Intelligent damage assessment methods, devices, equipment, media, and procedures.

Abstract

This invention provides an intelligent damage assessment method, apparatus, device, medium, and program product, relating to the field of artificial intelligence technology. The method includes: acquiring a target image of the target to be assessed and acquiring sensor data of the target; extracting features from the target image to obtain image features, and extracting features from the sensor data to obtain sensor features; inputting the image features and sensor features into a trained damage assessment model, and obtaining the damage assessment result output by the damage assessment model. The damage assessment model is trained based on multiple sets of training data, each set of training data including a sample feature set and corresponding damage assessment result labels. The sample feature set includes sample image features and sample sensor features. This invention eliminates the need for manual on-site inspection of the target to obtain the damage assessment result, achieving automated damage assessment of the target.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to intelligent damage assessment methods, devices, equipment, media, and program products. Background Technology

[0002] Damage assessment is used to evaluate the effectiveness of weapon strikes in a battlefield environment. It is a comprehensive analysis and evaluation process that considers factors such as operational objectives, battlefield environment, and destructive forces to assess the effectiveness of firepower strikes. Timely and accurate target damage assessment in a battlefield environment not only provides a basis for commanders to make correct decisions but also maximizes the optimization of firepower and improves the efficiency of operational resource utilization.

[0003] Traditional damage assessment mainly relies on manual on-site investigation. This method is not only greatly affected by human factors, but also inefficient and dangerous. Summary of the Invention

[0004] This invention provides intelligent damage assessment methods, devices, equipment, media, and program products to address the shortcomings of existing technologies that require manual on-site inspection for damage assessment, thereby achieving automated damage assessment.

[0005] This invention provides an intelligent damage assessment method, comprising:

[0006] Acquire a target image of the target to be evaluated, and acquire sensor data of the target to be evaluated;

[0007] Feature extraction is performed on the target image to obtain image features, and feature extraction is performed on the sensor data to obtain sensor features;

[0008] The image features and sensor features are input into the trained damage assessment model to obtain the damage assessment result output by the damage assessment model. The damage assessment model is trained based on multiple sets of training data. Each set of training data includes a sample feature group and a damage assessment result label corresponding to the sample feature group. The sample feature group includes sample image features and sample sensor features.

[0009] According to the intelligent damage assessment method provided by the present invention, acquiring the target image of the target to be assessed includes:

[0010] Multiple images of the target to be evaluated are acquired. Each image corresponds to a different time point in time at which the target is attacked.

[0011] Based on the display range of the multiple images, the multiple images are corrected to obtain multiple intermediate images, each of which has the same display range;

[0012] Keyframes are determined from the plurality of intermediate images and used as the target image.

[0013] According to the intelligent damage assessment method provided by the present invention, the step of acquiring multiple images of the target to be assessed includes:

[0014] Acquire images captured by the drone;

[0015] Multiple images of the target to be evaluated are obtained from the images captured by the UAV based on the geographical location of the target and the geographical location trajectory of the UAV.

[0016] According to the intelligent damage assessment method provided by the present invention, the step of acquiring sensor data of the target to be assessed includes:

[0017] Data collected by multiple sensors of the target to be evaluated at different target times are obtained, where each target time corresponds to the process in which the target to be evaluated is attacked.

[0018] The collected data are integrated based on the setting position of each sensor to obtain each sensor data, and each sensor data contains collected data from different types of sensors with the same setting position.

[0019] According to the intelligent damage assessment method provided by the present invention, the step of extracting features from the target image to obtain image features includes:

[0020] The image features are extracted using a preset feature extraction algorithm to obtain the first feature;

[0021] Feature selection is performed on the first feature to reduce its dimensionality, resulting in a second feature;

[0022] The second feature is transformed to obtain the image feature.

[0023] According to the intelligent damage assessment method provided by the present invention, after obtaining the damage assessment result output by the damage assessment model, the method includes:

[0024] Obtain expert evaluation results for the damage assessment results, wherein the expert evaluation results reflect the damage assessment experts' evaluation of the damage assessment results;

[0025] Based on the expert evaluation results, the damage assessment model is updated.

