A target positioning method and system based on thermal imaging analysis

Abstract

The embodiment of the application provides a target positioning method and system based on thermal imaging analysis, all image semantic features of target object surface thermal imaging data are clustered into target number of groups by a feature group strategy, all image semantic features contained by the groups are fused to generate group image features of the groups, so that the group image features contain feature relationships between all image semantic features, the group image features of the groups are input into a key feature area positioning model meeting a model convergence condition to generate image semantic features of estimated key feature areas of the groups, and a foreground image partition corresponding to an image semantic feature most matching the image semantic features of the estimated key feature areas of the groups in all image semantic features is taken as a key feature area of the groups, so that the image semantic features of the key feature area of the groups are extracted, and the reliability of key feature area extraction can be improved by combining the feature group strategy.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a target positioning method and system based on thermal imaging analysis. Background Art

[0002] At present, with the progress of socialization, people are gradually beginning to experience sub-health conditions. Symptoms such as waist and leg pain and stiff shoulders and backs seriously affect people's quality of life. Traditional Chinese medicine massage techniques can safely and effectively alleviate this situation and can achieve therapeutic effects by massaging specific acupuncture points on the human body. Currently, traditional Chinese medicine massage machines need to be combined with effective acupuncture point positioning during the massage process to achieve effective massage control accuracy. However, for the positioning method based on fixed coordinate points in the existing scheme, the positioning accuracy is poor due to the different actual conditions of each individual. Therefore, the inventors have discovered a target positioning scheme for key feature areas based on thermal imaging analysis technology. However, in the related technology, the reliability of key feature area extraction cannot meet actual needs. Summary of the Invention

[0003] In view of this, the purpose of this application is to provide a target positioning method and system based on thermal imaging analysis.

[0004] According to a first aspect of the present application, a target positioning method based on thermal imaging analysis is provided, which is applied to a target positioning system based on thermal imaging analysis. The method comprises:

[0005] Performing image segmentation on the target object surface thermal imaging data to generate a plurality of thermal imaging image segments of the target object surface thermal imaging data;

[0006] respectively parsing foreground image partitions in the plurality of thermal imaging image blocks into image semantic features;

[0007] Clustering all image semantic features of the target object surface thermal imaging data into a target number of clusters based on a feature clustering strategy;

[0008] Performing feature fusion on all image semantic features contained in each cluster in the surface thermal imaging data of the target object to determine the cluster image features of each cluster;

[0009] Inputting the cluster image features of each cluster into a key feature area positioning model that meets the model convergence condition to generate image semantic features of the estimated key feature areas of each cluster; wherein the key feature area positioning model is generated by iteratively optimizing the AI ​​network model based on the cluster image features of each cluster of each thermal imaging image sample in the first thermal imaging image sample set as model loading data and the image semantic features of the estimated key feature areas of the corresponding cluster as model output data;

[0010] Calculating the matching value between each image semantic feature of each cluster and the image semantic feature of the estimated key feature area of ​​each cluster respectively, and determining the foreground image partition corresponding to the image semantic feature with the largest matching value among all the image semantic features of each cluster as the key feature area of ​​each cluster;

[0011] The key feature area of ​​the target object surface thermal imaging data is extracted based on the key feature area of ​​each cluster of the target object surface thermal imaging data.

[0012] In a possible implementation of the first aspect, respectively calculating a matching value between each image semantic feature of each cluster and an image semantic feature of an estimated key feature area of ​​each cluster, and determining a foreground image partition corresponding to an image semantic feature having a maximum matching value among all image semantic features of each cluster as the key feature area of ​​each cluster, includes:

[0013] Calculating the feature costs of each image semantic feature of each cluster and the image semantic features of the estimated key feature areas of each cluster respectively;

[0014] The foreground image partition corresponding to the image semantic feature with the smallest feature cost between all the image semantic features of the clusters and the image semantic features of the estimated key feature area of ​​the clusters is determined as the key feature area of ​​the clusters.

