared images.

[0038] Therefore, in order to meet the requirements of real-time on-orbit forest fire detection, a forest fire detection and location method with low computational load, high efficiency and high accuracy based on mid-wave infrared images is needed.

[0039] The specific implementation of the above concept is described below.

[0040] Please refer to Figure 1 This invention provides a method for forest fire detection and location based on mid-wave infrared images, the method comprising:

[0041] Step 100: Perform a full-image filtering algorithm on the acquired mid-wave infrared image to obtain the coarse location of the fire point target in the mid-wave infrared image;

[0042] Step 102: Based on the coarse localization of the fire point target, determine the window image of the fire point target;

[0043] Step 104: Perform window filtering algorithm processing on the window image of the fire point target to obtain the precise location of the fire point target in the mid-wave infrared image;

[0044] Step 106: Perform geometric correction on the geometric distortion of the mid-wave infrared image, and obtain the spatial three-dimensional positioning result of the fire point target based on the precise positioning of the fire point target and the geometrically corrected mid-wave infrared image.

[0045] In this embodiment of the invention, a full-image filtering algorithm can quickly and roughly locate fire targets in mid-wave infrared images. Then, by applying a window filtering algorithm to the coarsely located fire targets, precise location can be obtained. Since the precise location process is performed on a window image of the fire target, there is no need to process pixel areas outside the fire target, greatly reducing the computational load. Furthermore, it allows for high-precision processing of small window image areas, thus achieving precise fire target location. After geometric correction, the spatial three-dimensional location result of the fire target is obtained. This solution not only has low computational load and high efficiency but also high accuracy, meeting the requirements for real-time on-orbit processing.

[0046] The following description Figure 1 The execution method for each step is shown.

[0047] First, for step 100, the acquired mid-wave infrared image is processed by a full-image filtering algorithm to obtain a coarse location of the fire point target in the mid-wave infrared image.

[0048] Since the peak infrared radiation of forest fires is located in the mid-wave spectrum, detecting fire targets in mid-wave infrared images can more clearly and accurately determine whether a fire target exists. Furthermore, because it is uncertain whether a fire target exists in the mid-wave infrared image after acquisition, a preliminary check can be performed to determine the presence of a fire target and, if present, to roughly locate it.

[0049] In this embodiment of the invention, since each pixel may contain a fire target, a full-image filtering algorithm can be used to process the mid-wave infrared image in order to avoid missing any pixels. Specifically, this processing method may include the following steps A1-A3:

[0050] A1. Based on the row filter, the acquired mid-wave infrared image is convolved according to the row pixels to obtain the background gray level estimate of the corresponding pixel position;

[0051] A2. Based on the pixel grayscale value of each pixel in the mid-wave infrared image and the estimated background grayscale value of the corresponding pixel position, determine whether each pixel is a fire point target;

[0052] A3. Cluster the pixels identified as fire targets to obtain the coarse localization of fire targets in the mid-wave infrared image.

[0053] In step A1, the mid-wave infrared image is convolved row by row using a row filter to obtain the background grayscale estimate for each row pixel position.

[0054] In step A2, if a pixel is a fire target, it indicates that there is a large difference between the gray value of the pixel and the estimated gray value of the background at the corresponding pixel position. By comparing the difference between the pixel gray value and the estimated gray value of the background at the corresponding pixel position, it can be determined whether the pixel is a fire target.

[0055] To improve the accuracy of determining whether a pixel is a fire target, in one embodiment of the present invention, step A2 can determine whether each pixel is a fire target in the following way:

[0056] A21. Based on the response rate calibration results of the gray-scale difference and temperature difference of each pixel on the ground, determine the relative gray-scale threshold between the normal temperature background and the fire point target.

[0057] A22. For each pixel, perform the following: determine whether the pixel grayscale value of the pixel is greater than the sum of the estimated background grayscale value of the corresponding pixel position and the relative grayscale threshold. If so, determine that the pixel is a fire target.

[0058] Based on ground-based relative radiometric calibration experiments, the response rate calibration results for pixel grayscale differences and temperature differences can be obtained, establishing a one-to-one correspondence between the grayscale difference and temperature difference between two pixels. For example, for every 1-degree difference between two pixels, the grayscale difference increases by 100; for instance, if the temperature difference is 10 degrees, the grayscale difference is 1000, and if the temperature difference is 100 degrees, the temperature difference is 10000. The temperature difference between a normal temperature background and a fire target can be known empirically, for example, if it is 400 degrees. Therefore, based on the grayscale difference and temperature difference response rate calibration results, the grayscale difference is 40000. Thus, the relative grayscale threshold between a normal temperature background and a fire target can be determined to be 40000.

