CN118446902B - Terahertz human body security inspection image optimization method and system - Google Patents

Abstract

The invention relates to the technical field of security inspection image optimization, in particular to a terahertz human body security inspection image optimization method and system, which improve the quality and the security inspection efficiency of security inspection images and ensure the accuracy and the high efficiency of security inspection; the method comprises the following steps: collecting real-time running state data information of security inspection equipment and a human body security inspection image to be optimized; acquiring standard running state data information of security inspection equipment; inputting the standard operation state data information and the real-time operation state data information of the security inspection equipment into a pre-trained equipment degradation operation parameter identification model to obtain a security inspection equipment degradation operation parameter set; constructing a correlation feature matrix between quality features of the human body security inspection image and operation parameters of the security inspection equipment based on the historical operation state data of the security inspection equipment and the human body security inspection image corresponding to the historical operation state data of the security inspection equipment; the correlation characteristic matrix comprises correlation between each security inspection equipment operation parameter and each human security inspection image quality characteristic.

Description

Terahertz human body security inspection image optimization method and system

Technical Field

The invention relates to the technical field of security inspection image optimization, in particular to a terahertz human body security inspection image optimization method and system.

Background

Along with the progress of science and technology and the improvement of safety requirements, the terahertz technology is widely applied in the field of human body security inspection because of the unique characteristics of non-ionization, strong penetrability, effective revealing of hidden objects and the like. However, the quality of terahertz human body security inspection images directly influences the security inspection efficiency and accuracy, how to effectively optimize such images, ensuring the accuracy and high efficiency of security inspection becomes an important subject of current research and practice.

Terahertz human body security inspection systems generally comprise complex hardware equipment and advanced imaging algorithms, and small changes of operation parameters of the terahertz human body security inspection systems can have significant influence on the quality of generated security inspection images. These parameters cover the light source intensity, detector sensitivity, scanning speed, signal processing strategies, imaging algorithms, and the like. In the actual operation process, due to factors such as equipment aging, environmental change, manual operation and the like, the running state of the security inspection equipment may deviate from ideal standards, so that problems such as unbalanced brightness, increased noise, reduced resolution, insufficient contrast and the like of images occur, and the identification and judgment of security inspection personnel on potential threat objects are seriously affected. The existing human body security inspection image optimization means often depend on an image post-processing algorithm in a fixed mode, and the method is not only lack of pertinence, but also is difficult to cope with complex and changeable equipment running states; therefore, development of an intelligent and self-adaptive terahertz human body security inspection image optimization method is urgently needed.

Disclosure of Invention

In order to solve the technical problems, the invention provides a terahertz human body security inspection image optimization method and system for improving the quality and the security inspection efficiency of security inspection images and ensuring the accuracy and the high efficiency of security inspection.

In a first aspect, the invention provides a terahertz human body security inspection image optimization method, which comprises the following steps:

Collecting real-time running state data information of security inspection equipment and a human body security inspection image to be optimized;

Acquiring standard running state data information of security inspection equipment;

Inputting the standard operation state data information and the real-time operation state data information of the security inspection equipment into a pre-trained equipment degradation operation parameter identification model to obtain a security inspection equipment degradation operation parameter set;

Constructing a correlation feature matrix between quality features of the human body security inspection image and operation parameters of the security inspection equipment based on the historical operation state data of the security inspection equipment and the human body security inspection image corresponding to the historical operation state data of the security inspection equipment; the correlation characteristic matrix comprises correlation between each security inspection equipment operation parameter and each human security inspection image quality characteristic;

Taking the degradation operation parameter set of the security inspection equipment as an index, and extracting a human body security inspection image quality feature set which corresponds to the degradation operation parameter set of the security inspection equipment and has a correlation larger than a preset correlation threshold value from a correlation feature matrix;

Inputting the quality feature set of the human body security inspection image into a pre-built security inspection image equipment interference optimization particle cluster to obtain first security inspection image optimization particles corresponding to the security inspection equipment degradation operation parameter set;

And carrying out optimization processing on the human body security inspection image to be optimized according to the first security inspection image optimization particles to obtain a human body security inspection optimization image.

Further, the construction method of the equipment degradation operation parameter identification model comprises the following steps:

Collecting operation state data of security inspection equipment, wherein the operation state data comprise light source intensity, detector sensitivity, scanning speed, signal processing parameters and imaging algorithm parameters under different working conditions;

Labeling corresponding equipment degradation levels for each group of security inspection equipment state data to form label data required by supervised learning;

Selecting the characteristic with the most distinguishing power for the degradation state of the equipment from the original equipment state data, and carrying out standardization and normalization processing on the selected characteristic;

Selecting a deep learning model as an infrastructure of a device degradation operation parameter identification model, wherein the deep learning model comprises a logistic regression, a support vector machine, a decision tree, a random forest, a neural network, a convolutional neural network, a cyclic neural network and a long-term and short-term memory network;

dividing the processed characteristics into a training set, a verification set and a test set;

Updating parameters through a back propagation algorithm by using the training set and the corresponding degradation labels, optimizing the identification capacity of the equipment degradation operation parameter identification model on the equipment degradation state, and periodically evaluating the model performance on the verification set;

evaluating the identification accuracy, recall rate and F1 score index of the equipment degradation operation parameter identification model on an independent test set;

And packaging the trained equipment degradation operation parameter identification model into an interface which is easy to call, and integrating the interface into a system.