[0015] In a possible implementation of the first aspect, performing image segmentation on the target object surface thermal imaging data to generate multiple thermal imaging image blocks of the target object surface thermal imaging data includes:

[0016] Segmenting the surface thermal imaging data of the target object based on a preset feature matching template to generate a plurality of initial thermal imaging image blocks;

[0017] Noise feature cleaning is performed on the multiple initial thermal imaging image blocks based on a preset noise feature template library to generate multiple thermal imaging image blocks of thermal imaging data of the target object surface.

[0018] In a possible implementation of the first aspect, the method further includes:

[0019] Based on the model training request of the key feature area positioning model, performing image segmentation on each thermal imaging image sample in the first thermal imaging image sample set to generate multiple thermal imaging image segments for each thermal imaging image sample;

[0020] Respectively parsing foreground image partitions in a plurality of thermal imaging image blocks of the thermal imaging image sample into image semantic features; clustering all image semantic features of the thermal imaging image sample into a preset number of clusters based on a feature clustering strategy; wherein the preset number is the number of estimated key feature areas in the thermal imaging image sample;

[0021] For each estimated key feature area of ​​the thermal imaging image sample, determining a cluster corresponding to an image semantic feature that best matches the image semantic feature of the corresponding estimated key feature area among all image semantic features of the thermal imaging image sample, and determining the cluster as the cluster corresponding to the image semantic feature of the corresponding estimated key feature area;

[0022] Performing feature fusion on all image semantic features contained in each cluster in the thermal imaging image sample to determine the cluster image features of each cluster in the thermal imaging image sample;

[0023] The AI ​​network model is iteratively optimized based on the cluster image features of each cluster of each thermal imaging image sample in the first thermal imaging image sample set as model loading data and the image semantic features of the estimated key feature area of ​​the corresponding cluster as model output data to output the key feature area positioning model.

[0024] In a possible implementation of the first aspect, respectively parsing the foreground image partitions in the plurality of thermal imaging image blocks into image semantic features includes:

[0025] Parsing the foreground image partitions in the thermal imaging image blocks into image semantic features based on a previously trained image semantic feature encoding model;

[0026] The method further comprises:

[0027] Based on the training request of the image semantic feature encoding model, obtaining a second thermal imaging image sample set;

[0028] Performing initialization quality improvement on the second thermal imaging image sample set, wherein the initialization quality improvement includes image segmentation and noise feature cleaning;

[0029] Based on a convolutional neural network algorithm, feature coding learning is performed on the second thermal imaging image sample set after initialization quality improvement to generate the image semantic feature coding model.

[0030] In a possible implementation of the first aspect, respectively parsing the foreground image partitions in the plurality of thermal imaging image blocks into image semantic features includes:

[0031] parsing a foreground image partition in a plurality of thermal imaging image blocks of the target object surface thermal imaging data into image semantic features having a set dimension;

[0032] The step of respectively parsing the foreground image partitions in the plurality of thermal imaging image blocks of the thermal imaging image sample into image semantic features comprises:

[0033] Parsing the foreground image partitions in the plurality of thermal imaging image blocks of the thermal imaging image sample into image semantic features having the set dimension.

[0034] The method further comprises:

[0035] Performing image blocking on each thermal imaging image sample in the first thermal imaging image sample set to generate a plurality of thermal imaging image blocks for each thermal imaging image sample;

[0036] respectively parsing foreground image partitions in a plurality of thermal imaging image blocks of the thermal imaging image sample into image semantic features;

[0037] Clustering all image semantic features of the thermal imaging image sample into a preset number of clusters based on a feature clustering strategy; wherein the preset number is the number of estimated key feature areas in the thermal imaging image sample;

[0038] For each estimated key feature area of ​​the thermal imaging image sample, determining a cluster corresponding to an image semantic feature that best matches the image semantic feature of the corresponding estimated key feature area among all image semantic features of the thermal imaging image sample, and determining the cluster as the cluster corresponding to the image semantic feature of the corresponding estimated key feature area;

[0039] Performing feature fusion on all image semantic features contained in each cluster in the thermal imaging image sample to determine the cluster image features of each cluster;

[0040] The AI ​​network model is iteratively optimized based on the cluster image features of each cluster of each thermal imaging image sample in the first thermal imaging image sample set as model loading data and the image semantic features of the estimated key feature area of ​​the corresponding cluster as model output data to output a key feature area positioning model.

