Infrared image splicing and labeling method and system based on image fusion and medium

Abstract

The invention discloses an infrared image splicing and labeling method and system based on image fusion and a medium, and relates to the technical field of image processing. The method comprises the steps that a target task is acquired, different target infrared collection devices associated with the target task are determined, and the target infrared collection devices comprise at least one static infrared collection device and at least one dynamic infrared collection device; acquiring a first infrared image acquired by each static infrared acquisition device and a second infrared image acquired by each dynamic infrared acquisition device; splicing and fusing each first infrared image and each second infrared image to obtain a fused image; performing target labeling on the fused image based on the target task to obtain a labeled target infrared image; and inputting the target infrared image into a preset model matched with the target task to obtain a task processing result corresponding to the target task. By implementing the technical scheme provided by the invention, the accuracy of monitoring the target object in the target task is improved.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to an infrared image stitching and annotation method, system, medium and device based on image fusion. Background Art

[0002] In the field of target tracking and detection, traditional visible light cameras and other sensors face significant performance limitations, especially in low light or bad weather conditions. In contrast, infrared imaging technology has the unique advantage of capturing the thermal information of targets in the absence of light or in low light conditions, and can track and detect targets at night or in bad weather conditions.

[0003] At present, infrared imaging technology mainly uses a single type of infrared device in target tracking and detection. These devices usually operate independently and collect infrared images before performing preliminary target recognition through artificial vision or basic computer vision algorithms. However, when a single infrared device is used to handle target tracking and detection tasks in complex environments, the image information may be incomplete due to the limitations of viewing angle and coverage, resulting in low accuracy in monitoring the target object in the target task. Summary of the invention

[0004] The present application provides an infrared image stitching and annotation method, system, medium and device based on image fusion, which improves the accuracy of monitoring target objects in target tasks.

[0005] In a first aspect, the present application provides an infrared image stitching and annotation method based on image fusion, the method comprising: Acquire a target task, and determine different target infrared acquisition devices associated with the target task, wherein the target infrared acquisition devices include at least one static infrared acquisition device and at least one dynamic infrared acquisition device; Acquire the first infrared image acquired by each of the static infrared acquisition devices and the second infrared image acquired by each of the dynamic infrared acquisition devices; splicing and fusing each of the first infrared images and each of the second infrared images to obtain a fused image; Performing target annotation on the fused image based on the target task to obtain an annotated target infrared image; The target infrared image is input into a preset model that matches the target task to obtain a task processing result corresponding to the target task.

[0006] By adopting the above technical scheme, the target task is determined, and the dynamic and static infrared acquisition devices are determined according to the target task. Then, the first infrared image and the second infrared image respectively acquired by the static and dynamic infrared devices are obtained, and the two types of images are effectively fused to generate a fused image integrating richer target information. With the support of the fused image, the target task is accurately labeled to obtain the labeled target infrared image, and then the target infrared image is input into the corresponding preset model, and finally a more accurate and reliable target task result is obtained, which effectively improves the performance of target detection and tracking based on infrared imaging, solves the problem of incomplete information in target monitoring of a single infrared device, and improves the accuracy of monitoring the target object in the target task.

[0007] Optionally, the determining of different target infrared acquisition devices associated with the target task includes: Analyze the target task, extract the activity area range, environmental feature points and target object features of the target object; Determine the static infrared acquisition equipment required to cover the activity area according to the environmental feature points; Based on the target object characteristics and the range of the activity area, determining at least one dynamic infrared acquisition device; Each of the static infrared acquisition devices and each of the dynamic infrared acquisition devices is used as a target infrared acquisition device.

[0008] By adopting the above technical solution, the target task is analyzed to extract information such as the target activity range and environmental characteristics, and then static and dynamic infrared devices are selectively determined based on this information to cover the target activity area. In this way, static devices can cover the entire activity range according to environmental characteristics, and dynamic devices can be key-tracked according to target characteristics, so that the determined infrared devices can take into account the monitoring needs of the target from multiple dimensions, so that the collected multiple infrared images contain comprehensive information from the environment to the target object, supporting subsequent image fusion and accurate monitoring.

[0009] Optionally, the acquiring of the first infrared image acquired by each of the static infrared acquisition devices and the second infrared image acquired by each of the dynamic infrared acquisition devices includes: Acquire first control data of the static infrared acquisition device and second control data of the dynamic infrared acquisition device; Calculating the collaborative coverage rate between the static infrared acquisition device and the dynamic infrared acquisition device based on the first control data and the second control data; Adjust the second control data based on the collaborative coverage ratio to obtain third control data; The static infrared acquisition device is controlled to acquire images according to the first control data to obtain a first infrared image, and the dynamic infrared acquisition device is controlled to acquire images according to the third control data to obtain a second infrared image.

[0010] By adopting the above technical solution, the control parameters of static and dynamic devices are obtained respectively, and the coordinated coverage of the two is calculated, and then the parameters of the dynamic device are adjusted based on the coverage to achieve better coordination. In this way, the static device can fully cover the basic range, and the dynamic device can flexibly supplement the key coverage. The two work together to obtain non-redundant and comprehensive infrared images. Ultimately, the multi-source infrared images obtained through parameter coordination can provide information-rich and non-repetitive support for subsequent image fusion, enhancing processing efficiency.

[0011] Optionally, the step of splicing and fusing the first infrared images and the second infrared images to obtain a fused image includes: Identifying common environmental feature points of each of the first infrared images and each of the second infrared images; Based on the common environmental feature points and the image acquisition time, the first infrared images and the second infrared images are grouped to obtain a plurality of groups of images; identifying a common image region between a first infrared image and a second infrared image in each set of images; Based on the common image area, image parameters and common environmental feature points, performing image alignment processing on each group of images; Each group of images after the alignment process is stitched and fused to obtain a fused image after the fusion process.