[0059] Since the estimated background grayscale values ​​differ at different pixel locations, in order to accurately determine whether a pixel is a fire target, it is necessary to compare the sum of the estimated background grayscale value and the relative grayscale threshold with the pixel grayscale value. If the pixel grayscale value is greater than the sum of the estimated background grayscale value and the relative grayscale threshold, then the pixel is determined to be a fire target; otherwise, the pixel is determined not to be a fire target.

[0060] By utilizing the relative grayscale threshold between the ambient background and the fire target, the estimated background grayscale value of the pixel location, and the pixel grayscale value, it is possible to more accurately determine whether a pixel is a fire target. This method uses the pixel as the smallest detection point; therefore, the detection result may contain multiple fire targets. These multiple fire targets can be continuous or discontinuous in the mid-wave infrared image. When multiple continuous pixels are all fire targets, it indicates that these continuous pixels belong to the same forest fire. Therefore, after determining whether a pixel is a fire target, step A3 needs to be performed to cluster multiple continuous pixels identified as fire targets into one category. This clustering method is that if a pixel identified as a fire target has an adjacent pixel that is also a fire target, then that pixel and its adjacent pixel are clustered together. After clustering, it is considered as one fire target, thus obtaining a coarse localization of fire targets in the mid-wave infrared image.

[0061] In this embodiment of the invention, the coarsely located fire point target can be one or multiple.

[0062] Then, the following explanations are given regarding step 102, "determine the window image of the fire point target based on the coarse localization of the fire point target" and step 104, "process the window image of the fire point target using a window filtering algorithm to obtain the fine localization of the fire point target in the mid-wave infrared image".

[0063] In this embodiment of the invention, fine localization is required for coarsely located fire targets to improve the accuracy of the detection and localization results. Fine localization involves refining the image of the coarsely located area of ​​the fire target, and this refining method is still achieved through background grayscale estimation.

[0064] In one implementation, the minimum envelope rectangle extracted from the fire point target can be used as the window image of the fire point target. However, in this method, the window image is always the fire point target, which will introduce errors in the background grayscale estimation results. Therefore, another implementation method can be used, which is to extract the minimum envelope rectangle from the coarsely located fire point target and expand this minimum envelope rectangle by a predetermined number of pixels to obtain the window image of the fire point target. This predetermined number is greater than 0, for example, one pixel, or 0.5 pixels.

[0065] In this embodiment of the invention, step 104 may specifically include the following steps B1-B4:

[0066] B1. Using the border pixels in the window image of the fire target as the background, perform background grayscale estimation on the window image of the fire target to obtain the background grayscale estimation value of the window image.

[0067] B2. Based on the grayscale value of each pixel in the window image and the estimated grayscale value of the background of the window image, determine whether each pixel in the window image is a fire point target;

[0068] B3. Cluster the pixels identified as fire points to obtain the final fire points;

[0069] B4. The centroid method is used to obtain the precise location of the fire point target in the mid-wave infrared image.

[0070] When performing background grayscale estimation in step B1, the border pixels in the window image of the fire target need to be used as the background. Therefore, when determining the window image, the minimum envelope rectangle is expanded by a set number of pixels, which can make the background grayscale estimation result more accurate.

[0071] In addition, in order to reduce the influence of blind pixels and other high-temperature targets that may exist at the edge of the window, outliers in the border pixels that serve as the background can be removed before the background grayscale estimation is performed on the window image of the fire target. The remaining border pixels are then used as the background for background grayscale estimation.

[0072] Outliers can be identified by using the standard deviation of the pixel grayscale values ​​of the border pixels used as the background.

[0073] It should be noted that the implementation methods of steps B2-B3 are the same as those of steps A2-A3, and will not be repeated here.

[0074] After obtaining the precise location of the fire point target, it is possible to determine which pixels in the mid-wave infrared image are the fire point target.

[0075] Finally, for step 106, the geometric distortion of the mid-wave infrared image is geometrically corrected, and the spatial three-dimensional positioning result of the fire point target is obtained based on the precise positioning of the fire point target and the geometrically corrected mid-wave infrared image.

[0076] In the process of Earth remote sensing imaging, the infrared images obtained are subject to geometric distortion due to the combined influence of various factors. By performing geometric correction on the geometric distortion of the mid-wave infrared image, the pixels of the corresponding fire point target in the geometrically corrected mid-wave infrared image become the final spatial three-dimensional positioning result.

[0077] In this embodiment of the invention, within the framework of infrared multispectral image decision-level fusion, mid-wave infrared images are rapidly and accurately used for forest fire detection and location. This enables the detection and location of small targets, and is not only computationally efficient and highly accurate, but also has on-orbit real-time processing capabilities.