[0041] In a possible implementation of the first aspect, respectively parsing foreground image partitions in the plurality of thermal imaging image blocks of the thermal imaging image sample into image semantic features includes:

[0042] Parsing the foreground image partitions in the thermal imaging image blocks into image semantic features based on a previously trained image semantic feature encoding model;

[0043] The method further comprises:

[0044] obtain a second thermal image sample set based on the training request of the image semantic feature coding model;

[0045] perform initialization quality improvement on the second thermal image sample set, the initialization quality improvement including image blocking and noise feature cleaning;

[0046] perform feature coding learning on the second thermal image sample set after the initialization quality improvement based on a convolutional neural network algorithm, to generate the image semantic feature coding model.

[0047] According to a second aspect of the present application, a server is provided, which includes a machine readable storage medium and a processor, the machine readable storage medium stores machine executable instructions, and the processor, when executing the machine executable instructions, implements the target positioning method based on thermal imaging analysis as described above.

[0048] According to a third aspect of the present application, a computer readable storage medium is provided, which stores computer executable instructions, and when the computer executable instructions are executed, the target positioning method based on thermal imaging analysis as described above is implemented.

[0049] According to the above aspects, the present application clusters all image semantic features of the target object surface thermal imaging data into a target number of clusters through feature clustering strategy, that is, each cluster is considered to contain a key feature area, all image semantic features contained in the cluster are fused to generate cluster image features of the cluster, so that the cluster image features contain the feature relationships between all image semantic features, the cluster image features of the cluster are input into the key feature area positioning model that meets the model convergence condition to generate the estimated key feature area image semantic features of the cluster, that is, the key feature area positioning model generated by model training of each cluster through pre-training AI network model is used to obtain the estimated key feature area of the cluster, and the foreground image partition corresponding to the image semantic feature that best matches the estimated key feature area image semantic features of the cluster is used as the key feature area of the cluster, thereby realizing the extraction of the key feature area image semantic features of the cluster, and through the combination of the feature clustering strategy, the reliability of the key feature area extraction can be improved. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without any creative work.

[0051] Figure 1 A schematic diagram of a target positioning method based on thermal imaging analysis provided in an embodiment of the present application;

[0052] Figure 2 A schematic diagram of the component structure of a server provided in an embodiment of the present application for implementing the above-mentioned target positioning method based on thermal imaging analysis is shown. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below based on the drawings in the embodiments of the present application. It should be understood that the drawings in the present application are only for the purpose of illustration and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn according to the physical occupancy rate. The flowcharts used in this application illustrate the operations implemented in some embodiments of the embodiments of the present application. It should be understood that the operations of the flowchart can be implemented out of sequence, and steps without logical context can be reversed or implemented simultaneously. In addition, those skilled in the art, under the guidance of the contents of this application, can add at least one other operation to the flowchart or remove at least one operation from the flowchart.

[0054] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of the present application. According to the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work fall within the scope of protection of the present application.

[0055] Figure 1 The following is a schematic flow chart of a target location method based on thermal imaging analysis provided in an embodiment of the present application. It should be understood that in other embodiments, the order of some steps in the target location method based on thermal imaging analysis of this embodiment may be interchanged, or some steps may be omitted or deleted, depending on actual needs. The detailed steps of the target location method based on thermal imaging analysis are described below.

[0056] Step 101, performing image segmentation on the thermal imaging data of the target object surface to generate a plurality of thermal imaging image segments of the thermal imaging data of the target object surface;

[0057] Step 102, respectively parsing the foreground image partitions in the plurality of thermal imaging image blocks into image semantic features;

[0058] Step 103, clustering all image semantic features of the target object surface thermal imaging data into a target number of clusters based on a feature clustering strategy;

[0059] Step 104: performing feature fusion on all image semantic features contained in each cluster in the thermal imaging data of the target object surface to determine the cluster image features of each cluster;

[0060] Step 105: Input the cluster image features of each cluster into a key feature area positioning model that meets the model convergence condition to generate image semantic features of the estimated key feature areas of each cluster; wherein the key feature area positioning model is generated by iteratively optimizing the AI ​​network model based on the cluster image features of each cluster of each thermal imaging image sample in the first thermal imaging image sample set as model loading data and the image semantic features of the estimated key feature areas of the corresponding cluster as model output data;

[0061] Step 106, respectively calculating the matching value between each image semantic feature of each cluster and the image semantic feature of the estimated key feature area of ​​each cluster, and determining the foreground image partition corresponding to the image semantic feature with the largest matching value among all the image semantic features of each cluster as the key feature area of ​​each cluster;

[0062] Step 107: extracting key feature areas of the target object surface thermal imaging data based on the key feature areas of each cluster of the target object surface thermal imaging data.