[0026] The present invention also provides an intelligent damage assessment device, comprising:

[0027] The data acquisition module is used to acquire the target image of the target to be evaluated and to acquire the sensor data of the target to be evaluated;

[0028] The feature extraction module is used to extract features from the target image to obtain image features and to extract features from the sensor data to obtain sensor features.

[0029] The model processing module is used to input the image features and the sensor features into the trained damage assessment model and obtain the damage assessment result output by the damage assessment model. The damage assessment model is trained based on multiple sets of training data. Each set of training data includes a sample feature group and a damage assessment result label corresponding to the sample feature group. The sample feature group includes sample image features and sample sensor features.

[0030] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the intelligent damage assessment method as described above.

[0031] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the intelligent damage assessment method as described above.

[0032] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the intelligent damage assessment method as described above.

[0033] The intelligent damage assessment method, apparatus, equipment, medium, and program products provided by this invention acquire image and sensor data of the target to be assessed, extract features from the image and sensor data of the target to be assessed, and input the extracted image features and sensor features into a trained damage assessment model to obtain the damage assessment result output by the model. This invention does not require manual entry into the target to be assessed for on-site investigation to obtain the damage assessment result of the target to be assessed; it can realize automated damage assessment of the target to be assessed based solely on the image and sensor data of the target to be assessed. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in this 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 this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0035] Figure 1This is one of the flowcharts of an intelligent damage assessment method provided by the present invention;

[0036] Figure 2 This is a second flowchart illustrating an intelligent damage assessment method provided by the present invention.

[0037] Figure 3 This is a schematic diagram of the intelligent damage assessment device provided by the present invention;

[0038] Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0040] In existing technologies, damage assessment requires manual entry into the site for investigation and evaluation. The assessment results obtained by this method are not only affected by the experience and subjective factors of the on-site investigators, but also very inefficient. In addition, there are certain dangers associated with manual entry into the weapon strike zone.

[0041] To address the shortcomings of existing technologies that require manual on-site inspection for damage assessment, this invention provides intelligent damage assessment methods, devices, equipment, media, and program products to achieve automated damage assessment.

[0042] The following is combined Figure 1 Describe the intelligent damage assessment method provided by this invention, such as... Figure 1 As shown, the method includes the following steps:

[0043] S110. Acquire the target image of the target to be evaluated and acquire the sensor data of the target to be evaluated;

[0044] S120. Extract features from the target image to obtain image features, and extract features from the sensor data to obtain sensor features;

[0045] S130. Input the image features and sensor features into the trained damage assessment model to obtain the damage assessment result output by the damage assessment model. The damage assessment model is trained based on multiple sets of training data. Each set of training data includes a sample feature group and the corresponding damage assessment result label. The sample feature group includes sample image features and sample sensor features.

[0046] The target to be evaluated is one that has been hit by a weapon, and the damage assessment result reflects the weapon's impact effect. The method provided by this invention acquires images and sensor data of the target to be evaluated, extracts features from these data, and inputs the extracted image and sensor features into a trained damage assessment model to obtain the damage assessment result output by the model. This invention eliminates the need for manual on-site investigation of the target to obtain damage assessment results; it achieves automated damage assessment of the target solely based on its image and sensor data.

[0047] Obtain target images of the target to be evaluated, including:

[0048] Multiple images of the target to be evaluated are acquired, each image corresponding to a different target time, and each target time corresponding to the process of the target being attacked.

[0049] Based on the display range of multiple images, the multiple images are corrected to obtain multiple intermediate images, and the display range of each intermediate image is consistent.

[0050] Keyframes are identified from multiple intermediate images and used as the target image.

[0051] In one possible implementation, acquiring the target image of the target to be evaluated can be an image obtained by a camera imaging the target after it has been attacked. In order to accurately assess the damage results of the target to be evaluated, the method provided by the present invention acquires images of the target to be evaluated at different times during the attack process. These images can reflect the changes of the target to be evaluated during the attack process and provide more information for damage assessment.