Further, the construction method of the correlation characteristic matrix comprises the following steps:

collecting state data of human body security inspection equipment in a historical operation state and corresponding generated human body security inspection images;

extracting quality characteristics of each historical security inspection image;

Measuring the association strength between the operation parameters of each security inspection device and the quality characteristics of each human security inspection image;

calculating a correlation coefficient between each set of device operating parameters and image quality features;

and organizing a correlation degree feature matrix according to the corresponding relation between the equipment operation parameters and the image quality features by the calculated correlation coefficients.

Further, the method for acquiring the human body security inspection optimized image comprises the following steps:

optimizing particles according to the first security inspection image, and determining an image processing strategy;

According to the first security inspection image optimization particles, parameter adjustment and optimization are carried out on the selected image processing strategy;

monitoring the quality change of the processed image in real time, and adjusting the image processing strategy according to the feedback information;

performing quality evaluation and verification on the optimized image after the processing is completed;

and outputting and recording the finally optimized human body security inspection image.

Further, the real-time operational status data information includes light source intensity, detector sensitivity, scanning speed, signal processing strategy, and imaging algorithm parameters.

Further, the acquisition way of the standard operation state data information of the security inspection equipment comprises equipment manual and technical documents, equipment calibration and authentication and experience accumulation.

Further, the setting influence factors of the preset relevant threshold value comprise safety requirements, image quality requirements, equipment running state change range, historical data analysis and performance balance.

On the other hand, the application also provides a terahertz human body security inspection image optimization system, which comprises:

The real-time data acquisition module is used for acquiring real-time running state data information of the security inspection equipment and a human body security inspection image to be optimized;

the equipment degradation operation parameter identification module is used for acquiring the standard operation state data information of the security inspection equipment, inputting the standard operation state data information of the security inspection equipment and the real-time operation state data information of the security inspection equipment into a pre-trained equipment degradation operation parameter identification model, and acquiring a security inspection equipment degradation operation parameter set;

The correlation characteristic matrix construction module is used for constructing a correlation characteristic matrix between the quality characteristics of the human body security inspection image and the operation parameters of the security inspection equipment according to the historical operation state data of the security inspection equipment and the human body security inspection image corresponding to the historical operation state data of the security inspection equipment; the correlation characteristic matrix comprises correlation between each security inspection equipment operation parameter and each human security inspection image quality characteristic;

The image screening module is used for taking the degradation operation parameter set of the security inspection equipment as an index, and extracting a human body security inspection image quality characteristic set which corresponds to the degradation operation parameter set of the security inspection equipment and has a correlation larger than a preset correlation threshold value from the correlation characteristic matrix;

the optimized particle generation module is used for inputting the quality characteristic set of the human body security inspection image into a pre-built security inspection image equipment interference optimized particle cluster to obtain first security inspection image optimized particles corresponding to the security inspection equipment degradation operation parameter set;

the image optimization processing module is used for carrying out optimization processing on the human body security inspection image to be optimized according to the first security inspection image optimization particles to obtain a human body security inspection optimization image.

In a third aspect, the present application provides an electronic device comprising a bus, a transceiver, a memory, a processor and a computer program stored on the memory and executable on the processor, the transceiver, the memory and the processor being connected by the bus, the computer program when executed by the processor implementing the steps of any of the methods described above.

In a fourth aspect, the application also provides a computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of any of the methods described above.

Compared with the prior art, the invention has the beneficial effects that: by acquiring real-time operation state data information and standard operation state data information of the security inspection equipment and applying a pre-trained equipment degradation operation parameter identification model, the method can realize real-time monitoring and identification of operation state change of the security inspection equipment, and adjust an image optimization strategy according to real-time conditions, so that the self-adaptability of the system is maintained;

By constructing a correlation feature matrix based on the historical operation state data of the security inspection equipment, the relation between the operation parameters of the security inspection equipment and the quality features of the security inspection images of the human body can be comprehensively considered, so that the comprehensiveness and the accuracy of the optimization of the image quality are ensured;

The optimized particle clusters are interfered by the pre-built security inspection image equipment, so that optimized particles can be intelligently generated and the image can be optimized, and the image quality and the security inspection effect are improved;

The image to be optimized can be efficiently screened out by extracting the image feature set which is related to the equipment degradation operation parameter set and has higher correlation degree, and meanwhile, the reliability and stability of the method are ensured by adopting a pre-trained model and a pre-trained cluster;

according to the method, the optimization strategy can be flexibly adjusted according to the running state and the image characteristics of the specific security inspection equipment so as to meet the requirements in different scenes, and the method has strong pertinence and flexibility;

In conclusion, the terahertz human body security inspection image optimization method can effectively cope with complex and changeable environment and equipment state changes in the security inspection field, improves the quality and the security inspection efficiency of security inspection images, and ensures the accuracy and the high efficiency of security inspection.