[0063] The estimated key feature areas may be key feature areas determined based on manually determined standards in the thermal imaging image samples, such as acupuncture points on the human body surface, etc. The first thermal imaging image sample set includes thermal imaging data of the human body surface.

[0064] For example, if the number of targets is 3; all image semantic features of the surface thermal imaging data of the target object are clustered into 3 clusters based on the feature clustering strategy, then each cluster includes multiple image semantic features; if the first cluster includes image semantic feature 1 (k1, k2, k3) and image semantic feature 2 (h1, h2, h3), then the cluster image features of the cluster are (k1+h1, k2+h2, k3+h3).

[0065] Based on the above steps, all image semantic features of the thermal imaging data of the target object surface are clustered into a target number of clusters through a feature clustering strategy, that is, it is considered that each cluster contains a key feature area, and all image semantic features contained in the cluster are feature fused to generate cluster image features of the cluster, so that the cluster image features contain the feature relationship between all image semantic features, and the cluster image features of the cluster are input into a key feature area positioning model that meets the model convergence conditions to generate image semantic features of the estimated key feature area of ​​the cluster, that is, the estimated key feature area of ​​the cluster is obtained by the key feature area positioning model generated by training each cluster with an AI network model trained in advance, and the foreground image partition corresponding to the image semantic feature that best matches the image semantic feature of the estimated key feature area of ​​the cluster among all image semantic features is used as the key feature area of ​​the cluster, thereby realizing the extraction of image semantic features of the key feature area of ​​the cluster, and by combining the feature clustering strategy, the reliability of the extraction of the key feature area can be improved.

[0066] In an alternative embodiment, the feature clustering strategy may be a k-means clustering algorithm.

[0067] In an alternative embodiment, determining the foreground image partition corresponding to the image semantic feature that best matches the image semantic feature of the estimated key feature area of ​​each cluster among all the image semantic features of each cluster as the key feature area of ​​each cluster includes:

[0068] Calculating the feature cost of each image semantic feature of each cluster and the image semantic feature of the estimated key feature area of ​​each cluster respectively; wherein the feature cost is a cosine feature cost or a Euclidean feature cost;

[0069] The foreground image partition corresponding to the image semantic feature with the smallest feature cost between all the image semantic features of the clusters and the image semantic features of the estimated key feature area of ​​the clusters is determined as the key feature area of ​​the clusters.

[0070] The feature cost may refer to a feature distance, which may be a cosine distance.

[0071] In an alternative embodiment, the step of performing image segmentation on the target object surface thermal imaging data to generate a plurality of thermal imaging image segments of the target object surface thermal imaging data includes:

[0072] Segmenting the surface thermal imaging data of the target object based on a preset feature matching template to generate a plurality of initial thermal imaging image blocks;

[0073] Noise feature cleaning is performed on the multiple initial thermal imaging image blocks based on a preset noise feature template library to generate multiple thermal imaging image blocks of thermal imaging data of the target object surface.