[0012] By adopting the above technical solution, the images are grouped using common features and time, and then the common areas of the images in each group are identified for targeted image alignment. In this way, infrared images from different sources can be accurately aligned by features and parameters in the region, and finally seamlessly fused. Images of static and dynamic devices can be fully aligned in time and space, and multi-source infrared information can be seamlessly connected to generate higher quality fused images.

[0013] Optionally, the first infrared images and the second infrared images are grouped based on the common environmental feature points and the image acquisition time to obtain multiple groups of images, including: Acquire the number of common environmental feature points between each of the first infrared images and each of the second infrared images and the time difference between image acquisition times; Determining whether the number of the common environmental feature points is greater than a preset number, and determining whether the time difference is less than a time difference threshold; The first infrared image and the second infrared image whose number of common environmental feature points is greater than a preset number and whose time difference is less than a time difference threshold are divided into one group to obtain multiple groups of images.

[0014] By adopting the above technical solution, the number of common feature points and the acquisition time difference between infrared images from different sources are obtained, and a threshold is set to determine whether the matching conditions are met. When there are enough common feature points and the time difference is not large, it can be determined that the two images have a corresponding relationship in scene and time, and they are automatically divided into a matching group. Through the automatic grouping judgment of the preset threshold, the extracted infrared image group with strong scene and time correlation can support subsequent accurate image alignment processing, and finally achieve seamless multi-source infrared image fusion that matches the target mission environment scene, improving the accuracy of task monitoring.

[0015] Optionally, the performing target labeling on the fused image based on the target task to obtain a labeled target infrared image includes: Determining thermal characteristics and shape characteristics of a target object according to the target task; determining a target area that meets the thermal and shape characteristics in the fused image; Calculating the bounding box coordinates and area of ​​the target area; According to the coordinates and area of ​​the bounding box, a labeling box is drawn in the fused image, and a target category label is added to obtain a labeled target infrared image.

[0016] By adopting the above technical solution, the thermal and morphological characteristics of the target are obtained based on target task analysis, and the target area is accurately locked in the information-rich infrared image after fusion. The key information such as the target area coordinates and category are annotated, which enables the subsequent model to quickly lock the target features and identify them, thereby improving the effect of monitoring various types of targets based on infrared imaging.

[0017] Optionally, inputting the target infrared image into a preset model matching the target task to obtain a task processing result corresponding to the target task includes: Matching a corresponding preset model according to the type of the target task, wherein the preset model includes one of an anomaly detection model, a trajectory generation model, and a target positioning model; Inputting the target infrared image into the preset model, generating a task processing report including target abnormal state or motion trajectory or positioning information; The task processing report is associated with the target infrared image and stored to obtain a task processing result.

[0018] By adopting the above technical solution, the preset model corresponding to the target task is matched, and the annotated image is input into the model, so that it can automatically output task reports such as detecting anomalies and generating trajectories. Finally, the image and its report are associated and stored to form a standard and complete result output.

[0019] In a second aspect of the present application, an infrared image stitching and annotation system based on image fusion is provided, the system comprising: A device determination module, used to obtain a target task and determine different target infrared acquisition devices associated with the target task, wherein the target infrared acquisition devices include at least one static infrared acquisition device and at least one dynamic infrared acquisition device; An image acquisition module, used to acquire the first infrared image acquired by each of the static infrared acquisition devices, and the second infrared image acquired by each of the dynamic infrared acquisition devices; An image fusion module, used for splicing and fusing each of the first infrared images and each of the second infrared images to obtain a fused image; An image annotation module, used for performing target annotation on the fused image based on the target task to obtain an annotated target infrared image; The task processing module is used to input the target infrared image into a preset model matching the target task to obtain a task processing result corresponding to the target task.

[0020] In a third aspect of the present application, a computer storage medium is provided, wherein the computer storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing the above method steps.

[0021] In a fourth aspect of the present application, an electronic device is provided, comprising: a processor and a memory; wherein the memory stores a computer program, and the computer program is suitable for being loaded by the processor and executing the above-mentioned method steps.

[0022] In summary, one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: The present application determines the target task, determines the dynamic and static infrared acquisition devices according to the target task, then obtains the first infrared image and the second infrared image respectively acquired by the static and dynamic infrared devices, and effectively fuses the two types of images to generate a fused image integrating richer target information. With the support of the fused image, accurate annotation is performed for the target task to obtain the annotated target infrared image, and then the target infrared image is input into the corresponding preset model, and finally a more accurate and reliable target task result is obtained, which effectively improves the performance of target detection and tracking based on infrared imaging, solves the problem of incomplete information in target monitoring with a single infrared device, and improves the accuracy of monitoring the target object in the target task. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flowchart of an infrared image stitching and annotation method based on image fusion provided in an embodiment of the present application; Figure 2 It is a module schematic diagram of an infrared image stitching and annotation system based on image fusion provided in an embodiment of the present application; Figure 3 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application.

[0024] Description of reference numerals: 300, electronic device; 301, processor; 302, communication bus; 303, user interface; 304, network interface; 305, memory. DETAILED DESCRIPTION

[0025] In order to enable technicians in this field to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0026] In the description of the embodiments of the present application, words such as "for example" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "for example" or "for example" is intended to present related concepts in a specific way.

[0027] In the description of the embodiments of the present application, the meaning of the term "multiple" refers to two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "include", "comprise", "have" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0028] The following will provide a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments.

[0029] Please refer to Figure 1 , a flowchart of an infrared image stitching and annotation method based on image fusion is proposed. The method can be implemented by a computer program, a single-chip microcomputer, or run on an infrared image stitching and annotation system based on image fusion. The computer program can be integrated into a computer device or run as an independent tool application. Specifically, the method includes steps 10 to 50, and the steps are as follows: Step 10: Acquire the target task, and determine different target infrared acquisition devices associated with the target task, where the target infrared acquisition devices include at least one static infrared acquisition device and at least one dynamic infrared acquisition device.

