Damage prediction method, device, equipment and storage medium

Abstract

The present application relates to a method, device, equipment and storage medium for damage estimation, the method comprising: obtaining image features of a target object; obtaining posture features of the target object corresponding to the image features; and estimating the damaged part of the target object and the damage value corresponding to the damaged part according to the posture features. The present application is used to timely estimate the degree of damage to people in traffic accidents.

Description

Technical Field

[0001] The present application relates to the field of blockchain technology, and in particular to a damage prediction method, apparatus, device, and storage medium. Background Art

[0002] Injury estimation in traffic accidents is crucial, as it can provide guidance for subsequent rescue efforts. Existing injury estimation methods require medical personnel to arrive at the scene before assessing the extent of injury, which can be time-consuming. Summary of the Invention

[0003] The present application provides a damage estimation method, apparatus, device and storage medium for timely estimating the degree of damage to people in traffic accidents.

[0004] In a first aspect, the present application provides a damage prediction method, which is applied to a server, including:

[0005] Obtain image features of the target object;

[0006] Acquire a posture feature of a target object corresponding to the image feature;

[0007] The damaged part of the target object and the damage value corresponding to the damaged part are estimated according to the posture feature.

[0008] Optionally, obtaining a posture feature of the target object corresponding to the image feature includes:

[0009] Obtaining a posture recovery model from a blockchain node, and inputting the image features into the posture recovery model;

[0010] Obtaining N key points of the target object corresponding to the image features through the posture recovery model, where N is an integer equal to or equal to 1;

[0011] Obtaining the vector corresponding to each key point respectively, and forming a vector set, and using the vector set as the posture feature of the target object;

[0012] Outputting the posture feature of the target object corresponding to the image feature.

[0013] Optionally, the training process of the posture recovery model includes:

[0014] Obtaining an image feature sample set, wherein the image feature sample set includes: M image feature samples, N preset key points corresponding to each of the image feature samples, a preset vector corresponding to each of the preset key points, and a preset vector set consisting of the N preset vectors, wherein M is an integer greater than or equal to 1;

[0015] The following training process is performed on each image feature sample in the image feature sample set: the image feature sample is input into an initial posture recovery model; N sample key points corresponding to the image feature sample and vectors of the N sample key points are obtained through the initial posture recovery model; the vectors of the N sample key points are combined into a sample vector set;

[0016] Calculating a first consistency rate between the sample vector set and the preset vector set;

[0017] If the first consistency rate is not greater than the first preset threshold, after adjusting the parameters in the initial posture recovery model, the training process is repeated until the first consistency rate is greater than the first preset threshold, and the initial posture recovery model is used as the final posture recovery model.

[0018] Optionally, estimating the damaged part of the target object and the damage value corresponding to the damaged part according to the posture feature includes:

[0019] Obtaining a damage prediction model from the blockchain node, and inputting the posture feature into the damage prediction model;

[0020] The damage prediction model outputs the damaged part of the target object and the damage value corresponding to the damaged part.

[0021] Optionally, the training process of the damage prediction model includes:

[0022] Acquire a posture feature sample set, wherein the posture feature sample set includes: S posture feature samples, a preset damage location corresponding to each of the posture feature samples, and a preset damage value corresponding to the preset damage location, wherein S is an integer greater than or equal to 1;

[0023] The following training process is performed on each of the posture feature samples in the posture feature sample set: the posture feature sample is input into an initial damage prediction model; the initial damage prediction model is used to output a sample damage part of the target object and a sample damage value corresponding to the sample damage part;

[0024] Calculating a second consistency rate between the sample damage location and the sample damage value and the preset damage location and the preset damage value;

[0025] If the second consistency rate is not greater than the second preset threshold, after adjusting the parameters in the initial damage prediction model, the training process is repeated until the second consistency rate is greater than the second preset threshold, and the initial damage prediction model is used as the final damage prediction model.

[0026] Optionally, after estimating the damaged part of the target object and the damage value corresponding to the damaged part according to the posture feature, the method further includes:

[0027] Extracting a first hash value of the image feature, a second hash value of the posture feature, and a third hash value of the damage value;

[0028] The first hash value, the second hash value, and the third hash value are stored in the blockchain node.