Drawings

FIG. 1 is a flow chart of the present invention;

FIG. 2 is a flow chart of a method of constructing a device degradation operating parameter identification model;

Fig. 3 is a block diagram of a terahertz human body security inspection image optimization system.

Detailed Description

In the description of the present application, those skilled in the art will appreciate that the present application may be embodied as methods, apparatus, electronic devices, and computer-readable storage media. Accordingly, the present application may be embodied in the following forms: complete hardware, complete software (including firmware, resident software, micro-code, etc.), a combination of hardware and software. Furthermore, in some embodiments, the application may also be embodied in the form of a computer program product in one or more computer-readable storage media, which contain computer program code.

Any combination of one or more computer-readable storage media may be employed by the computer-readable storage media described above. The computer-readable storage medium includes: an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of the computer readable storage medium include the following: portable computer magnetic disks, hard disks, random access memories, read-only memories, erasable programmable read-only memories, flash memories, optical fibers, optical disk read-only memories, optical storage devices, magnetic storage devices, or any combination thereof. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, device.

The technical scheme of the application obtains, stores, uses, processes and the like the data, which all meet the relevant regulations of national laws.

The application provides a method, a device and electronic equipment through flow charts and/or block diagrams.

It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer-readable program instructions. These computer-readable program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks.

These computer readable program instructions may also be stored in a computer readable storage medium that can cause a computer or other programmable data processing apparatus to function in a particular manner. Thus, instructions stored in a computer-readable storage medium produce an instruction means which implement the functions/acts specified in the flowchart and/or block diagram block or blocks.

The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks.

The present application will be described below with reference to the drawings in the present application.

Embodiment one: as shown in fig. 1 to 2, the terahertz human body security inspection image optimization method provided by the invention specifically comprises the following steps:

S1, acquiring real-time running state data information of security inspection equipment and a human body security inspection image to be optimized;

S1, acquiring comprehensive and accurate real-time running state data of security inspection equipment and a human body security inspection image to be optimized, and providing necessary data support for identifying equipment degradation parameters, constructing a correlation characteristic matrix, generating optimized particles and performing image optimization processing in the subsequent steps;

The real-time running state data information covers various key parameters in the running process of the terahertz security inspection system, and specifically comprises the following steps:

The light source intensity records the output power of the currently used terahertz light source, ensures that the light source is in a stable and proper range, and generates high-quality terahertz waves; too strong or too weak light sources can cause overexposure or underexposure of the image, affecting the presentation of image details;

The sensitivity of the detector is used for monitoring the responsivity and the signal-to-noise ratio of the terahertz detector under the current working condition; the high-sensitivity detector can more effectively capture weak terahertz echo signals, and reduce the influence of noise on image quality;

scanning speed, recording the speed of the equipment when performing terahertz scanning on a human body; too fast or too slow scanning speed can cause motion blurring of the image or too long acquisition time, and the definition and the overall imaging efficiency of the image are affected;

The signal processing strategy is to collect preprocessing parameters such as signal filtering, gain control, background suppression and the like adopted by the equipment; the reasonable signal processing strategy can effectively reduce environmental interference, improve the signal-to-noise ratio of images and enhance the contrast ratio of hidden objects and the background;

Imaging algorithm parameters, recording a reconstruction algorithm adopted by the system and related parameters thereof, wherein the parameters directly influence the resolution, contrast and noise level of the image;

The human body security inspection images to be optimized are directly generated by a terahertz security inspection system in the process of executing security inspection tasks, and the images are presented in two-dimensional and three-dimensional forms and contain terahertz reflection or transmission information on the body surface of the inspected person and below clothes; the image acquisition follows the following principle:

The complete coverage ensures that the image covers the whole body of the detected personnel without obvious missing areas so as to comprehensively check potential threat objects;

Proper resolution, image resolution should meet the need for identifying small and hidden items while considering data processing and transmission efficiency;

the standardized format is that the image data should be stored according to the unified standard format, so that the subsequent image analysis and optimization processing are facilitated;

and the timestamp marks are used for adding accurate timestamps for each image and are associated with corresponding equipment running state data, so that equipment states during image generation can be tracked conveniently.