[0030] In the present application, the target task refers to a specific detection or tracking task that needs to be completed, which includes determining the environmental feature points for tracking / detection, the associated infrared acquisition device control data (such as device identification, acquisition control time), and the control route of the dynamic mobile device, etc. The target task can be a specific task type such as anomaly detection, mobile trajectory generation, target positioning, etc.

[0031] In the embodiment of the present application, the target infrared acquisition device refers to a collection of infrared image acquisition devices associated with a specific target task, including at least one static infrared acquisition device and at least one dynamic infrared acquisition device, which are determined to be the most suitable device combination to complete the specific target task.

[0032] In the embodiment of the present application, the static infrared acquisition device refers to an infrared image acquisition device fixed at a specific location. These devices are usually used to cover the activity area of ​​the target object and are arranged according to environmental feature points. The static device is responsible for acquiring the first infrared image.

[0033] In the embodiments of the present application, dynamic infrared acquisition devices refer to infrared image acquisition devices that can be moved or adjusted in position. These devices can be dynamically adjusted according to the characteristics of the target object and the range of the activity area to optimize the image acquisition effect. The dynamic devices are responsible for acquiring the second infrared image.

[0034] Specifically, in order to achieve the purpose of image stitching and annotation, it is necessary to first obtain the target task, which can be to detect abnormal targets in the environment, track the target movement trajectory, or locate specific targets. Then, according to the target task, different infrared acquisition devices related to it are determined. These devices are used to collect infrared images of the environmental area concerned by the target task. The purpose of determining the relevant equipment is to collect as much effective infrared information as possible so as to obtain the environmental area required to fully cover the target task. The target infrared acquisition equipment includes static infrared camera equipment and dynamic infrared camera equipment. The static infrared camera equipment is used to cover the key areas and feature points of the environment to obtain the overall environmental information. The dynamic infrared camera equipment is used to track moving targets and obtain detailed information of the target. The two are used in coordination to obtain panoramic information of the environment where the target is located. The determination of relevant equipment also needs to consider factors such as environmental feature points, target movement range, and target characteristics, and select the appropriate type and number of equipment for deployment to ensure that the obtained infrared image can meet the processing requirements of the target task. By obtaining the target task, clarifying the environmental area to be covered and the target features of concern, and selecting static and dynamic infrared devices for collaborative acquisition in a targeted manner, sufficient and effective image information can be obtained, laying the foundation for subsequent image fusion processing and target recognition tasks.

[0035] Based on the above embodiment, as an optional embodiment, the step of determining different target infrared acquisition devices associated with the target task may further include the following steps: Step 101: Analyze the target task and extract the activity area range, environmental feature points and target object features of the target object.

[0036] In order to determine the relevant infrared images required to fully cover the target task, it is necessary to parse the key information in the target task, mainly including the activity area range of the target object, the characteristic points in the environment, and the characteristics of the target object itself. The purpose of parsing the target task is to accurately determine the type and layout of the acquisition equipment. Specifically, firstly, the activity area range of the target object is parsed according to the target task to determine the spatial range that needs to be covered and monitored. Then, the characteristic points in the environment are extracted. These points can be used as references for image fusion between different devices, and can also be used to locate the target object. Finally, the characteristics of the target object, such as body shape, heat source distribution, etc., are parsed to determine the image clarity required for identifying and tracking the target. By parsing the target task to extract key information to guide the selection and layout of the equipment, the obtained infrared image can accurately match the processing requirements of the target task, and provide reliable and complete image support for subsequent image stitching, target recognition and behavior analysis algorithms, thereby improving the final task processing effect.

[0037] Step 102: Determine the static infrared acquisition equipment required to cover the activity area based on the environmental feature points.

[0038] Specifically, environmental feature points can be representative landmarks in the environment, such as bridges, forks, building entrances, etc. These points can determine the outline and range of the target activity area, and are also the coincidence positioning nodes when different devices are fused. According to the range of the activity area and the distribution of environmental feature points, a matrix layout with multiple rows and columns can be used to plan the position of static devices to ensure that each feature point is within the common coverage of at least two devices. At the same time, there is a certain overlap between adjacent devices. Static devices deployed in this way can completely cover the range of the activity area. Using feature point positioning to deploy static infrared devices can clarify the spatial relationship between devices and calculate accurate relative position parameters. At the same time, planning based on the layout of feature points can also optimize the number of devices and avoid waste of resources.

[0039] Step 103: Determine at least one dynamic infrared acquisition device based on the target object characteristics and the range of the activity area.

[0040] Specifically, after deploying static infrared equipment to cover the entire activity area, dynamic infrared equipment needs to be determined to track and collect detailed information of the target object. The determination of dynamic equipment needs to be based on the characteristics of the target object, such as body size, whether there is a heat source, and movement speed. These characteristics determine the type and number of dynamic collection equipment required. For example, for small targets without heat sources, higher resolution infrared cameras are required, and for targets that move quickly, multiple devices may be required for collaborative tracking. Specifically, the proportion of the target in the image can be estimated based on its size, and the focus tracking requirements can be calculated with reference to the movement speed to determine the model of a single dynamic device. Then, based on the scope of the activity area and the number of targets, determine the number of dynamic devices that need to be deployed to ensure that all targets in the entire area are covered.

[0041] Step 104: taking each static infrared acquisition device and each dynamic infrared acquisition device as a target infrared acquisition device.

[0042] Specifically, after determining the location of static devices and the number and model of dynamic devices, all selected static and dynamic infrared devices are confirmed and recorded as target infrared acquisition devices. A collection device information table can be established to record the device's identification number, model parameters, location coordinates and other information. Among them, dynamic devices also need to record motion range and speed parameters.