[0078] like Figure 2 , Figure 3 As shown, this embodiment of the invention provides a forest fire detection and location device based on mid-wave infrared images. The device embodiment can be implemented through software, hardware, or a combination of both. From a hardware perspective, as... Figure 2 The diagram shown is a hardware architecture diagram of an electronic device for a forest fire detection and location device based on mid-wave infrared images, provided in an embodiment of the present invention. Except for... Figure 2 In addition to the processor, memory, network interface, and non-volatile memory shown, the electronic device in the embodiment may also include other hardware, such as a forwarding chip responsible for processing packets. Taking software implementation as an example, such as... Figure 3 As shown, a device in a logical sense is formed by the CPU of its host electronic device reading the corresponding computer program from non-volatile memory into memory and running it. This embodiment provides a forest fire detection and location device based on mid-wave infrared images, comprising:

[0079] The coarse localization unit 301 is used to perform a full-image filtering algorithm on the acquired mid-wave infrared image to obtain the coarse localization of the fire point target in the mid-wave infrared image.

[0080] The window image determination unit 302 is used to determine the window image of the fire point target based on the coarse localization of the fire point target;

[0081] The fine positioning unit 303 is used to perform window filtering algorithm processing on the window image of the fire point target to obtain the fine positioning of the fire point target in the mid-wave infrared image;

[0082] The result determination unit 304 is used to perform geometric correction on the geometric distortion of the mid-wave infrared image, and obtain the spatial three-dimensional positioning result of the fire point target based on the precise positioning of the fire point target and the geometrically corrected mid-wave infrared image.

[0083] In one embodiment of the present invention, the coarse positioning unit is specifically used for:

[0084] Based on the row filter, the acquired mid-wave infrared image is convolved according to the row pixels to obtain the background gray level estimate of the corresponding pixel position;

[0085] Based on the pixel grayscale value of each pixel in the mid-wave infrared image and the estimated background grayscale value of the corresponding pixel position, determine whether each pixel is a fire point target;

[0086] The pixels identified as fire points are clustered to obtain a coarse localization of the fire points in the mid-wave infrared image.

[0087] In one embodiment of the present invention, the coarse positioning unit, when determining whether each pixel is a fire point target, specifically includes:

[0088] Based on the response rate calibration results of the gray-scale difference and temperature difference of each pixel on the ground, the relative gray-scale threshold between the normal temperature background and the fire point target is determined.

[0089] For each pixel, the following steps are performed: determine whether the pixel grayscale value of the pixel is greater than the sum of the estimated background grayscale value of the corresponding pixel position and the relative grayscale threshold. If so, the pixel is determined to be a fire target.

[0090] In one embodiment of the present invention, the window image determination unit is specifically used to extract the minimum envelope rectangle of the coarsely located fire point target, and expand the minimum envelope rectangle by a set number of pixels to obtain the window image of the fire point target.

[0091] In one embodiment of the present invention, the precise positioning unit is specifically used for:

[0092] Using the border pixels in the window image of the fire target as the background, the background grayscale of the window image of the fire target is estimated to obtain the background grayscale estimate of the window image.

[0093] Based on the grayscale value of each pixel in the window image and the estimated grayscale value of the background of the window image, determine whether each pixel in the window image is a fire target;

[0094] The pixels identified as fire points are clustered to obtain the final fire point targets;

[0095] The centroid method is used to obtain the precise location of the fire point target in the mid-wave infrared image.

[0096] In one embodiment of the present invention, the fine positioning unit is further configured to remove outliers in the border pixels that serve as the background, and use the remaining border pixels as the background, so as to perform background grayscale estimation on the window image of the fire target.

[0097] It is understood that the structures illustrated in the embodiments of the present invention do not constitute a specific limitation on a forest fire detection and location device based on mid-wave infrared images. In other embodiments of the present invention, a forest fire detection and location device based on mid-wave infrared images may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0098] The information interaction and execution process between the modules in the above-mentioned device are based on the same concept as the method embodiment of the present invention, and the specific details can be found in the description of the method embodiment of the present invention, and will not be repeated here.

[0099] This invention also provides an electronic device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements a forest fire detection and location method based on mid-wave infrared images according to any embodiment of this invention.

[0100] This invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform a forest fire detection and location method based on mid-wave infrared images according to any embodiment of this invention.

[0101] Specifically, a system or apparatus equipped with a storage medium may be provided, on which software program code implementing the functions of any of the embodiments described above is stored, and the computer (or CPU or MPU) of the system or apparatus may read and execute the program code stored in the storage medium.