[0029] In a second aspect, the present application provides a damage prediction method, which is applied to a client, including:

[0030] Collect images of the target object when it touches the target object;

[0031] Extract image features of the image, and send the extracted image features to a server.

[0032] In a third aspect, the present application provides a damage prediction device, comprising:

[0033] A first acquisition module is used to acquire image features of a target object;

[0034] A second acquisition module is used to acquire the posture feature of the target object corresponding to the image feature;

[0035] The estimation module is used to estimate the damaged part of the target object and the damage value corresponding to the damaged part according to the posture feature.

[0036] In a fourth aspect, the present application provides an electronic device comprising: a processor, a communication component, a memory and a communication bus, wherein the processor, the communication component and the memory communicate with each other through the communication bus; the memory is used to store computer programs; the processor is used to execute the programs stored in the memory to implement the damage prediction method described in the first aspect and the damage prediction method described in the second aspect.

[0037] In a fifth aspect, the present application provides a computer-readable storage medium storing a computer program for implementing the damage prediction method described in the first aspect and the damage prediction method described in the second aspect.

[0038] The above-mentioned technical solution provided by the embodiment of the present application has the following advantages over the prior art: the method provided by the embodiment of the present application obtains the image features of the target object. It can be seen that the present application obtains the image features of the target object and does not directly obtain the image of the target object, which effectively protects the user's privacy and data security and reduces the amount of data transmission; further, the posture features of the target object corresponding to the image features are obtained, and based on the posture features, the damaged part of the target object and the damage value corresponding to the damaged part are estimated, which effectively solves the problem that the existing problem of personal injuries in traffic accidents requires medical personnel to arrive at the scene of the accident to estimate the extent of the injury, and can timely estimate the extent of personal injuries in traffic accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0041] Figure 1 This is a schematic diagram of the query system structure based on blockchain in an embodiment of the present application;

[0042] Figure 2 This is a schematic diagram of the blockchain structure in the embodiment of this application;

[0043] Figure 3 This is a schematic diagram of the functional structure of the blockchain network in the embodiment of this application;

[0044] Figure 4 Schematic diagram of the damage prediction method applied to the client in an embodiment of the present application;

[0045] Figure 5 Schematic diagram of the damage prediction method applied to the server in an embodiment of the present application;

[0046] Figure 6 A schematic diagram of a process for obtaining posture features of a target object corresponding to image features in an embodiment of the present application;

[0047] Figure 7 This is a schematic diagram of the structure of the key points of the target object in the embodiment of this application;

[0048] Figure 8 This is a flow chart of estimating the damaged part of a target object and the damage value corresponding to the damaged part in an embodiment of the present application;

[0049] Figure 9 This is a schematic diagram of the structure of the damage estimation device in an embodiment of the present application;

[0050] Figure 10 This is a schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0051] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0052] In the following description, reference is made to “one specific embodiment”, which describes a subset of all possible embodiments, but it is understood that “one specific embodiment” describes the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0053] Unless otherwise defined, all technical terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention pertains. The terms used herein are for the purpose of describing embodiments of the present invention only and are not intended to limit the present invention.

[0054] Before further describing the embodiments of the present invention in detail, the nouns and terms involved in the embodiments of the present invention are explained. The nouns and terms involved in the embodiments of the present invention are subject to the following explanations.

[0055] (1) Blockchain is an encrypted, chained transaction storage structure formed by blocks.

[0056] (2) Blockchain Network: A collection of nodes that incorporate new blocks into the blockchain through consensus.

[0057] (3) Ledger is a general term for the blockchain (also known as ledger data) and the state database synchronized with the blockchain. The blockchain records transactions in the form of files in the file system; the state database records transactions in the blockchain in the form of different types of key-value pairs, which is used to support fast query of transactions in the blockchain.

[0058] (4) Smart Contracts, also known as chaincode or application code, are programs deployed in the nodes of the blockchain network. Nodes (or blockchain nodes) execute smart contracts called in received transactions to update or query key-value pairs in the state database.