[0043] Step 20: Acquire the first infrared image acquired by each static infrared acquisition device and the second infrared image acquired by each dynamic infrared acquisition device.

[0044] In the embodiment of the present application, the first infrared image refers to an infrared image collected by a static infrared collection device, which is captured from a fixed position and usually covers a predefined monitoring area. The first infrared image provides stable background information and continuous monitoring data.

[0045] The second infrared image in the embodiment of the present application refers to an infrared image captured by a dynamic infrared acquisition device. These images are captured from a movable or adjustable device. The dynamic infrared acquisition device can move or adjust according to the characteristics and activity range of the target object to obtain more detailed or better angle images. The second infrared image provides a more flexible monitoring capability and can capture details or dynamic changes that static devices may miss.

[0046] Specifically, the purpose of acquiring infrared images is to collect various infrared information required for the target task, including the overall environment image acquired by the static infrared acquisition device, and the image of the specific target object acquired by the dynamic infrared acquisition device. These images will provide data support for subsequent information fusion and target recognition. Specifically, all static infrared devices can be started according to the calculated control parameters, and they can continue to collect images at the planned positions to obtain the first infrared image. Then, the dynamic infrared devices in the corresponding area are started for tracking and acquisition to obtain the second infrared image. The acquisition time of these two images can be different.

[0047] Based on the above embodiment, as an optional embodiment, the step of acquiring the first infrared image acquired by each static infrared acquisition device and the second infrared image acquired by each dynamic infrared acquisition device may further include the following steps: Step 201: Acquire first control data of a static infrared acquisition device and second control data of a dynamic infrared acquisition device.

[0048] Before acquiring an infrared image, it is necessary to determine the first control data of a static device and the second control data of a dynamic device respectively. The reason for determining them separately is that static devices and dynamic devices have different focuses and key parameters in image acquisition. For static infrared devices, more attention is paid to the global coverage of the scene and long-term stable acquisition, so the first control data will determine a lower image resolution to ensure the real-time data transmission, and at the same time determine a longer exposure time to improve the image signal-to-noise ratio. Dynamic infrared devices need to focus on the ability to identify and track moving targets, so the second control data will determine a higher image resolution to obtain more target details, and at the same time determine a shorter exposure time to reduce the blur of moving targets, and the gain parameter also needs to be appropriately increased.

[0049] Specifically, we can first analyze the environmental scope and target features that need to be paid attention to according to the target task, and clarify the image quality requirements for the two devices, such as scene information integrity, target detail clarity, etc. Then, combined with the equipment performance parameter table, calculate the optimal parameter combination corresponding to the two, which are used as the first and second control data respectively. For example, for a scene with a monitoring range of 200 meters × 200 meters, first determine the resolution of each static device to be 320 × 240 according to the coverage requirements. Then, based on the imaging quality data of the device at night, select an exposure time of 20 milliseconds and a gain of 10 times. For a moving target with a size of 2 meters × 1 meter, determine the resolution of the dynamic device to be 640 × 480, an exposure time of 10 milliseconds, and a gain of 20 times.

[0050] Step 202: Calculate the collaborative coverage between the static infrared acquisition device and the dynamic infrared acquisition device based on the first control data and the second control data.

[0051] Specifically, the purpose of calculating the collaborative coverage rate is to evaluate whether there is excessive overlap or gap in the image coverage of dynamic devices and static devices, so as to optimize the equipment deployment plan. The coverage boundary of each device can be inferred using a geometric model based on the location information, speed parameters, field of view, etc. of the two types of equipment, and then the ratio of the intersection of these coverage ranges to the combined part is calculated to obtain the collaborative coverage rate. If the collaborative coverage rate is too high, it means that there is a large amount of unnecessary repeated coverage, and you can consider reducing the number of dynamic devices. If the collaborative coverage rate is too low, there are missed monitoring blind spots, and dynamic devices need to be added to supplement the coverage.

[0052] Step 203: Adjust the second control data based on the cooperative coverage ratio to obtain third control data.

[0053] Specifically, it is necessary to judge whether the coverage range of the dynamic infrared acquisition equipment is reasonable based on the previously calculated collaborative coverage rate. If the collaborative coverage rate is too high, it means that the coverage areas of the dynamic equipment and the static equipment overlap too much, resulting in a waste of resources. At this time, you can consider reducing the acquisition frequency and resolution of the dynamic equipment, or narrowing its monitoring range to reduce redundant data collection. On the contrary, if the collaborative coverage rate is too low, it means that the dynamic equipment has failed to effectively supplement the coverage blind spots of the static equipment, which may cause the target to be lost. At this time, it is necessary to increase the acquisition frequency and resolution of the dynamic equipment, or expand its monitoring range to ensure effective coverage of key areas. These adjusted parameters ultimately constitute the third control data.

[0054] Assume that in a perimeter security system, the static infrared acquisition device has covered 80% of the area, and the initial second control data of the dynamic infrared acquisition device is set to a high-resolution scan of the entire area. After calculation, the collaborative coverage rate is as high as 90%, indicating that the coverage range of the dynamic device overlaps a lot with the static device. At this time, according to the results of the collaborative coverage analysis, the second control data of the dynamic device can be adjusted to only perform a high-resolution scan of the remaining 20% ​​uncovered area, and the scanning resolution of the covered area can be reduced to generate the third control data. This can not only ensure effective monitoring of the entire area, but also avoid waste of resources and improve system efficiency. On the contrary, if the initial collaborative coverage rate is only 50%, it means that the coverage range of the dynamic device is insufficient. At this time, the monitoring range of the dynamic device can be expanded, or its scanning frequency and resolution can be increased to make up for the coverage defects of the static device, and finally form a more effective third control data.