In the step, through collecting various key parameter data of the security inspection equipment in real time, the running state of the equipment can be comprehensively known, so that the abnormal running or faults of the equipment can be found in time; the human body security inspection image to be optimized is obtained in real time, the body structure and the article carrying condition of the inspected person can be intuitively displayed, basic data are provided for subsequent image optimization processing, and the definition, contrast and resolution of the image are improved; by analyzing the real-time running state data information, degradation parameters possibly existing in the equipment can be identified, and target parameters are provided for subsequent optimization processing; the real-time running state data is associated with the human body security inspection image, so that a correlation characteristic matrix is built, the degree of association between the running parameters of the equipment and the image quality is analyzed, and guidance is provided for subsequent image optimization; by adding an accurate time stamp to each image and associating the accurate time stamp with corresponding equipment running state data, the equipment state during image generation can be tracked and managed, and subsequent data analysis and optimization processing are facilitated;

In summary, the beneficial effects of the step S1 include comprehensively knowing the device state, optimizing the image quality, accurately identifying the degradation parameters, establishing the correlation feature matrix, and realizing data tracking and management, thereby providing necessary data support and guidance for the subsequent optimization of the terahertz human body security inspection image.

S2, acquiring standard running state data information of security inspection equipment;

The standard operation state data information refers to the recommended range of each key operation parameter of the terahertz security inspection equipment under ideal conditions; the standard running state data information of the security inspection equipment is consistent with the real-time running state data information, and the security inspection equipment comprises the following components:

The light source intensity standard ensures that the image quality is ensured and the discomfort caused by excessive irradiation to the human body is avoided according to the technical specification provided by the equipment manufacturer and the optimum light source output power determined by laboratory tests;

The sensitivity standard of the detector, the optimal responsivity and signal-to-noise ratio obtained by measurement when the equipment leaves the factory are used for evaluating whether the performance of the current detector accords with the expectation or not and whether calibration or maintenance is needed or not;

the scanning speed standard is a proper scanning speed set according to the equipment design principle and the human body moving speed factor, so that the integrity of image acquisition can be ensured, and image blurring or overlong acquisition time caused by too high or too low speed can be avoided;

A signal processing standard strategy, and a preprocessing parameter combination such as optimal signal filtering, gain control, background suppression and the like recommended by equipment manufacturers or verified by a large number of experiments, is used for reducing noise to the maximum extent, improving signal-to-noise ratio and keeping image details clear;

the imaging algorithm standard parameters, the reconstruction algorithm parameters optimized for specific equipment models and application scenes, are aimed at ensuring that images with high resolution, good contrast and low noise can be generated under different conditions;

The method for acquiring the standard running state data information of the security inspection equipment comprises the following steps:

Equipment manuals and technical documentation, standard operating parameters are obtained from the official materials of user manuals, technical white books and performance reports provided by equipment manufacturers;

the method comprises the steps of equipment calibration and authentication, and in the equipment installation and debugging stage, performing on-site calibration and performance test through a professional institution to obtain standard running state data actually reached by equipment in the current environment;

Experience accumulation, combined with the actual performance of long-term operation of the equipment, user feedback and latest research results, continuously revises and perfects standard operation state data, ensures that the standard operation state data is advanced with time and reflects best practice.

In the step, the standard operation state data information of the security inspection equipment is obtained, so that a reference range can be provided for actual operation; these data not only serve as benchmarks for equipment performance, but also help operators to understand the expected behavior of the equipment in an ideal situation, thereby better performing equipment operation and maintenance; the standard running state data information provides guidance for parameter setting of the equipment; by knowing the optimal parameter configuration recommended by the device manufacturer, operators can more accurately adjust the device to meet the requirements of security inspection while ensuring image quality and security; the standardized setting of parameters such as light source intensity, detector sensitivity and the like is beneficial to ensuring that the security inspection equipment cannot cause discomfort or injury to inspected personnel in the operation process; by following standardized running state data information, the safety of the security inspection process can be ensured to the greatest extent; the standardized signal processing strategy and imaging algorithm parameters can improve the quality of the image; by using the verified optimal parameter combination, noise in the image can be reduced to the maximum extent, and definition and detail of the image are maintained, so that accuracy and efficiency of security inspection are improved; experience accumulation and continuous optimization enable standard operating state data information to be in progress; along with the continuous development of technology and the accumulation of practical experience, standard parameters can be revised and updated at any time to reflect best practice, and the performance and efficiency of security inspection equipment are continuously improved;

In summary, by acquiring the standard running state data information of the security inspection equipment, the operation efficiency and the security of the equipment can be improved, the image quality is optimized, and more reliable support is provided for security inspection.