[0059] (5) Consensus is a process in the blockchain network used to reach an agreement on the transactions in the block among the multiple nodes involved. The block that reaches an agreement will be appended to the end of the blockchain. The mechanisms for achieving consensus include Proof of Work (PoW), Proof of Stake (PoS), Delegated Proof-of-Stake (DPoS), Proof of Elapsed Time (PoET), etc.

[0060] The following describes an exemplary application of the blockchain network provided by the embodiment of the present invention, such as Figure 1 As shown, Figure 1 1 is a schematic diagram of a blockchain-based query system provided by an embodiment of the present invention, including a blockchain network 101, a consensus node 102, an authentication center 103, a business entity 104, a client node 104-1, a business entity 105, and a client node 105-1, which are described below respectively:

[0061] The types of blockchain network 101 are flexible and diverse, and can be, for example, public, private, or consortium blockchains. For example, with a public blockchain, any business entity's electronic device, such as a user terminal or server, can access blockchain network 101 without authorization. For example, with a consortium blockchain, a business entity can access blockchain network 101 with authorization, and its electronic devices (such as terminals / servers) become client nodes in blockchain network 101.

[0062] In some embodiments, client node 104 may only serve as an observer of blockchain network 101, providing support for business entities to initiate transactions (e.g., storing data on-chain or querying on-chain data). The client node may implement the functions of consensus node 102 of blockchain network 101, such as sorting, consensus services, and ledger functions, by default or selectively (e.g., depending on the specific business needs of the business entity). This allows the business entity's data and business processing logic to be migrated to blockchain network 101 to the greatest extent possible, ensuring the trustworthiness and traceability of data and business processing processes through blockchain network 101.

[0063] The consensus nodes in the blockchain network 101 receive data from different business entities, such as Figure 1 The transaction submitted by the client node 104 - 1 of the business entity 104 shown in the figure is executed to update the ledger or query the ledger. Various intermediate results or final results of the transaction can be returned to the client node 104 - 1 of the business entity 104 for display.

[0064] For example, client node 104-1 can subscribe to events of interest in blockchain network 101, such as transactions occurring in a specific organization / channel in blockchain network 101, and consensus node 102 pushes corresponding transaction notifications to client node 104-1, thereby triggering corresponding business logic in client node 104-1.

[0065] As an example of blockchain, Figure 2 As shown, Figure 2 This is a schematic diagram of the structure of the blockchain in the blockchain network 101 provided by an embodiment of the present invention. The header of each block can include the hash values ​​of all transactions in the block as well as the hash values ​​of all transactions in the previous block. The records of newly generated transactions are filled into the block and, after consensus among the nodes in the blockchain network, are appended to the end of the blockchain to form a chain-like growth. The chain structure between blocks based on hash values ​​ensures that transactions in the block are tamper-proof and anti-forgery.

[0066] The following describes an exemplary functional architecture of the blockchain network provided by an embodiment of the present invention. Figure 3 As shown, Figure 3 1 is a functional architecture diagram of a blockchain network 101 provided in an embodiment of the present invention, including an application layer 301, a consensus layer 302, a network layer 303, a data layer 304, and a resource layer 305, which are described below:

[0067] The application layer 301 encapsulates various services that can be implemented by the blockchain network, including transaction traceability, evidence storage, and verification.

[0068] The consensus layer 302 encapsulates the mechanism by which nodes 102 in the blockchain network 101 reach consensus on blocks (i.e., the consensus mechanism), as well as transaction management and ledger management functions. Consensus mechanisms include consensus algorithms such as POS, POW, and DPOS, and support pluggable consensus algorithms. Transaction management verifies the digital signatures carried in transactions received by node 101, verifies the identity information of business entities 104, and, based on this identity information, determines whether they have the authority to conduct transactions (by reading relevant information from the business entity identity management). Business entities authorized to access the blockchain network 101 possess digital certificates issued by a certification center. They use the private keys in their digital certificates to sign submitted transactions, thereby asserting their legal identity. Ledger management maintains the blockchain and state database. Consensus blocks are appended to the end of the blockchain. Transactions in these consensus blocks are executed. If the transaction includes an update operation, the key-value pairs in the state database are updated. If the transaction includes a query operation, the key-value pairs in the state database are queried and the query results are returned to the business entity's client node. Supports query operations on multiple dimensions of the state database, including: querying blocks based on block serial numbers (such as transaction hash values); querying blocks based on block hash values; querying blocks based on transaction serial numbers; querying transactions based on transaction serial numbers; querying business entity account data based on business entity account (serial number); and querying the blockchain in a channel based on the channel name.