[0055] Step 204: Control the static infrared acquisition device to acquire an image according to the first control data to obtain a first infrared image, and control the dynamic infrared acquisition device to acquire an image according to the third control data to obtain a second infrared image.

[0056] Specifically, in the infrared image acquisition system, an instruction carrying the first control data is sent to the static infrared acquisition device, and an instruction containing the third control data is sent to the dynamic infrared acquisition device. The two types of devices independently set the working mode and acquisition configuration according to the corresponding control parameters. Finally, the static infrared acquisition device outputs the first infrared image to provide scene information, and the dynamic infrared acquisition device outputs the second infrared image to provide target information.

[0057] Step 30: stitching and fusing the first infrared images and the second infrared images to obtain a fused image.

[0058] Specifically, the purpose of image fusion is to achieve an organic combination of environmental panoramic information and target detail information, and obtain a fused infrared view suitable for target identification and positioning. The separate first image can provide a scene framework, and the second image can provide target features. Fusion of the two can give full play to the advantages of both. Specifically, the two images can be preprocessed first, including coordinate correction, denoising, etc., to improve the fusion effect; then, based on the coordinate information of the two, a common reference point is determined to correctly align the second image with the first image; then, an image fusion algorithm, such as weighted superposition, is used in the overlapping area to obtain a seamless transition effect. Finally, image enhancement is performed to correct local brightness differences and output the fused image.

[0059] Based on the above embodiment, as an optional embodiment, the step of splicing and fusing the first infrared images and the second infrared images to obtain a fused image may further include the following steps: Step 301: Identify common environmental feature points of each first infrared image and each second infrared image.

[0060] Specifically, the purpose of identifying common feature points is to find the corresponding matching relationship between the two images in order to perform correct image alignment. Since the two images come from different sources, there are differences in perspective and range, and it is necessary to establish a connection through common static elements in the environment. An image processing algorithm can be used to first extract features from the two images and detect significant features such as corners and edges in the images. Then, through the feature matching method, the descriptors of the feature points in the two images are compared to determine the feature point pairs with a high degree of matching, that is, to confirm them as common environmental feature points. These points usually correspond to significant static elements in the environment, such as building corners, road intersections, etc.

[0061] Step 302: grouping the first infrared images and the second infrared images based on the common environmental feature points and the image acquisition time to obtain a plurality of groups of images.

[0062] Specifically, the purpose of grouping is to distinguish two infrared images that are related in time and space, so as to facilitate the subsequent splicing of each group of images. Since there are differences between images obtained at different times and different viewing angles, they cannot be directly spliced, so they need to be grouped according to certain rules so that the images in each group can be fused. First, it is necessary to detect the number of common environmental feature points between each first infrared image and the second infrared image, and calculate the time difference between the acquisition time of the two images. Then set two thresholds, the preset number and the time difference threshold, to determine whether the number of common environmental feature points is greater than the preset number, and whether the time difference is less than the time difference threshold. Only when the number of common points of a pair of images is greater than the preset number, and the acquisition time difference is less than the time difference threshold, the pair of images is considered to be fusible and divided into the same group.

[0063] For example, the preset number is 10, and the time difference threshold is set to 100 milliseconds. For a pair of images with an acquisition time difference of 80 milliseconds and 12 common points, they are classified into the same group; while for a pair of images with an acquisition time difference of 120 milliseconds and only 8 common points, they are divided into two different groups. After traversal, all image pairs with strong correlation are divided into the same group, and finally multiple image groups that can be fused in time and space are obtained.

[0064] Step 303: Identify a common image region between the first infrared image and the second infrared image in each set of images.

[0065] Specifically, the purpose of identifying common areas is to clarify which specific parts of the two infrared images have content that can be fused. Since the acquisition angles and ranges of the two images are not exactly the same, there is often only overlap in local areas. Identifying these commonly covered areas can guide the subsequent image fusion process so that it is only processed in areas with fusible content. First, feature extraction can be performed on the two images in each group, and then the features of the two can be matched and compared to determine where the two images have more common features. This shows that the two images have common scene content at this location. Then the contour range of these position areas can be calculated and output as common image areas.

[0066] Step 304: performing image alignment processing on each group of images based on the common image area, image parameters and common environmental feature points.

[0067] Specifically, the purpose of image alignment is to make the two images coincide correctly in the coordinate system, laying the foundation for subsequent fusion processing. Due to the differences in viewing angle, distance, etc. between the two infrared images, direct splicing will result in positional offset problems, and image alignment can correct this offset. Some common feature points in the common area can be selected, and the spatial mapping transformation relationship between these points can be calculated based on the parameters of the two images, such as pixel size and field of view. Then, an image coordinate transformation model is established, and the coordinate system adjustment of the second infrared image is guided according to the model parameters to achieve alignment with the first infrared image.

[0068] Step 305: performing image stitching and fusion on each group of images after the alignment process to obtain a fused image after the fusion process.

[0069] Specifically, pixel-level image fusion techniques, such as weighted superposition algorithms, can be applied to the common areas of the two images. In the common area, different weights are given according to the input characteristics of the two images to smoothly transition the two images. For non-common areas, the content of a single image is directly taken, and then global visual effects are adjusted, such as illumination adjustment, and the fused image is output. Based on the alignment results, seamless and natural content combination is achieved, and image quality is also improved due to the integration of different sources.

[0070] Step 40: Annotate the fused image based on the target task to obtain an annotated target infrared image.