[0074] In an alternative embodiment, the method further comprises:

[0075] Based on the model training request of the key feature area positioning model, performing image segmentation on each thermal imaging image sample in the first thermal imaging image sample set to generate multiple thermal imaging image segments for each thermal imaging image sample;

[0076] respectively parsing foreground image partitions in a plurality of thermal imaging image blocks of the thermal imaging image sample into image semantic features;

[0077] Clustering all image semantic features of the thermal imaging image sample into a preset number of clusters based on a feature clustering strategy; wherein the preset number is the number of estimated key feature areas in the thermal imaging image sample;

[0078] For each estimated key feature area of ​​the thermal imaging image sample, determining a cluster corresponding to an image semantic feature that best matches the image semantic feature of the corresponding estimated key feature area among all image semantic features of the thermal imaging image sample, and determining the cluster as the cluster corresponding to the image semantic feature of the corresponding estimated key feature area;

[0079] Performing feature fusion on all image semantic features contained in each cluster in the thermal imaging image sample to determine the cluster image features of each cluster in the thermal imaging image sample;

[0080] The AI ​​network model is iteratively optimized based on the cluster image features of each cluster of each thermal imaging image sample in the first thermal imaging image sample set as model loading data and the image semantic features of the estimated key feature area of ​​the corresponding cluster as model output data to output the key feature area positioning model.

[0081] In an alternative embodiment, the model training request based on the key feature area positioning model performs image segmentation on each thermal imaging image sample in the first thermal imaging image sample set to generate multiple thermal imaging image segments for each thermal imaging image sample, including:

[0082] Segmenting the thermal imaging image sample into blocks based on a preset feature matching template to generate a plurality of initial thermal imaging image blocks of the thermal imaging image sample;

[0083] Noise feature cleaning is performed on multiple initial thermal imaging image blocks of the thermal imaging image sample based on a preset noise feature template library to generate multiple thermal imaging image blocks of the thermal imaging image sample.

[0084] In an alternative embodiment, the dictionary used to block the thermal imaging image samples is the same as the dictionary used to block the thermal imaging data of the target object surface; the stop word list used to clean the noise features of the thermal imaging image samples is the same as the stop word list used to clean the noise features of the thermal imaging data of the target object surface.

[0085] In an alternative embodiment, the step of respectively parsing the foreground image partitions in the plurality of thermal imaging image blocks into image semantic features includes:

[0086] Parsing the foreground image partitions in the thermal imaging image blocks into image semantic features based on a previously trained image semantic feature encoding model;

[0087] The method further comprises:

[0088] Based on the training request of the image semantic feature encoding model, obtaining a second thermal imaging image sample set;

[0089] Performing initialization quality improvement on the second thermal imaging image sample set, wherein the initialization quality improvement includes image segmentation and noise feature cleaning;

[0090] Based on a convolutional neural network algorithm, feature coding learning is performed on the second thermal imaging image sample set after initialization quality improvement to generate the image semantic feature coding model.

[0091] In an alternative embodiment, the step of respectively parsing the foreground image partitions in the plurality of thermal imaging image blocks into image semantic features includes:

[0092] parsing a foreground image partition in a plurality of thermal imaging image blocks of the target object surface thermal imaging data into image semantic features having a set dimension;

[0093] The step of respectively parsing the foreground image partitions in the plurality of thermal imaging image blocks of the thermal imaging image sample into image semantic features comprises:

[0094] Parsing the foreground image partitions in the plurality of thermal imaging image blocks of the thermal imaging image sample into image semantic features having the set dimension.

[0095] That is, by ensuring that the image semantic features of the target object surface thermal imaging data and the image semantic features of the thermal imaging image sample have the same dimension, processing of image semantic features of different dimensions is avoided, thereby improving processing efficiency.

[0096] Figure 2An example thermal imaging analysis based target location system 100 that can be used to implement various embodiments described in the present application is shown schematically.

[0097] For one embodiment, Figure 2 A thermal imaging analysis based target location system 100 is shown having at least one processor 102, a control module (chipset) 104 coupled to at least one of the processor(s) 102, a memory 106 coupled to the control module 104, a non-volatile memory (NVM) / storage device 108 coupled to the control module 104, at least one input / output device 110 coupled to the control module 104, and a network interface 112 coupled to the control module 106.

[0098] The processor(s) 102 can include at least one single-core or multi-core processor, which can include any combination of general-purpose processors or special-purpose processors (e.g., graphics processors, application processors, baseband processors, etc.). In some embodiments, the thermal imaging analysis based target location system 100 can be capable of operating as a server device, such as a gateway, as described in the embodiments of the present application.