[0069] The network layer 303 encapsulates the functions of point-to-point (P2P) network protocol, data transmission mechanism and data verification mechanism, access authentication mechanism and business subject identity management.

[0070] Among them, the P2P network protocol realizes the communication between nodes 102 in the blockchain network 101, the data propagation mechanism ensures the propagation of transactions in the blockchain network 101, and the data verification mechanism is used to realize the reliability of data transmission between nodes 102 based on cryptographic methods (such as digital certificates, digital signatures, public / private key pairs); the access authentication mechanism is used to authenticate the identity of the business entity joining the blockchain network 101 according to the actual business scenario, and grant the business entity the right to access the blockchain network 101 when the authentication is passed; the business entity 104 identity management is used to store the identity of the business entity 104 allowed to access the blockchain network 101, and the authority (such as the type of transaction that can be initiated).

[0071] The data layer 304 encapsulates various data structures that implement the ledger, including the blockchain implemented as files in the file system, the key-value state database, and the existence proof (such as the hash tree of transactions in the block).

[0072] The resource layer 305 encapsulates the computing resources, storage resources, and communication resources of each node 102 in the blockchain network 101.

[0073] Based on the above architecture, the embodiment of the present invention proposes the following implementation methods.

[0074] The embodiment of the present application provides a damage prediction method, which is applied in the client. The specific implementation of the method is as follows: Figure 4 As shown:

[0075] Step 401: Capture an image of the target object when it is touched.

[0076] The target object can be a person or a vehicle. The following example uses a person as the target object. However, it should be noted that the examples in the embodiments of this application are only examples and are not used to limit the scope of protection. In addition, the examples in other embodiments of this application are also only examples and are not used to limit the scope of protection. They will not be explained one by one below.

[0077] For example, a vehicle dashcam or road video surveillance equipment is used to capture images of the target object when it collides with another object. The captured images are then preprocessed, such as resizing them to a preset size, for example, 64 ps x 64 ps, and converting the resized images into grayscale images.

[0078] Step 402: extract image features of the image and send the extracted image features to the server.

[0079] For example, an autoencoder can be used to extract image features. An autoencoder is an unsupervised neural network model that learns the implicit features of the input data. Specifically, an image is input to the autoencoder, and the autoencoder outputs the image features. For example, a five-layer autoencoder can be used, and the implicit features output by the last layer are used as the image features of the image. These features are represented as 64*64-dimensional vectors.

[0080] Furthermore, the extracted image features are sent to the server. It can be seen that this application does not send the image directly to the server, but sends the extracted image features to the server, which effectively protects the user's privacy and data security, reduces the amount of data transmission, and improves the transmission rate.

[0081] The embodiment of the present application provides a damage prediction method, which is applied in a server. The specific implementation of the method is as follows: Figure 5 As shown:

[0082] Step 501: Acquire image features of a target object.

[0083] Specifically, the image features of the target object sent by the client are obtained.

[0084] Step 502: Acquire the posture features of the target object corresponding to the image features.

[0085] In a specific embodiment, a posture recovery model is trained to obtain the posture features of the target object corresponding to the image features through the posture recovery model. The posture recovery model is specifically obtained as follows:

[0086] First, an initial posture recovery model is constructed.

[0087] Then, an image feature sample set is obtained, wherein the image feature sample set includes: M image feature samples, N preset key points corresponding to each image feature sample, a preset vector corresponding to each preset key point, and a preset vector set consisting of N preset vectors, wherein M is an integer greater than or equal to 1, and N is an integer greater than or equal to 1.

[0088] Secondly, the following training process is performed on each image feature sample in the image feature sample set: the image feature sample is input into the initial posture recovery model; through the initial posture recovery model, N sample key points corresponding to the image feature sample and the vectors of the N sample key points are obtained; the vectors of the N sample key points are combined into a sample vector set.

[0089] Then, a first consistency rate between the sample vector set and the preset vector set is calculated.