[0071] Specifically, according to the target category that needs to be identified in the target task analysis, the typical features of this type of target are extracted, such as heat source distribution, boundary contour, etc. On the fused image, image processing methods are used to detect potential target areas that meet the above characteristics, such as heat source segmentation, boundary detection, etc. For each potential target area, its bounding box coordinates are further calculated, and the area of ​​the area is calculated, and the target is screened according to the target size range in the task. For each screened target area, a rectangular box annotation representing the bounding box is superimposed on the image, and text annotations such as target category and serial number are added inside and around the box. Output the labeled target infrared images of all target areas to complete the target labeling process.

[0072] Based on the above embodiment, as an optional embodiment, the step of performing target labeling on the fused image based on the target task to obtain the labeled target infrared image may further include the following steps: Step 401: Determine the thermal characteristics and shape characteristics of the target object according to the target task.

[0073] Specifically, according to the target task analysis, the target category to be identified, such as people, vehicles, etc., is identified; typical infrared image samples of this type of target are consulted, and the heat source distribution range, shape, outline, etc. are statistically summarized; representative thermal features that distinguish this target from other types of targets are determined, such as the heat distribution of the head and torso; representative shape contour features are determined, such as the three-dimensional outline of a person, the rectangular outline of a vehicle, etc.; features are expressed with vocabulary, such as "the head and torso have stronger heat sources", "three-dimensional outline with a large top and a small bottom", etc.; a text description file of thermal features and shape features is generated to provide a basis for subsequent target positioning.

[0074] Step 402: Determine a target area that meets the thermal and shape characteristics in the fused image.

[0075] Specifically, the fused image is processed by using image processing algorithms such as heat source segmentation and edge detection. Among all detected potential target areas, the image area that matches the thermal distribution characteristics and contour characteristics of the predetermined target category is selected as the formal target area, completing the target positioning in the complex fused image. This method, which combines pre-feature analysis with image processing technology, can make the target area determine and focus on the specific category set in the matching target task, exclude other interference types, and ensure the accuracy of subsequent target labeling.

[0076] Step 403: Calculate the bounding box coordinates and area of ​​the target area.

[0077] Specifically, the image processing algorithm is used to extract the edges of each candidate area, fit the boundary contour rectangle, and then calculate the coordinate values ​​of the intersection of the rectangle diagonals to obtain the bounding box coordinate data. At the same time, the area of ​​the rectangle is calculated to obtain the area of ​​the target area. After obtaining accurate numerical data in these two aspects, the area of ​​the correct target can be filtered out according to constraints such as the target size range, and the coordinate and area parameters can be used to draw the annotation box, thereby improving the accuracy and effect of target annotation.

[0078] Step 404: Draw a labeling box in the fused image according to the coordinates and area of ​​the bounding box, and add a target category label to obtain a labeled target infrared image.

[0079] Specifically, a rectangular annotation box is drawn at the corresponding image position according to the bounding box coordinate parameters to frame the target area; then a text label is added according to the target category, and the mislabeled boxes are filtered out in combination with the area parameter; finally, an infrared image of the target with clear annotations is output. This method of combining target geometric information with annotation drawing can distinguish different targets, while providing intuitive information such as target category and size, improving the expression of target information and providing accurate and complete target input data for subsequent algorithm analysis and processing.

[0080] Step 50: Input the target infrared image into a preset model that matches the target task to obtain a task processing result corresponding to the target task.

[0081] The preset model in the present application embodiment refers to a processing algorithm or analysis model that is preset according to the type of target task. The preset model includes but is not limited to the following types: anomaly detection model, trajectory generation model, and target positioning model.

[0082] The task processing result in the embodiment of the present application refers to the output information obtained after the marked target infrared image is input into the preset model for processing. Depending on different preset models and target tasks, the task processing result may include but is not limited to the following: target abnormal state, motion trajectory and positioning information.

[0083] Specifically, after obtaining the annotated target infrared image, it is necessary to submit it as input to the preset model that matches the target task to obtain the final task processing result. The purpose of this step is to use the features extracted from the target image to realize special task analysis and processing such as detection, recognition, and positioning of the target. You can select a suitable preset model according to different types of target tasks, such as a detection model, trajectory model, etc., and submit the annotated image as input data to the model. The model will automatically analyze and process the target in the image and output results, such as anomaly detection results or target trajectory. In this way, through professional target analysis and processing by a model that matches the target task, accurate and reliable task result output can be obtained to meet the final processing requirements.

[0084] Based on the above embodiment, as an optional embodiment, the step of inputting the target infrared image into a preset model matching the target task to obtain the task processing result corresponding to the target task may also include the following steps: Step 501: matching a corresponding preset model according to the type of the target task, where the preset model includes one of an anomaly detection model, a trajectory generation model, and a target positioning model.

[0085] Specifically, in order to obtain corresponding specialized processing results for different types of target tasks, it is necessary to match different preset models for image analysis according to the task type. Because different tasks such as anomaly detection, trajectory generation, and target positioning have different requirements for image analysis, different models are needed to achieve the best results. You can select matching preset models, including anomaly detection models, trajectory generation models, and target positioning models, according to whether the target task belongs to the three types of anomaly detection, trajectory generation, or positioning. These models can be obtained through different sample training and are suitable for the corresponding task types. At run time, after receiving the target task, match its corresponding model category, input the annotated infrared image into the selected model, and obtain the result output that meets the task requirements through the preset model processing. This method of automatically matching models according to task types can make the entire image analysis process efficiently and accurately adapt to different target tasks and realize customized processing of various tasks.

[0086] Step 502: Input the target infrared image into a preset model to generate a task processing report including the target abnormal state or movement trajectory or positioning information.

[0087] Specifically, after selecting the matching preset model through the target task type, the annotated infrared target image needs to be input into the selected model for processing, and finally a task processing report is generated. The model's professional image analysis capabilities can be used to extract the key state information of the target and form the result output of tasks such as detection and positioning. For example, if the anomaly detection model is selected, the model will analyze the abnormal signs of the target in the image sequence and generate a detection report to determine the abnormal state of the target. If the trajectory generation model is selected, the model will analyze the movement law of the target and generate a report reflecting the target's movement trajectory. If the positioning model is selected, the model will analyze the precise coordinates of the target area and generate a report representing the target positioning information. After inputting the image to be processed, the model will automatically output the corresponding task result report obtained by professional analysis.