[0052] It is understandable that the more images of the target to be evaluated, the more information they carry. However, more images also mean more processing steps and higher computational resource consumption. In the method provided by this invention, after acquiring multiple images of the target to be evaluated, not all images are used as the target image for subsequent processing. Instead, keyframes are determined from the multiple images, and the target image is obtained based on the keyframes.

[0053] A keyframe is a crucial image representing a change in the state of a target within a series of images over a period of time. Determining keyframes from multiple images can be achieved using existing keyframe extraction techniques, which will not be elaborated upon here. Extracting keyframes from multiple images of the target to be evaluated reduces image processing workload and resource consumption.

[0054] Since the target to be evaluated is a target subjected to fire, obtaining multiple images of it may have certain limitations, resulting in multiple images corresponding to different imaging ranges, i.e., the display range of multiple images may be different. Since keyframe detection is based on the features of image display content, in order to prevent the display range of the image from affecting the keyframe detection results, the method provided in this invention first performs correction processing on multiple images of the target to be evaluated before determining the keyframe, so that their corresponding display ranges are consistent. This can improve the detection accuracy of keyframes and ensure the accuracy of the damage assessment results obtained based on keyframes.

[0055] It is understandable that the target to be assessed may be large, such as a building or an area. To accurately assess the damage to the target, acquiring images from more angles clearly reflects a more comprehensive picture and improves the accuracy of the damage assessment. One possible implementation involves using multiple fixed cameras positioned on the target to acquire images from multiple angles. However, since the target is an object subjected to attack, setting up fixed cameras requires consideration of the attack range, attack intensity, and the influence of the surrounding environment. Adopting different fixed camera setups for different targets requires significant time and financial resources. To reduce the cost of image acquisition for the target, the method provided in this invention uses a drone to acquire images of the target. Specifically, it acquires multiple images of the target, including:

[0056] Acquire images captured by the drone;

[0057] Multiple images are acquired from the images transmitted back by the drone, based on the geographic location of the target to be evaluated and the geographic trajectory of the drone.

[0058] Multiple images are acquired from the images transmitted back by the drone, based on the geographic location of the target to be evaluated and the geographic trajectory of the drone, including:

[0059] The geographic area to be filmed is determined based on the geographic location of the target to be evaluated and the camera parameters of the drone.

[0060] The images captured when the drone's geographical location is within the captured geographical area are used as multiple images of the target to be evaluated.

[0061] Based on the drone's camera parameters, it can be determined how far away the drone can capture a clear image of the target being evaluated. Combining the drone's camera parameters with the geographical location of the target, the shooting geographical area can be determined. The shooting geographical area is the geographical area where the drone's camera is suitable for shooting the target. In other words, when the drone is in the shooting geographical area, it can capture a clear image of the target.

[0062] During the acquisition and transmission of UAV images, noise and data loss may occur. After obtaining multiple images, they can be preprocessed before keyframe extraction and subsequent processing. Image preprocessing can be implemented through integrated modularity, that is, by calling image preprocessing algorithms through open interfaces. This supports the import of external algorithms, enhancing the adaptability of the intelligent damage assessment method provided by this invention.

[0063] The method provided by this invention uses a drone to capture images of the target to be evaluated for subsequent assessment. This method can image the target from multiple angles, providing more comprehensive information on the damage effect and achieving a more accurate damage assessment.

[0064] Acquire sensor data of the target to be evaluated, including:

[0065] Acquire data collected by multiple sensors on the target under evaluation at different target times, with each target time corresponding to the process of the target under evaluation being hit;

[0066] The collected data is integrated based on the setting location of each sensor to obtain individual sensor data. Each sensor data includes data collected by different types of sensors with the same setting location.

[0067] Using images alone for damage analysis has certain limitations. The method provided by this invention also incorporates sensor data of the target to be evaluated for damage assessment, thereby improving the accuracy of the assessment.

[0068] The multiple sensors installed on the target to be evaluated can be of different types, such as pressure sensors, temperature sensors, and vibration sensors. Different types of sensors collecting different types of data can provide a more comprehensive picture of the impact damage suffered by the target. Combining images of the target with its sensor data to achieve damage assessment can improve the accuracy of the evaluation.