S3, inputting the standard operation state data information of the security inspection equipment and the real-time operation state data information of the security inspection equipment into a pre-trained equipment degradation operation parameter identification model to obtain a security inspection equipment degradation operation parameter set;

Step S3, utilizing a pre-trained equipment degradation operation parameter identification model, accurately identifying and quantifying performance degradation parameters of equipment under the current working condition by comparing the real-time operation state of the security inspection equipment with standard operation state data, and providing a key basis for the subsequent targeted optimization of security inspection images;

the construction method of the equipment degradation operation parameter identification model comprises the following steps:

S31, collecting a large amount of security inspection equipment operation state data, including light source intensity, detector sensitivity, scanning speed, signal processing parameters and imaging algorithm parameters under different working conditions;

s32, labeling corresponding equipment degradation grades for each group of security inspection equipment state data to form label data required by supervised learning;

S33, selecting the characteristic with the most distinguishing power for the degradation state of the equipment from the original equipment state data, and carrying out standardization and normalization processing on the selected characteristic to ensure that different characteristics are on the same scale;

S34, selecting a deep learning model as an infrastructure of a device degradation operation parameter identification model, wherein the deep learning model comprises a logistic regression, a support vector machine, a decision tree, a random forest, a neural network, a convolutional neural network, a cyclic neural network and a long-term and short-term memory network;

s35, dividing the processed characteristics into a training set, a verification set and a test set;

S36, updating parameters through a back propagation algorithm by using the training set and the corresponding degradation labels, optimizing the identification capacity of the equipment degradation operation parameter identification model on the equipment degradation state, and periodically evaluating the model performance on the verification set to prevent over-fitting;

S37, S: evaluating the identification accuracy, recall rate and F1 score index of the equipment degradation operation parameter identification model on an independent test set, and comprehensively evaluating the performance of the model;

s38, packaging the trained equipment degradation operation parameter identification model into an interface which is easy to call, so that real-time equipment state data can be received in actual application conveniently and degradation parameter identification results can be output quickly.

In the step, by comparing the real-time running state of the security inspection equipment with the standard running state data, the model can accurately identify and quantify performance degradation parameters of the equipment under the current working condition, thereby being beneficial to timely finding out the running problem of the equipment and reducing the degradation of the security inspection image quality and the security inspection efficiency; by acquiring the degradation operation parameter set of the security inspection equipment, a key basis can be provided for the subsequent targeted optimization of the security inspection image; knowing the degradation state of the equipment can help to adjust parameters such as an imaging algorithm, a signal processing strategy and the like, so that the quality of security inspection images is improved, and the security inspection accuracy and efficiency are improved; the deep learning model is adopted as the basic framework, so that the degradation state of the equipment can be effectively identified, and the accuracy and generalization capability are higher; the accuracy and stability of recognition are improved, so that the subsequent image optimization work is guided more reliably; the trained model is packaged into an interface which is easy to call, so that real-time equipment state data can be conveniently received, and degradation parameter identification results can be rapidly output; the model can be easily applied to an actual security inspection system, and timely support and guidance are provided for security inspection work;

In summary, the equipment degradation operation parameter identification model constructed in the step S3 can accurately and efficiently identify the degradation state of the equipment, provides important basis and support for subsequent security inspection image optimization, and is beneficial to improving security inspection efficiency and accuracy.

S4, constructing a correlation feature matrix between quality features of the human body security inspection image and operation parameters of the security inspection equipment based on the historical operation state data of the security inspection equipment and the human body security inspection image corresponding to the historical operation state data of the security inspection equipment; the correlation characteristic matrix comprises correlation between each security inspection equipment operation parameter and each human security inspection image quality characteristic;

the constructed correlation characteristic matrix clearly shows the correlation degree between the operation parameters of each security inspection device and the quality characteristics of each human security inspection image; the correlation feature matrix is the basis for determining the quality features of the key images, generating optimized particles and performing image optimization in the subsequent steps; the correlation characteristic matrix reflects the internal relation between the running state of the equipment and the quality of the security inspection image, and is helpful for rapidly positioning the image quality characteristic highly correlated with the equipment when the equipment is subjected to the degradation running parameter set, so that the targeted image optimization is realized;

the construction method of the correlation characteristic matrix comprises the following steps:

S41, collecting state data of human body security inspection equipment in a historical operation state and corresponding generated human body security inspection images; ensuring that the data overlay device operates normally, slightly deteriorates to severely deteriorates in various states to reflect the influence of device parameter changes on image quality;

S42, extracting quality features of each historical security inspection image, wherein the quality features comprise brightness distribution, contrast ratio, signal to noise ratio, resolution, edge definition, texture features, gray level histogram and frequency domain features;

S43, selecting a correlation measurement method, and measuring the correlation strength between the operation parameters of each security inspection device and the quality characteristics of each human security inspection image; the correlation measurement method comprises a Pierson correlation coefficient, a Szellman grade correlation coefficient, a Kendell tau coefficient and a distance correlation;

s44, calculating a correlation coefficient between each group of equipment operation parameters and image quality characteristics by using the selected correlation measurement method; for each device parameter, obtaining a vector associated with all image quality features; conversely, for each image quality feature, a vector is also obtained that is related to all device parameters;

S45, organizing the calculated correlation coefficient into a two-dimensional matrix according to the corresponding relation between the equipment operation parameter and the image quality characteristic, namely a correlation characteristic matrix.