[0099] In some embodiments, the thermal imaging analysis based target location system 100 can include at least one computer-readable medium (e.g., the memory 106 or the NVM / storage device 108) having instructions 114 and at least one processor 102 integrated with the at least one computer-readable medium configured to execute the instructions 114 to implement modules to perform the actions described in the present disclosure.

[0100] For one embodiment, the control module 104 can include any suitable interface controllers to provide any suitable interface to at least one of the processor(s) 102 and / or any suitable device or component in communication with the control module 104.

[0101] The control module 104 can include a memory controller module to provide an interface to the memory 106. The memory controller module can be a hardware module, a software module, and / or a firmware module.

[0102] The memory 106 can be used, for example, to load and store data and / or instructions 114 for the thermal imaging analysis based target location system 100. For one embodiment, the memory 106 can include any suitable volatile memory, such as suitable DRAM. In some embodiments, the memory 106 can include double data rate type four synchronous dynamic random access memory (DDR4 FbRAM).

[0103] For one embodiment, the control module 104 may include at least one input / output controller to provide an interface to the NVM / storage device 108 and the (at least one) input / output device 110 .

[0104] For example, NVM / storage 108 may be used to store data and / or instructions 114. NVM / storage 108 may include any suitable non-volatile memory (e.g., flash memory) and / or may include any suitable (at least one) non-volatile storage device (e.g., at least one hard disk drive (HDD), at least one compact disk (CD) drive, and / or at least one digital versatile disk (DVD) drive).

[0105] The NVM / storage device 108 may include storage resources that are physically part of the device on which the thermal imaging analysis-based target location system 100 is installed, or it may be accessible to the device without being part of the device. For example, the NVM / storage device 108 may be accessible via (at least one) input / output device 110 over a network.

[0106] The (at least one) input / output device 110 may provide an interface for the thermal imaging analysis-based target location system 100 to communicate with any other appropriate device. The input / output device 110 may include a communication component, a phonetic component, a sensor component, etc. The network interface 112 may provide an interface for the thermal imaging analysis-based target location system 100 to communicate via at least one network. The thermal imaging analysis-based target location system 100 may wirelessly communicate with at least one component of a wireless network based on any of at least one wireless network standard and / or protocol, for example, accessing a wireless network based on a communication standard such as WiFi, 2G, 3G, 4G, 5G, etc., or a combination thereof for wireless communication.

[0107] For one embodiment, at least one of the (at least one) processors 102 may be packaged together with the logic of at least one controller of the control module 104 (e.g., a memory controller module). For one embodiment, at least one of the (at least one) processors 102 may be packaged together with the logic of at least one controller of the control module 104 to form a system-in-package (SiD). For one embodiment, at least one of the (at least one) processors 102 may be integrated on the same die with the logic of at least one controller of the control module 104. For one embodiment, at least one of the (at least one) processors 102 may be integrated on the same die with the logic of at least one controller of the control module 104 to form a system-on-chip (SoC).

[0108] In various embodiments, the thermal imaging analysis based target positioning system 100 can be, but is not limited to, a server, a desktop computing device, or a mobile computing device (e.g., a laptop computing device, a handheld computing device, a tablet, a netbook, etc.), and the like. In various embodiments, the thermal imaging analysis based target positioning system 100 can have more or less components, and / or different architectures. For example, in some embodiments, the thermal imaging analysis based target positioning system 100 includes at least one camera, a keyboard, a liquid crystal display (LCD) screen (including touch screen displays), a non-removable storage memory port, a plurality of antennas, a graphics chip, an application specific integrated circuit (ASIC), and a speaker.

[0109] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0110] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchain. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), data processing logic devices based on quantum computing, and the like.

[0111] In some embodiments, computer-executable instructions may be in the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0112] By way of example, computer executable instructions can comprise but are not limited to, instructions and data which, when executed by a processor, perform a certain function or implement an abstract data type, processing object, or other programming structure. The computer executable instructions may, for example, be implemented using software, hardware, firmware, or a combination of software, hardware, and firmware. Computer executable instructions can be stored in one or more computer readable media and executed by one or more computer processors.

[0113] The technical features of the above embodiments can be combined in any manner. For brevity, not all possible combinations of the technical features in the above embodiments are described, however, as long as the combinations of the technical features do not contradict each other, they should be considered within the scope of the present disclosure.