[0069] For sensor-collected data, it can be cleaned before integration. Data cleaning can include processing abnormal data and missing values, and eliminating noise.

[0070] Furthermore, similar to the process of acquiring images of the target to be evaluated described above, in the method provided by this invention, the sensor acquires data at different times during the process of the target being evaluated being hit. This can reflect the changes in the physical quantities of the target being evaluated during the process of being hit, providing more comprehensive information for damage assessment.

[0071] To collect data on changes in physical quantities of a target under evaluation from multiple locations during an attack, the method provided by this invention involves placing sensors at multiple locations on the target. To ensure that the sensor data more accurately reflects the damage at each location of the target, the method integrates the sensor data, combining data from different types of sensors at the same locations to obtain a single sensor data point. This allows for the differentiation of the impact conditions at different locations on the target, improving evaluation accuracy.

[0072] The precision of sensor placement within the target being evaluated can be determined based on the size of the target. Larger targets require lower precision, and vice versa. For example, when the target is a building, sensor placement precision can be down to the meter level, meaning sensors within one meter of each other are considered to be in the same position. Conversely, when the target is a vehicle, sensor placement precision can be down to the centimeter level, meaning sensors within one centimeter of each other are considered to be in the same position. Those skilled in the art can determine the appropriate sensor placement precision based on the specific circumstances.

[0073] After obtaining the target image and sensor data, feature extraction is performed to obtain image features and sensor features.

[0074] Specifically, feature extraction is performed on the target image to obtain image features, including:

[0075] The target image is subjected to feature extraction using a preset feature extraction algorithm to obtain the first feature;

[0076] Feature selection is performed on the first feature to reduce its dimensionality, resulting in the second feature;

[0077] The second feature is transformed to obtain the image features.

[0078] The purpose of feature extraction on a target image is to extract relevant information about the damage. This involves using a pre-defined feature processing algorithm to extract features from the target image, such as texture, shape, and edges. The pre-defined feature processing algorithm can be non-negative matrix factorization, convolutional neural networks, etc.

[0079] Furthermore, such as Figure 2As shown, after extracting features from the target image and obtaining the first feature, feature selection is performed on the first feature to reduce its dimensionality, resulting in the second feature. Feature selection filters the initially extracted features, removing redundant and irrelevant features and retaining those most relevant to the damage. Feature selection can improve the generalization ability and computational efficiency of the damage assessment model. The second feature obtained after feature selection is further transformed to make it more suitable for pattern recognition and classification. Feature transformation methods include feature scaling and feature encoding.

[0080] Feature extraction is performed on sensor data to obtain sensor features. This can be done by directly extracting features using feature extraction algorithms (such as principal component analysis, linear discriminant analysis, nonnegative matrix factorization, convolutional neural networks, etc.) and then performing feature transformations (such as feature scaling, feature encoding, etc.) to obtain sensor features.

[0081] Image and sensor features are input into a trained damage assessment model to obtain the damage assessment results output by the model. The damage assessment model is a deep learning model trained on multiple sets of training data. Each set of training data includes a sample feature set and corresponding damage assessment result labels. The sample feature set includes sample image features and sample sensor features. The acquisition methods for sample image features and sample sensor features are consistent with those described earlier. The damage assessment result labels can be manually annotated by experts.

[0082] Multiple sets of training data constitute the training dataset. The damage assessment model is trained using the training dataset, and the model parameters are adjusted to optimize model performance. The trained model is evaluated using the validation dataset, and metrics such as precision, recall, and F1 score are calculated. Based on the evaluation results, the model is optimized and adjusted.

[0083] After the damage assessment model is trained, it can output damage assessment results by inputting image features and sensor features. In other words, after the damage assessment model is trained, it can not only improve the automation and efficiency of damage assessment and avoid the influence of human factors on the assessment results, but also realize real-time damage assessment and provide timely support for battlefield decision-making.