[0090] Finally, if the first consistency rate is not greater than the first preset threshold, the parameters in the initial posture recovery model are adjusted and the training process is repeated until the first consistency rate is greater than the first preset threshold, and the initial posture recovery model is used as the final posture recovery model.

[0091] In addition, after the posture recovery model training is completed, the hash value of the posture recovery model is obtained through the hash algorithm, and the hash value of the posture recovery model is stored in the block through the blockchain node to ensure the security of the data.

[0092] In a specific embodiment, the specific process of obtaining the posture feature of the target object corresponding to the image feature is as follows: Figure 6 As shown:

[0093] Step 601: Obtain a posture recovery model from a blockchain node and input image features into the posture recovery model.

[0094] Step 602: Obtain N key points of the target object corresponding to the image features through the posture recovery model.

[0095] Here, we take N equal to 15 as an example to illustrate that 15 key points of the target object corresponding to the image features are obtained through the posture recovery model, as follows: Figure 7 As shown, they are key point 701, key point 702, key point 703, key point 704, key point 705, key point 706, key point 707, key point 708, key point 709, key point 7010, key point 7011, key point 7012, key point 7013, key point 7014, and key point 7015.

[0096] Step 603: Obtain the vectors corresponding to each key point respectively, form a vector set, and use the vector set as the posture feature of the target object.

[0097] Specifically, the vector corresponding to each key point is a three-dimensional vector. A three-dimensional rectangular coordinate system is established with an arbitrary point as the center point, and the three-dimensional coordinates of each key point are obtained respectively to form a 45-dimensional vector set.

[0098] Step 604: Output the posture features of the target object corresponding to the image features.

[0099] Step 503 : Estimate the damaged part of the target object and the damage value corresponding to the damaged part according to the posture feature.

[0100] In a specific embodiment, a damage prediction model is trained to be able to estimate the damage location of the target object and the damage value corresponding to the damage location through the damage prediction model. The specific damage prediction model is obtained as follows:

[0101] First, an initial damage prediction model is established.

[0102] Then, a posture feature sample set is obtained, wherein the posture feature sample set includes: S posture feature samples, a preset damage location corresponding to each posture feature sample, and a preset damage value corresponding to the preset damage location, wherein S is an integer greater than or equal to 1.

[0103] Secondly, the following training process is performed on each posture feature sample in the posture feature sample set: the posture feature sample is input into the initial damage prediction model; through the initial damage prediction model, the sample damage part of the target object and the sample damage value corresponding to the sample damage part are output.

[0104] Furthermore, a second consistency rate between the sample damage location and the sample damage value and the preset damage location and the preset damage value is calculated.

[0105] Finally, if the second consistency rate is not greater than the second preset threshold, the parameters in the initial damage prediction model are adjusted and the training process is repeated until the second consistency rate is greater than the second preset threshold, and the initial damage prediction model is used as the final damage prediction model.

[0106] The second consistency rate between the sample damage site and the preset damage site, and the second consistency rate between the sample damage value and the preset damage value are both no greater than the second preset threshold.

[0107] In addition, after the damage prediction model training is completed, the hash value of the damage prediction model is obtained through the hash algorithm, and the hash value of the damage prediction model is stored in the block through the blockchain node to ensure the security of the data.

[0108] In a specific embodiment, the specific process of estimating the damage part of the target object and the damage value corresponding to the damage part according to the posture characteristics is as follows: Figure 8 As shown:

[0109] Step 801: Obtain a damage prediction model from a blockchain node and input posture features into the damage prediction model.

[0110] Specifically, a damage prediction model is obtained from the blockchain node, and a 45-dimensional vector set is input into the damage prediction model.

[0111] Step 802: Output the damaged part of the target object and the damage value corresponding to the damaged part through the damage prediction model.

[0112] Specifically, the damage value is the probability of damage. Through the damage prediction model, a 6-dimensional vector is output, which respectively represents the damage values ​​of the human head, torso, left hand to left arm, right hand to right arm, left foot to left leg, and right foot to right leg. The value can be expressed in percentage or in the form of a fraction.