In the step, a correlation degree characteristic matrix is constructed, so that the correlation degree between the operation parameters of each security inspection device and the quality characteristics of each human security inspection image is systematically quantized, a structured knowledge base is formed, and a complex relation network between the operation state of the device and the quality of the security inspection image is intuitively displayed; the correlation feature matrix provides basis for determining key image quality features highly correlated with the equipment degradation operation parameters in the subsequent steps; when the degradation condition of specific equipment is faced, the image characteristics with the largest influence can be quickly positioned through the query matrix, so that the optimization work is ensured to be focused on the most critical problem point, and the pertinence and the efficiency of the optimization are improved; based on the correlation characteristic matrix, an optimization strategy can be dynamically adjusted according to the running state of the real-time equipment, and optimized particles which are adaptive to the current equipment state are generated, so that the intelligent and self-adaptive optimization of the security inspection image is realized; the adaptability of the system is enhanced, so that the system can effectively cope with equipment state fluctuation caused by factors such as equipment aging, environmental change, manual operation and the like, and the stability and the accuracy of the quality of the security inspection image are ensured; through deep mining and analysis of historical data, the constructed correlation feature matrix not only serves an immediate image optimization task, but also provides a basis for long-term knowledge accumulation and technology iteration; along with the continuous accumulation and updating of data, the matrix can be continuously perfected, so that the understanding and modeling capacity of the image relationship of the equipment is further improved, and the development of security inspection image optimization technology is promoted; the correlation feature matrix provides data-driven decision support for security inspection image optimization, reduces the dependence on subjective experience, and improves the objectivity and scientificity of decisions; has important significance for improving the overall performance of the security inspection system and ensuring the accuracy and the high efficiency of security inspection work;

In summary, the construction of the correlation feature matrix is a key link in the terahertz human body security inspection image optimization method, so that the deep correlation analysis between the running state of equipment and the security inspection image quality is realized, powerful data support and decision basis are provided for subsequent targeted optimization, and the accurate control and efficient improvement of the security inspection image quality are promoted.

S5, taking the degradation operation parameter set of the security inspection equipment as an index, and extracting a human body security inspection image quality feature set which corresponds to the degradation operation parameter set of the security inspection equipment and has a correlation larger than a preset correlation threshold value from the correlation feature matrix;

s5, screening out an image quality feature set which is highly relevant to the correlation feature matrix and has obvious influence on the image quality according to the identified degradation operation parameter set of the security inspection equipment; the pertinence of the subsequent image optimization work is ensured, so that the optimization measures can directly act on the image characteristics most obviously affected by equipment degradation, thereby more effectively improving the image quality and improving the security inspection accuracy;

screening out image quality features with obvious relevance to the equipment degradation operation parameter set based on a preset relevance threshold;

For each parameter in the degradation operation parameter set, searching a corresponding row or column in the correlation feature matrix, and acquiring correlation coefficients of the parameter and all image quality features;

screening out image quality features with correlation coefficients larger than a preset correlation threshold, wherein the features have strong correlation with degradation parameters of current equipment, namely the change of equipment parameters has great influence on the degradation parameters;

Combining all the screened image quality features with the correlation degree higher than a threshold value to form a human body security inspection image quality feature set which is closely related to the current equipment degradation operation parameter set; the set represents the image attribute that most needs to be focused and optimized in the current equipment state;

the setting influence factors of the preset relevant threshold value comprise:

The safety requirement is considered, and the preset related threshold value is set to ensure that the identification and judgment of the potential threat articles have enough accuracy and reliability; too high setting of the relevant threshold value can lead to missed detection or misjudgment, while too low setting can introduce too many interference factors to influence the security inspection efficiency and accuracy;

The image quality requirement, the setting of the preset relevant threshold value should also consider the requirement on the image quality; too low a correlation threshold setting can result in too many images being excluded from the optimization range, affecting the effectiveness of image optimization; if the setting is too high, excessive noise or interference images are contained, and the quality of the optimized images is affected;

The equipment operation state change range is considered, the operation state of the security inspection equipment is influenced by various factors, and the change range of the equipment operation state is required to be considered for setting a preset related threshold; the related threshold value is set outside the range of the running state change of the equipment, so that the running state change of the equipment cannot be effectively identified and adapted, and the image optimization effect is affected;

Historical data analysis, namely analyzing historical operation state data of security inspection equipment and corresponding human security inspection images to obtain actual correlation conditions between different operation parameters and image quality; the setting of the preset relevant threshold value can refer to the analysis result of the historical data so as to ensure that the historical data has better applicability and effectiveness in practical application;

the performance is balanced, and the setting of the preset relevant threshold value needs to be balanced and balanced among the factors such as safety, image quality requirements, equipment running state change range and the like so as to realize the optimal effect of optimizing the security inspection image.