[0084] Furthermore, such as Figure 2 As shown, after obtaining the damage assessment results output by the damage assessment model, a damage assessment report can be generated based on the damage assessment results. The damage assessment report includes the damage assessment results, target images, sensor data, etc., and is displayed in the form of charts.

[0085] Furthermore, to prevent the limitations of the training data from affecting the robustness of the damage assessment model, the method provided by this invention allows for real-time updates to the damage assessment model during the damage assessment process. That is, after obtaining the damage assessment result output by the damage assessment model, the method further includes:

[0086] Obtain expert evaluation results on the damage assessment findings;

[0087] The damage assessment model was updated based on expert evaluation results.

[0088] Expert evaluation results reflect the assessment of damage assessment results by damage assessment experts. After the damage assessment results are generated, they can be provided to damage assessment experts for evaluation, resulting in expert evaluation results. In essence, expert evaluation results reflect the quality of the damage assessment results output by the damage assessment model. Updating the damage assessment model based on these evaluation results allows for real-time online updates, ensuring the model's accuracy and robustness.

[0089] The intelligent damage assessment device provided by this invention is described below. The intelligent damage assessment device described below can be referred to in correspondence with the intelligent damage assessment method described above. Figure 3 As shown, the intelligent damage assessment device provided by the present invention includes:

[0090] The data acquisition module 310 is used to acquire the target image of the target to be evaluated and to acquire the sensor data of the target to be evaluated.

[0091] The feature extraction module 320 is used to extract features from the target image to obtain image features and to extract features from the sensor data to obtain sensor features.

[0092] The model processing module 330 is used to input image features and sensor features into the trained damage assessment model and obtain the damage assessment results output by the damage assessment model. The damage assessment model is trained based on multiple sets of training data. Each set of training data includes a sample feature group and a damage assessment result label corresponding to the sample feature group. The sample feature group includes sample image features and sample sensor features.

[0093] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4As shown, the electronic device may include a processor 410, a communication interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communication interface 420, and the memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute an intelligent damage assessment method, which includes: acquiring a target image of the target to be assessed and acquiring sensor data of the target to be assessed; performing feature extraction on the target image to obtain image features, performing feature extraction on the sensor data to obtain sensor features; inputting the image features and the sensor features into a trained damage assessment model, and obtaining the damage assessment result output by the damage assessment model, wherein the damage assessment model is trained based on multiple sets of training data, each set of training data includes a sample feature set and a damage assessment result label corresponding to the sample feature set, and the sample feature set includes sample image features and sample sensor features.

[0094] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0095] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the intelligent damage assessment method provided by the above methods. The method includes: acquiring a target image of a target to be assessed and acquiring sensor data of the target to be assessed; extracting features from the target image to obtain image features and extracting features from the sensor data to obtain sensor features; inputting the image features and the sensor features into a trained damage assessment model and obtaining a damage assessment result output by the damage assessment model. The damage assessment model is trained based on multiple sets of training data. Each set of training data includes a sample feature set and a damage assessment result label corresponding to the sample feature set. The sample feature set includes sample image features and sample sensor features.

[0096] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the intelligent damage assessment method provided by the methods described above. The method includes: acquiring a target image of a target to be assessed and acquiring sensor data of the target to be assessed; extracting features from the target image to obtain image features, and extracting features from the sensor data to obtain sensor features; inputting the image features and the sensor features into a trained damage assessment model, and obtaining a damage assessment result output by the damage assessment model. The damage assessment model is trained based on multiple sets of training data, each set of training data including a sample feature set and a damage assessment result label corresponding to the sample feature set. The sample feature set includes sample image features and sample sensor features.