[0113] In a specific embodiment, after estimating the damage location of the target object and the damage value corresponding to the damage location through a damage prediction model, a first hash value of the image feature, a second hash value of the posture feature, and a third hash value of the damage value are extracted; and the first hash value, the second hash value, and the third hash value are stored in a blockchain node, and stored in a block through the blockchain node to ensure data security.

[0114] The method provided in the embodiment of the present application obtains the image features of the target object. It can be seen that the present application obtains the image features of the target object, and does not directly obtain the image of the target object, which effectively protects the user's privacy and data security and reduces the amount of data transmission; further, the posture features of the target object corresponding to the image features are obtained, and based on the posture features, the damaged part of the target object and the damage value corresponding to the damaged part are estimated, which effectively solves the problem that the existing problem of personal injuries in traffic accidents requires medical personnel to arrive at the scene of the accident to estimate the extent of the injury, and can timely estimate the extent of personal injuries in traffic accidents.

[0115] The present application also provides a damage prediction device. The specific implementation of the device can be found in the description of the method embodiment part, and the repeated parts will not be repeated. Figure 9 As shown, the device mainly includes:

[0116] The first acquisition module 901 is used to acquire image features of a target object.

[0117] The second acquisition module 902 is used to acquire the posture feature of the target object corresponding to the image feature.

[0118] The estimation module 903 is used to estimate the damaged part of the target object and the damage value corresponding to the damaged part according to the posture characteristics.

[0119] Based on the same concept, an electronic device is also provided in the embodiment of the present application, such as Figure 10 As shown, the electronic device mainly includes: a processor 1001, a communication component 1002, a memory 1003 and a communication bus 1004, wherein the processor 1001, the communication component 1002 and the memory 1003 communicate with each other via the communication bus 1004. The memory 1003 stores a program that can be executed by the processor 1001. The processor 1001 executes the program stored in the memory 1003 to implement the following steps: obtaining image features of the target object; obtaining posture features of the target object corresponding to the image features; estimating the damage part of the target object and the damage value corresponding to the damage part based on the posture features, and collecting images of the target object when it is touched; extracting image features from the image, and sending the extracted image features to the server.

[0120] The communication bus 1004 mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The communication bus 1004 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 10 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0121] The communication component 1002 is used for communication between the electronic device and other devices.

[0122] The memory 1003 may include a random access memory (RAM) or a non-volatile memory, such as at least one disk storage. Alternatively, the memory may be at least one storage device located away from the processor 1001.

[0123] The above-mentioned processor 1001 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc., and can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components.

[0124] In another embodiment of the present application, a computer-readable storage medium is provided, in which a computer program is stored. When the computer program runs on a computer, the computer executes the damage prediction method described in the above embodiment.

[0125] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When software is used for implementation, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions are transmitted from a website, computer, server or data center to another website, computer, server or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, microwave, etc.) method. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrations. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape, etc.), an optical medium (e.g., a DVD) or a semiconductor medium (e.g., a solid-state hard disk), etc.

[0126] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0127] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is intended to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A damage prediction method, characterized in that: Applied to the server, including: Obtain image features of the target object; Obtaining the posture feature of the target object corresponding to the image feature, including: obtaining a posture recovery model from a blockchain node, and inputting the image feature into the posture recovery model; obtaining N key points of the target object corresponding to the image feature through the posture recovery model, where N is an integer equal to or equal to 1; respectively obtaining vectors corresponding to each of the key points, and forming a vector set, and using the vector set as the posture feature of the target object; and outputting the posture feature of the target object corresponding to the image feature; According to the posture feature, the damage part of the target object and the damage value corresponding to the damage part are estimated, including: obtaining a damage estimation model from the blockchain node, and inputting the posture feature into the damage estimation model; outputting the damage part of the target object and the damage value corresponding to the damage part through the damage estimation model; the training process of the damage estimation model includes: obtaining a posture feature sample set, wherein the posture feature sample set includes: S posture feature samples, a preset damage part corresponding to each of the posture feature samples, and a preset damage value corresponding to the preset damage part, wherein S is an integer greater than or equal to 1; respectively training the posture feature samples, wherein the posture feature samples are trained on the preset damage part, and the damage value corresponding to the preset damage part is trained on the preset damage part. Each of the posture feature samples in the feature sample set executes the following training process: inputting the posture feature sample into an initial damage prediction model; outputting the sample damage part of the target object and the sample damage value corresponding to the sample damage part through the initial damage prediction model; calculating the second consistency rate between the sample damage part and the sample damage value and the preset damage part and the preset damage value; if the second consistency rate is not greater than the second preset threshold, adjusting the parameters in the initial damage prediction model, and repeating the training process until the second consistency rate is greater than the second preset threshold, and then using the initial damage prediction model as the final damage prediction model.