[0102] In this case, the program code read from the storage medium can itself implement the function of any of the above embodiments, and therefore the program code and the storage medium storing the program code constitute part of the present invention.

[0103] Examples of storage media used to provide program code include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards, and ROMs. Alternatively, program code can be downloaded from a server computer via a communication network.

[0104] Furthermore, it should be clear that not only can the program code read by the computer be executed, but also the operating system or other components operating on the computer can be instructed based on the program code to perform some or all of the actual operations, thereby realizing the function of any of the embodiments described above.

[0105] Furthermore, it is understood that the program code read from the storage medium is written to the memory set in the expansion board inserted into the computer or to the memory set in the expansion module connected to the computer. Then, based on the instructions of the program code, the CPU or other components installed on the expansion board or expansion module execute some and all of the actual operations, thereby realizing the function of any of the above embodiments.

[0106] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0107] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as ROM, RAM, magnetic disk, or optical disk.

[0108] 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 them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A forest fire detection and positioning method based on mid-wave infrared images, characterized in that, A method for detecting and positioning a forest fire target in real time in orbit, comprising: Convolving the collected mid-wave infrared image by a line filter to obtain a background gray scale estimation value of a corresponding pixel position; determining whether each pixel is a fire point target according to a pixel gray scale value of each pixel in the mid-wave infrared image and the background gray scale estimation value of the corresponding pixel position; clustering the pixels determined as fire point targets to obtain a coarse positioning of the fire point target in the mid-wave infrared image; the clustering mode is that if a pixel determined as a fire point target has an adjacent pixel which is a fire point target, the pixel and the adjacent pixel are clustered into one class; after clustering, the pixel and the adjacent pixel are taken as one fire point target; the determination of whether each pixel is a fire point target comprises: determining a relative gray scale threshold of a normal temperature background and a fire point target based on a response rate calibration result of a gray scale difference and a temperature difference of each pixel on the ground; for each pixel, it is determined whether the pixel gray scale value of the pixel is greater than the sum of the background gray scale estimation value of the corresponding pixel position of the pixel and the relative gray scale threshold, and if yes, the pixel is determined as a fire point target; Determining a window image of the fire point target based on the coarse positioning of the fire point target; Performing a window filtering algorithm on the window image of the fire point target to obtain a fine positioning of the fire point target in the mid-wave infrared image; Geometrically correcting the geometric distortion of the mid-wave infrared image, and obtaining a spatial three-dimensional positioning result of the fire point target according to the fine positioning of the fire point target and the geometrically corrected mid-wave infrared image.

2. The method of claim 1, wherein, The determination of the window image of the fire point target based on the coarse positioning of the fire point target comprises: Extracting a minimum envelope rectangle of the coarsely positioned fire point target, and extending the minimum envelope rectangle by a set number of pixels to obtain a window image of the fire point target.

3. The method of claim 1, wherein, The performing of the window filtering algorithm on the window image of the fire point target to obtain the fine positioning of the fire point target in the mid-wave infrared image comprises: Taking a border pixel in the window image of the fire point target as a background to estimate a background gray scale of the window image to obtain a background gray scale estimation value of the window image; Determining whether each pixel in the window image is a fire point target according to a pixel gray scale value of each pixel in the window image and the background gray scale estimation value of the window image; Clustering the pixels determined as fire point targets to obtain a final fire point target; Obtaining the fine positioning of the fire point target in the mid-wave infrared image by using a centroid method for the final fire point target.

4. The method of claim 3, wherein, Before the background gray scale estimation of the window image of the fire point target, the method further comprises: removing outliers in the border pixels as the background, and taking the remaining border pixels as the background to estimate the background gray scale of the window image of the fire point target.

5. A forest fire detection and positioning apparatus based on mid-wave infrared images, characterized by A device for performing the method for detecting and positioning a forest fire target based on a mid-wave infrared image according to any one of claims 1-4, comprising: a coarse positioning unit configured to perform a full-image filtering algorithm on the collected mid-wave infrared image to obtain a coarse positioning of a fire point target in the mid-wave infrared image; a window image determining unit configured to determine a window image of the fire point target based on the coarse positioning of the fire point target; a fine positioning unit configured to perform a window filtering algorithm on the window image of the fire point target to obtain fine positioning of the fire point target in the MWIR image; a result determining unit configured to perform geometric correction on the geometric distortion of the MWIR image, and obtain a spatial three-dimensional positioning result of the fire point target according to the fine positioning of the fire point target and the geometric corrected MWIR image. 6.An electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method of any one of claims 1-4. 7.A computer readable storage medium, having stored thereon a computer program, which when executed in a computer, causes the computer to perform the method of any one of claims 1-4.

Images