In the step, the pertinence of image optimization work is ensured by extracting an image quality feature set highly related to a degradation operation parameter set of security inspection equipment from a correlation feature matrix; the method can directly act on the image characteristics most obviously affected by equipment degradation, so that the image quality is improved more effectively and the security inspection accuracy is improved; by optimizing the image characteristics most significantly affected by the equipment degradation, the image quality degradation caused by the equipment running state change can be effectively reduced; the definition, contrast and resolution of the security inspection image are improved, and the accuracy of identification and judgment of potential threat objects by security inspection personnel is improved; by combining the image quality characteristics highly correlated with the equipment degradation operation parameter set, a closely correlated 'human body security inspection image quality characteristic set' is formed, and the image characteristics can be subjected to targeted optimization treatment, so that the image quality is remarkably improved; when the preset relevant threshold value is set, factors such as safety requirements, image quality requirements, equipment running state change range, historical data analysis, performance balance and the like are comprehensively considered, so that the preset relevant threshold value is ensured to have better applicability and effectiveness in practical application; the method is beneficial to balancing the relation between security inspection accuracy and efficiency and balancing the trade-off between the effect of image quality optimization and cost;

In conclusion, the beneficial effects of the step S5 are embodied in the aspects of improving security inspection accuracy, optimizing image quality, improving targeted optimizing effect, comprehensively considering various factors and the like, and are beneficial to improving the effect and practicality of terahertz human body security inspection image optimization.

S6, inputting the quality feature set of the human body security inspection image into a pre-built security inspection image equipment interference optimization particle cluster to obtain first security inspection image optimization particles corresponding to the security inspection equipment degradation operation parameter set;

S6, utilizing a pre-built security inspection image device to interfere with an optimized particle cluster, and generating an optimized strategy with strong pertinence and adapting to the state change of the device according to the key image quality characteristics of the device in the degradation state; the method overcomes the limitation of the traditional fixed mode image post-processing algorithm, realizes intelligent response to complex and changeable equipment states, and meets the requirements of the security inspection image optimization field on accurate and efficient optimization technology;

taking the image quality feature set as input, and sending the image quality feature set into a security inspection image equipment interference optimization particle cluster;

According to the characteristics, the security inspection image equipment interference optimization particle cluster starts an internal optimization algorithm, and an optimization strategy capable of improving the image quality to the greatest extent is searched through iterative computation and information sharing; the process involves the parameter adjustment of image processing techniques such as image enhancement, denoising, sharpening, equalization, resolution enhancement, etc., to correct image defects caused by equipment degradation;

The security inspection image equipment interferes with the optimized particle cluster to generate a first security inspection image optimized particle capable of optimizing an image aiming at the degradation state of the current equipment;

the method for establishing the interference optimization particle cluster of the security inspection image equipment comprises the following steps:

Defining a specific image quality improvement target to be achieved by the optimization cluster, and constructing objective evaluation indexes suitable for terahertz human body security inspection images according to the optimization target, wherein the evaluation indexes comprise a signal-to-noise ratio, a contrast entropy, a structural similarity index and a peak signal-to-noise ratio, and are used for quantitatively evaluating the optimization effect;

Selecting an optimization algorithm, wherein the optimization algorithm comprises particle swarm optimization, a genetic algorithm, simulated annealing and a deep reinforcement learning algorithm; the adaptability, convergence speed, easiness in implementation and other factors of the algorithm are considered during selection, so that the method is applicable to the terahertz human body security inspection image optimization scene;

Determining how the particle represents an image optimization strategy, each dimension of the particle representing a set of parameter values; defining how to convert the particle codes into actual image processing operations, designing a decoding function, and mapping the particle vectors into specific image processing parameter sets so as to be applied to the image to be optimized;

generating initial particle groups according to the dimension of the optimization problem, wherein each particle represents a possible optimization strategy;

setting proper control parameters, influencing the particle search range and the convergence speed, setting termination criteria, and preventing the algorithm from falling into local excessive iteration;

embedding the evaluation index into an optimization algorithm to form a fitness function, and evaluating the optimization effect of each particle according to the fitness function in each iteration process to guide particle updating;

Training and verifying the built optimization cluster by using a representative terahertz human body security inspection image sample set and a corresponding standard optimization result; and (3) adjusting algorithm parameters, optimizing targets or evaluation indexes to optimize cluster performance, and ensuring that the image quality can be effectively improved under various equipment degradation conditions.