[0114] The above embodiments only express several implementation manners of the present application, and the description is specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that, for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application. Therefore, the scope of protection of the present application should be subject to the appended claims.

[0115] The present application has been described in detail above, and the principle and implementation manner of the present application have been described by applying specific examples; the above embodiment description is only used to help understand the method of the present application and its core idea; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation manner and application range can be changed, and the above description should not be understood as a limitation on the present application.

Claims

1. A target positioning method based on thermal imaging analysis, characterized in that: Applied to a server, the method includes: Performing image segmentation on the target object surface thermal imaging data to generate a plurality of thermal imaging image segments of the target object surface thermal imaging data; respectively parsing foreground image partitions in the plurality of thermal imaging image blocks into image semantic features; Clustering all image semantic features of the target object surface thermal imaging data into a target number of clusters based on a feature clustering strategy; Performing feature fusion on all image semantic features contained in each cluster in the surface thermal imaging data of the target object to determine the cluster image features of each cluster; Inputting the cluster image features of each cluster into a key feature area positioning model that meets the model convergence condition to generate image semantic features of the estimated key feature areas of each cluster; The key feature area positioning model is generated by iteratively optimizing the AI ​​network model based on the cluster image features of each cluster of each thermal imaging image sample in the first thermal imaging image sample set as model loading data, and the image semantic features of the estimated key feature areas of the corresponding cluster as model output data; Calculating the matching value between each image semantic feature of each cluster and the image semantic feature of the estimated key feature area of ​​each cluster respectively, and determining the foreground image partition corresponding to the image semantic feature with the largest matching value among all the image semantic features of each cluster as the key feature area of ​​each cluster; The key feature area of ​​the target object surface thermal imaging data is extracted based on the key feature area of ​​each cluster of the target object surface thermal imaging data.

2. The target positioning method based on thermal imaging analysis according to claim 1, characterized in that: The step of respectively calculating a matching value between each image semantic feature of each cluster and an image semantic feature of an estimated key feature area of ​​each cluster, and determining a foreground image partition corresponding to an image semantic feature having a maximum matching value among all image semantic features of each cluster as the key feature area of ​​each cluster, comprises: Calculating the feature costs of each image semantic feature of each cluster and the image semantic features of the estimated key feature areas of each cluster respectively; The foreground image partition corresponding to the image semantic feature with the smallest feature cost between all the image semantic features of the clusters and the image semantic features of the estimated key feature area of ​​the clusters is determined as the key feature area of ​​the clusters.

3. The target positioning method based on thermal imaging analysis according to claim 1, characterized in that: The step of performing image segmentation on the target object surface thermal imaging data to generate a plurality of thermal imaging image segments of the target object surface thermal imaging data includes: Segmenting the surface thermal imaging data of the target object based on a preset feature matching template to generate a plurality of initial thermal imaging image blocks; Noise feature cleaning is performed on the multiple initial thermal imaging image blocks based on a preset noise feature template library to generate multiple thermal imaging image blocks of thermal imaging data of the target object surface.

4. The target positioning method based on thermal imaging analysis according to claim 1, characterized in that: The method further comprises: Based on the model training request of the key feature area positioning model, performing image segmentation on each thermal imaging image sample in the first thermal imaging image sample set to generate multiple thermal imaging image segments for each thermal imaging image sample; respectively parsing foreground image partitions in a plurality of thermal imaging image blocks of the thermal imaging image sample into image semantic features; Clustering all image semantic features of the thermal imaging image sample into a preset number of clusters based on a feature clustering strategy; Wherein, the preset number is the number of estimated key feature areas in the thermal imaging image sample; For each estimated key feature area of ​​the thermal imaging image sample, determining a cluster corresponding to an image semantic feature that best matches the image semantic feature of the corresponding estimated key feature area among all image semantic features of the thermal imaging image sample, and determining the cluster as the cluster corresponding to the image semantic feature of the corresponding estimated key feature area; Performing feature fusion on all image semantic features contained in each cluster in the thermal imaging image sample to determine the cluster image features of each cluster in the thermal imaging image sample; The AI ​​network model is iteratively optimized based on the cluster image features of each cluster of each thermal imaging image sample in the first thermal imaging image sample set as model loading data and the image semantic features of the estimated key feature area of ​​the corresponding cluster as model output data to output the key feature area positioning model.