2. The damage prediction method according to claim 1, characterized in that: The training process of the posture recovery model includes: Acquire an image feature sample set, wherein the image feature sample set includes: M image feature samples, N preset key points corresponding to each of the image feature samples, a preset vector corresponding to each of the preset key points, and a preset vector set consisting of the N preset vectors, wherein M is an integer greater than or equal to 1; The following training process is performed on each of the image feature samples in the image feature sample set: the image feature sample is input into an initial posture recovery model; through the initial posture recovery model, N sample key points corresponding to the image feature sample and the vectors of the N sample key points are obtained; the vectors of the N sample key points are combined into a sample vector set; Calculating a first consistency rate between the sample vector set and the preset vector set; If the first consistency rate is not greater than the first preset threshold, after adjusting the parameters in the initial posture recovery model, the training process is repeated until the first consistency rate is greater than the first preset threshold, and the initial posture recovery model is used as the final posture recovery model.

3. The damage prediction method according to any one of claims 1 to 2, characterized in that: After estimating the damaged part of the target object and the damage value corresponding to the damaged part according to the posture feature, the method further includes: Extracting a first hash value of the image feature, a second hash value of the posture feature, and a third hash value of the damage value; The first hash value, the second hash value, and the third hash value are stored in the blockchain node.

4. A damage prediction method, characterized in that: Applied to the client, including: Collecting images of the target object when it touches; Extract image features of the image, and send the extracted image features to a server so that the server executes the damage prediction method applied to the server according to any one of claims 1 to 3.

5. A damage prediction device, characterized in that: include: A first acquisition module, used to acquire image features of a target object; The second acquisition module is used to obtain the posture feature of the target object corresponding to the image feature, including: obtaining a posture recovery model from a blockchain node, and inputting the image feature into the posture recovery model; obtaining N key points of the target object corresponding to the image feature through the posture recovery model, where N is an integer equal to or equal to 1; respectively obtaining vectors corresponding to each of the key points, and forming a vector set, and using the vector set as the posture feature of the target object; and outputting the posture feature of the target object corresponding to the image feature. The estimation module is used to estimate the damage part of the target object and the damage value corresponding to the damage part according to the posture feature, including: obtaining a damage estimation model from the blockchain node, and inputting the posture feature into the damage estimation model; outputting the damage part of the target object and the damage value corresponding to the damage part through the damage estimation model; the training process of the damage estimation model includes: obtaining a posture feature sample set, wherein the posture feature sample set includes: S posture feature samples, a preset damage part corresponding to each of the posture feature samples, and a preset damage value corresponding to the preset damage part, wherein S is an integer greater than or equal to 1; respectively Each posture feature sample in the posture feature sample set performs the following training process: inputting the posture feature sample into an initial damage prediction model; outputting the sample damage part of the target object and the sample damage value corresponding to the sample damage part through the initial damage prediction model; calculating the second consistency rate between the sample damage part and the sample damage value and the preset damage part and the preset damage value; if the second consistency rate is not greater than the second preset threshold, adjusting the parameters in the initial damage prediction model, and repeating the training process until the second consistency rate is greater than the second preset threshold, and then using the initial damage prediction model as the final damage prediction model.

6. An electronic device, characterized in that: include: A processor, a communication component, a memory and a communication bus, wherein the processor, the communication component and the memory communicate with each other via the communication bus; The memory is used to store computer programs; The processor is used to execute the program stored in the memory to implement the damage prediction method applied to the server as described in any one of claims 1-3, and to implement the damage prediction method applied to the client as described in claim 4.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it implements the damage prediction method applied to the server as described in any one of claims 1-3, and implements the damage prediction method applied to the client as described in claim 4.

Images