[0088] Step 503: The task processing report is associated with the target infrared image and stored to obtain the task processing result.

[0089] Specifically, an index association can be established between image data and report data so that each target image points to the corresponding processing report. When saving, the two are stored in the database according to the association relationship, such as binding by file name, ID, etc. When extracting later, the corresponding report can be quickly found according to the target image to verify the results. This method of storing images and reports in association can make each step in the task processing flow clearly recorded and reconstruct the connection between the image source and the processing result.

[0090] See also Figure 2, is a module schematic diagram of an infrared image stitching and annotation system based on image fusion provided in an embodiment of the present application. The infrared image stitching and annotation system based on image fusion may include: a device determination module, an image acquisition module, an image fusion module, an image annotation module and a task processing module, wherein: A device determination module, used to obtain a target task and determine different target infrared acquisition devices associated with the target task, wherein the target infrared acquisition devices include at least one static infrared acquisition device and at least one dynamic infrared acquisition device; An image acquisition module, used to acquire the first infrared image acquired by each of the static infrared acquisition devices, and the second infrared image acquired by each of the dynamic infrared acquisition devices; An image fusion module, used for splicing and fusing each of the first infrared images and each of the second infrared images to obtain a fused image; An image annotation module, used for performing target annotation on the fused image based on the target task to obtain an annotated target infrared image; The task processing module is used to input the target infrared image into a preset model matching the target task to obtain a task processing result corresponding to the target task.

[0091] Optionally, the device determination module is also used to parse the target task, extract the activity area range, environmental feature points and target object characteristics of the target object; determine the static infrared acquisition equipment required to cover the activity area range according to the environmental feature points; determine at least one dynamic infrared acquisition equipment based on the target object characteristics and the activity area range; and use each of the static infrared acquisition equipment and each of the dynamic infrared acquisition equipment as a target infrared acquisition equipment.

[0092] Optionally, the image acquisition module is also used to obtain first control data of the static infrared acquisition device and second control data of the dynamic infrared acquisition device; calculate the collaborative coverage rate between the static infrared acquisition device and the dynamic infrared acquisition device based on the first control data and the second control data; adjust the second control data based on the collaborative coverage rate to obtain third control data; control the static infrared acquisition device to perform image acquisition according to the first control data to obtain a first infrared image, and control the dynamic infrared acquisition device to perform image acquisition according to the third control data to obtain a second infrared image.

[0093] Optionally, the image fusion module is further used to identify common environmental feature points of each of the first infrared images and each of the second infrared images; based on the common environmental feature points and image acquisition time, group each of the first infrared images and each of the second infrared images to obtain multiple groups of images; identify the common image area between the first infrared image and the second infrared image in each group of images; and perform image alignment processing on each group of images based on the common image area, image parameters and common environmental feature points; Each group of images after the alignment process is stitched and fused to obtain a fused image after the fusion process.

[0094] Optionally, the image fusion module is also used to obtain the number of common environmental feature points between each of the first infrared images and each of the second infrared images and the time difference between image acquisition times; determine whether the number of common environmental feature points is greater than a preset number, and determine whether the time difference is less than a time difference threshold; and divide the first infrared image and the second infrared image whose number of common environmental feature points is greater than a preset number and whose time difference is less than the time difference threshold into a group to obtain multiple groups of images.

[0095] Optionally, the image annotation module is also used to determine the thermal characteristics and shape characteristics of the target object according to the target task; determine the target area that meets the thermal characteristics and shape characteristics in the fused image; calculate the bounding box coordinates and area of ​​the target area; draw a annotation box in the fused image according to the bounding box coordinates and area, and add a target category label to obtain a labeled target infrared image.

[0096] Optionally, the task processing module is also used to match a corresponding preset model according to the type of the target task, the preset model including one of an anomaly detection model, a trajectory generation model and a target positioning model; input the target infrared image into the preset model to generate a task processing report including the target abnormal state or motion trajectory or positioning information; and store the task processing report in association with the target infrared image to obtain a task processing result.

[0097] It should be noted that: when the system provided in the above embodiment realizes its functions, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0098] An embodiment of the present application also provides a computer storage medium, which can store multiple instructions. The instructions are suitable for being loaded by a processor and executing an infrared image stitching and annotation method based on image fusion of the above embodiment. The specific execution process can be found in the specific description of the above embodiment, which will not be repeated here.

[0099] Please refer to Figure 3 The application also discloses an electronic device. Figure 3 The electronic device 300 may include: at least one processor 301 , at least one network interface 304 , a user interface 303 , a memory 305 , and at least one communication bus 302 .

[0100] The communication bus 302 is used to realize the connection and communication between these components.

[0101] The user interface 303 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 303 may also include a standard wired interface and a wireless interface.

[0102] The network interface 304 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0103] Among them, the processor 301 may include one or more processing cores. The processor 301 uses various interfaces and lines to connect various parts in the entire server, and executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 305, and calling data stored in the memory 305. Optionally, the processor 301 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 301 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 301, and it can be implemented separately through a chip.

[0104] Among them, the memory 305 may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read-Only Memory). Optionally, the memory 305 includes a non-transitory computer-readable storage medium. The memory 305 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 305 may optionally also be at least one storage device located away from the aforementioned processor 301. Refer to Figure 3 The memory 305 as a computer storage medium may include an operating system, a network communication module, a user interface module, and an application program of an infrared image stitching and annotation method based on image fusion.