5. The target positioning method based on thermal imaging analysis according to any one of claims 1 to 4, characterized in that: The step of respectively parsing the foreground image partitions in the plurality of thermal imaging image blocks into image semantic features comprises: Parsing the foreground image partitions in the thermal imaging image blocks into image semantic features based on a previously trained image semantic feature encoding model; The method further comprises: Based on the training request of the image semantic feature encoding model, obtaining a second thermal imaging image sample set; Performing initialization quality improvement on the second thermal imaging image sample set, wherein the initialization quality improvement includes image segmentation and noise feature cleaning; Based on a convolutional neural network algorithm, feature coding learning is performed on the second thermal imaging image sample set after initialization quality improvement to generate the image semantic feature coding model.

6. The target positioning method based on thermal imaging analysis according to claim 4, characterized in that: The step of respectively parsing the foreground image partitions in the plurality of thermal imaging image blocks into image semantic features comprises: parsing a foreground image partition in a plurality of thermal imaging image blocks of the target object surface thermal imaging data into image semantic features having a set dimension; The step of respectively parsing the foreground image partitions in the plurality of thermal imaging image blocks of the thermal imaging image sample into image semantic features comprises: Parsing the foreground image partitions in the plurality of thermal imaging image blocks of the thermal imaging image sample into image semantic features having the set dimension.

7. The target positioning method based on thermal imaging analysis according to claim 6, characterized in that: The method further comprises: Performing image blocking on each thermal imaging image sample in the first thermal imaging image sample set to generate a plurality of thermal imaging image blocks for each thermal imaging image sample; respectively parsing foreground image partitions in a plurality of thermal imaging image blocks of the thermal imaging image sample into image semantic features; Clustering all image semantic features of the thermal imaging image sample into a preset number of clusters based on a feature clustering strategy; Wherein, the preset number is the number of estimated key feature areas in the thermal imaging image sample; For each estimated key feature area of ​​the thermal imaging image sample, determining a cluster corresponding to an image semantic feature that best matches the image semantic feature of the corresponding estimated key feature area among all image semantic features of the thermal imaging image sample, and determining the cluster as the cluster corresponding to the image semantic feature of the corresponding estimated key feature area; Performing feature fusion on all image semantic features contained in each cluster in the thermal imaging image sample to determine the cluster image features of each cluster; The AI ​​network model is iteratively optimized based on the cluster image features of each cluster of each thermal imaging image sample in the first thermal imaging image sample set as model loading data and the image semantic features of the estimated key feature area of ​​the corresponding cluster as model output data to output a key feature area positioning model.

8. The target positioning method based on thermal imaging analysis according to claim 6, characterized in that: The step of respectively parsing the foreground image partitions in the plurality of thermal imaging image blocks of the thermal imaging image sample into image semantic features comprises: Parsing the foreground image partitions in the thermal imaging image blocks into image semantic features based on a previously trained image semantic feature encoding model; The method further comprises: Based on the training request of the image semantic feature encoding model, obtaining a second thermal imaging image sample set; Performing initialization quality improvement on the second thermal imaging image sample set, wherein the initialization quality improvement includes image segmentation and noise feature cleaning; Based on a convolutional neural network algorithm, feature coding learning is performed on the second thermal imaging image sample set after initialization quality improvement to generate the image semantic feature coding model.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program, which, when executed by a processor, implements the target positioning method based on thermal imaging analysis according to any one of claims 1 to 7.

10. A computer program product, characterized in that The method comprises a computer program or a computer executable instruction, which, when executed by a processor, implements the target positioning method based on thermal imaging analysis according to any one of claims 1 to 7.

11. A target positioning system based on thermal imaging analysis, characterized in that: The target positioning system based on thermal imaging analysis includes a processor and a readable storage medium, wherein the readable storage medium stores a program, and when the program is executed by the processor, the target positioning method based on thermal imaging analysis according to any one of claims 1 to 7 is implemented.

Images