[0097] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0098] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0099] 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. An intelligent damage assessment method, characterized in that, include: Acquire a target image of the target to be evaluated, and acquire sensor data of the target to be evaluated; Feature extraction is performed on the target image to obtain image features, and feature extraction is performed on the sensor data to obtain sensor features; The image features and sensor features are input into the trained damage assessment model to obtain the damage assessment result output by the damage assessment model. The damage assessment model is trained based on multiple sets of training data. Each set of training data includes a sample feature set and a damage assessment result label corresponding to the sample feature set. The sample feature set includes sample image features and sample sensor features. The process of acquiring the target image of the target to be evaluated includes: Multiple images of the target to be evaluated are acquired, each image corresponding to a different target time, and each target time corresponding to the process in which the target to be evaluated is attacked. Based on the display range of the multiple images, the multiple images are corrected to obtain multiple intermediate images, each of which has the same display range; Keyframes are determined from the plurality of intermediate images and used as the target image; The acquisition of multiple images of the target to be evaluated includes: Acquire images captured by the drone; Multiple images of the target to be evaluated are obtained from the images captured by the drone based on the geographical location of the target to be evaluated and the geographical location trajectory of the drone. The process of acquiring multiple images of the target to be evaluated from the images transmitted back by the UAV based on the target's geographical location and the UAV's geographical trajectory includes: The geographic area to be filmed is determined based on the geographic location of the target to be evaluated and the camera parameters of the drone; The images captured when the drone's geographical location is within the captured geographical area are used as multiple images of the target to be evaluated; The acquisition of sensor data of the target to be evaluated includes: Data collected by multiple sensors of the target to be evaluated at different target times are obtained, where each target time corresponds to the process in which the target to be evaluated is attacked. The collected data are integrated based on the setting position of each sensor to obtain each sensor data line, and each sensor data line contains collected data from different types of sensors with the same setting position.

2. The intelligent damage assessment method according to claim 1, characterized in that, The step of extracting features from the target image to obtain image features includes: The image features are extracted using a preset feature extraction algorithm to obtain the first feature; Feature selection is performed on the first feature to reduce its dimensionality, resulting in a second feature; The second feature is transformed to obtain the image feature.

3. The intelligent damage assessment method according to claim 1, characterized in that, After obtaining the damage assessment results output by the damage assessment model, the process includes: Obtain expert evaluation results for the damage assessment results, wherein the expert evaluation results reflect the damage assessment experts' evaluation of the damage assessment results; Based on the expert evaluation results, the damage assessment model is updated.

4. An intelligent damage assessment device, characterized in that, include: The data acquisition module is used to acquire the target image of the target to be evaluated and to acquire the sensor data of the target to be evaluated; The feature extraction module is used to extract features from the target image to obtain image features and to extract features from the sensor data to obtain sensor features. The model processing module is used to input the image features and the sensor features into the trained damage assessment model and obtain the damage assessment result output by the damage assessment model. The damage assessment model is trained based on multiple sets of training data. Each set of training data includes a sample feature set and a damage assessment result label corresponding to the sample feature set. The sample feature set includes sample image features and sample sensor features. The process of acquiring the target image of the target to be evaluated includes: Multiple images of the target to be evaluated are acquired, each image corresponding to a different target time, and each target time corresponding to the process in which the target to be evaluated is attacked. Based on the display range of the multiple images, the multiple images are corrected to obtain multiple intermediate images, each of which has the same display range; Keyframes are determined from the plurality of intermediate images and used as the target image; The acquisition of multiple images of the target to be evaluated includes: Acquire images captured by the drone; Multiple images of the target to be evaluated are obtained from the images captured by the drone based on the geographical location of the target to be evaluated and the geographical location trajectory of the drone. The process of acquiring multiple images of the target to be evaluated from the images transmitted back by the UAV based on the target's geographical location and the UAV's geographical trajectory includes: The geographic area to be filmed is determined based on the geographic location of the target to be evaluated and the camera parameters of the drone; The images captured when the drone's geographical location is within the captured geographical area are used as multiple images of the target to be evaluated; The acquisition of sensor data of the target to be evaluated includes: Data collected by multiple sensors of the target to be evaluated at different target times are obtained, where each target time corresponds to the process in which the target to be evaluated is attacked. The collected data are integrated based on the setting position of each sensor to obtain each sensor data, and each sensor data contains collected data from different types of sensors with the same setting position.

5. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the intelligent damage assessment method as described in any one of claims 1 to 3.

6. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the intelligent damage assessment method as described in any one of claims 1 to 3.

7. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the intelligent damage assessment method as described in any one of claims 1 to 3.

Images