[0105] exist Figure 3 In the electronic device 300 shown, the user interface 303 is mainly used to provide an input interface for the user and obtain the data input by the user; and the processor 301 can be used to call the application program stored in the memory 305 for an infrared image stitching and annotation method based on image fusion. When executed by one or more processors 301, the electronic device 300 executes one or more of the methods described in the above embodiments. It should be noted that for the aforementioned method embodiments, for the sake of simple description, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for the present application.

[0106] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0107] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0108] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0109] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0110] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: various media that can store program codes, such as USB flash drives, mobile hard drives, magnetic disks or optical disks.

[0111] The above is only an exemplary embodiment of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and practice, those skilled in the art will easily think of other embodiments of the present disclosure.

[0112] This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art not described in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. An infrared image stitching and annotation method based on image fusion, It is characterized in that The method comprises: Acquire a target task, and determine different target infrared acquisition devices associated with the target task, wherein the target infrared acquisition devices include at least one static infrared acquisition device and at least one dynamic infrared acquisition device; Acquire the first infrared image acquired by each of the static infrared acquisition devices and the second infrared image acquired by each of the dynamic infrared acquisition devices; splicing and fusing each of the first infrared images and each of the second infrared images to obtain a fused image; Performing target annotation on the fused image based on the target task to obtain an annotated target infrared image; The target infrared image is input into a preset model that matches the target task to obtain a task processing result corresponding to the target task.

2. According to the infrared image stitching and annotation method based on image fusion according to claim 1, It is characterized in that The determining of different target infrared acquisition devices associated with the target task includes: Analyze the target task, extract the activity area range, environmental feature points and target object features of the target object; Determine the static infrared acquisition equipment required to cover the activity area according to the environmental feature points; Based on the target object characteristics and the range of the activity area, determining at least one dynamic infrared acquisition device; Each of the static infrared acquisition devices and each of the dynamic infrared acquisition devices is used as a target infrared acquisition device.

3. The infrared image stitching and annotation method based on image fusion according to claim 1, It is characterized in that The acquiring of the first infrared image acquired by each of the static infrared acquisition devices and the second infrared image acquired by each of the dynamic infrared acquisition devices comprises: Acquire first control data of the static infrared acquisition device and second control data of the dynamic infrared acquisition device; Calculating the collaborative coverage rate between the static infrared acquisition device and the dynamic infrared acquisition device based on the first control data and the second control data; Adjust the second control data based on the collaborative coverage ratio to obtain third control data; The static infrared acquisition device is controlled to acquire images according to the first control data to obtain a first infrared image, and the dynamic infrared acquisition device is controlled to acquire images according to the third control data to obtain a second infrared image.

4. The infrared image stitching and annotation method based on image fusion according to claim 1, It is characterized in that The step of stitching and fusing the first infrared images and the second infrared images to obtain a fused image includes: Identifying common environmental feature points of each of the first infrared images and each of the second infrared images; Based on the common environmental feature points and the image acquisition time, the first infrared images and the second infrared images are grouped to obtain a plurality of groups of images; identifying a common image region between a first infrared image and a second infrared image in each set of images; Based on the common image area, image parameters and common environmental feature points, performing image alignment processing on each group of images; Each group of images after the alignment process is stitched and fused to obtain a fused image after the fusion process.

5. The infrared image stitching and annotation method based on image fusion according to claim 4, It is characterized in that The first infrared images and the second infrared images are grouped based on the common environmental feature points and the image acquisition time to obtain multiple groups of images, including: Acquire the number of common environmental feature points between each of the first infrared images and each of the second infrared images and the time difference between image acquisition times; Determining whether the number of the common environmental feature points is greater than a preset number, and determining whether the time difference is less than a time difference threshold; The first infrared image and the second infrared image whose number of common environmental feature points is greater than a preset number and whose time difference is less than a time difference threshold are divided into one group to obtain multiple groups of images.

6. The infrared image stitching and annotation method based on image fusion according to claim 1, It is characterized in that The target labeling of the fused image based on the target task to obtain a labeled target infrared image includes: Determining thermal characteristics and shape characteristics of a target object according to the target task; determining a target area that meets the thermal and shape characteristics in the fused image; Calculating the bounding box coordinates and area of ​​the target area; According to the coordinates and area of ​​the bounding box, a labeling box is drawn in the fused image, and a target category label is added to obtain a labeled target infrared image.

7. The infrared image stitching and annotation method based on image fusion according to claim 1, It is characterized in that The step of inputting the target infrared image into a preset model matching the target task to obtain a task processing result corresponding to the target task includes: Matching a corresponding preset model according to the type of the target task, wherein the preset model includes one of an anomaly detection model, a trajectory generation model, and a target positioning model; Inputting the target infrared image into the preset model, generating a task processing report including target abnormal state or motion trajectory or positioning information; The task processing report is associated with the target infrared image and stored to obtain a task processing result.

8. An infrared image stitching and annotation system based on image fusion, It is characterized in that The system comprises: A device determination module, used to obtain a target task and determine different target infrared acquisition devices associated with the target task, wherein the target infrared acquisition devices include at least one static infrared acquisition device and at least one dynamic infrared acquisition device; An image acquisition module, used to acquire the first infrared image acquired by each of the static infrared acquisition devices, and the second infrared image acquired by each of the dynamic infrared acquisition devices; An image fusion module, used for splicing and fusing each of the first infrared images and each of the second infrared images to obtain a fused image; An image annotation module, used for performing target annotation on the fused image based on the target task to obtain an annotated target infrared image; The task processing module is used to input the target infrared image into a preset model matching the target task to obtain a task processing result corresponding to the target task.

9. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the method according to any one of claims 1 to 7.

10. An electronic device, It is characterized in that It includes a processor, a memory, a user interface and a network interface, the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes the method as described in any one of claims 1-7.

Images