In the step, an optimization strategy with strong pertinence is generated according to the key image quality characteristics in the equipment degradation state, so that the optimized particle cluster can be better adapted to the equipment state change; compared with the traditional fixed pattern image post-processing algorithm, the method has more flexibility and adaptability; through iterative computation and information sharing of an internal optimization algorithm, an optimization particle cluster can intelligently find an optimization strategy capable of improving image quality to the greatest extent; the method can be better adapted to complex and changeable equipment states, so that the quality and accuracy of security inspection images are improved; optimizing the image by the particle cluster according to the degradation state of the equipment, and correcting the image defect caused by the degradation of the equipment through the parameter adjustment of image processing technologies such as image enhancement, denoising, sharpening, equalization, resolution improvement and the like; the quality of the security inspection image can be improved, so that the security inspection efficiency and accuracy are improved; the optimization effect can be quantitatively evaluated by defining a specific image quality improvement target and an objective evaluation index suitable for terahertz human body security inspection images; the method is favorable for more accurately evaluating the effect of the optimization algorithm and guiding the subsequent optimization process; considering the characteristics of the terahertz human body security inspection image optimization scene, a plurality of optimization algorithms are selected, and the most suitable optimization algorithm can be selected according to actual conditions so as to achieve better optimization effect;

In summary, the method for establishing the interference optimization particle cluster of the security inspection image equipment in the S6 can effectively improve the quality of the terahertz human body security inspection image, so that the security inspection efficiency and accuracy are improved, and the requirements of the security inspection image optimization field on accurate and efficient optimization technology are met.

S7, optimizing the human body security inspection image to be optimized according to the first security inspection image optimizing particles to obtain a human body security inspection optimized image;

S7, based on an image processing strategy indicated by the optimized particles of the first security inspection image, carrying out targeted optimization processing on the terahertz human body security inspection image to be optimized, enabling a power-assisted security inspection personnel to identify potential threat objects more accurately and efficiently, and improving the overall efficiency of a security inspection system;

The method for acquiring the optimized image of the human body security inspection comprises the following steps:

S71, determining an image processing strategy applicable to the current security inspection image according to an optimization strategy provided by the first security inspection image optimization particles;

S72, parameter adjustment and optimization are carried out on the selected image processing strategy according to the parameter values contained in the first security inspection image optimization particles; parameters are adjusted according to specific image characteristics and optimization targets so as to ensure that the image quality is improved to the greatest extent;

s73, monitoring quality change of the processed image in real time in the process of processing the human body security inspection image to be optimized, and adjusting an image processing strategy according to feedback information;

S74, performing quality evaluation and verification on the processed optimized image to ensure that the improvement of the image quality accords with the expectation and can meet the security inspection requirement;

and S75, outputting and recording the finally optimized human body security inspection image for subsequent security inspection tasks.

In the step, the terahertz human body security inspection image to be optimized is subjected to targeted processing and parameter optimization according to the optimization strategy indicated by the first security inspection image optimization particles, so that the definition, contrast and detail presentation of the image can be remarkably improved, and the identification capability of security inspection personnel on potential threat objects is enhanced; the optimized human body security inspection image can help security inspection personnel to detect and judge potential threat articles more accurately and rapidly, so that the conditions of missing inspection and false inspection are reduced, and the overall efficiency and the working efficiency of the security inspection system are improved; according to the optimization strategy provided by the first security inspection image optimization particles, the image processing strategy is adjusted in real time so as to cope with security inspection image characteristics and optimization requirements under different situations, and the adaptability and flexibility of the processing method are ensured; in the processing process, the quality change of the processed image is monitored in real time, and the processing strategy is adjusted according to the feedback information, so that the instantaneity of the processing process and the accuracy of the effect can be ensured; the quality evaluation and verification are carried out on the optimized image, so that the improvement of the image quality is ensured to meet the expectations, the security inspection requirement is met, and the reliability and the practicability of the optimization method are enhanced; outputting and recording the finally optimized human body security inspection image, providing reliable image resources for subsequent security inspection task use, and providing basis for tracing and optimizing the processing process;

In summary, the beneficial effect of the step S7 is that the quality and the identification accuracy of the security inspection image are improved, the overall efficiency and the working efficiency of the security inspection system are enhanced, and meanwhile, the flexibility and the reliability of the processing method are ensured.

Embodiment two: as shown in fig. 3, the terahertz human body security inspection image optimization system of the invention specifically comprises the following modules;

The real-time data acquisition module is used for acquiring real-time running state data information of the security inspection equipment and a human body security inspection image to be optimized;

the equipment degradation operation parameter identification module is used for acquiring the standard operation state data information of the security inspection equipment, inputting the standard operation state data information of the security inspection equipment and the real-time operation state data information of the security inspection equipment into a pre-trained equipment degradation operation parameter identification model, and acquiring a security inspection equipment degradation operation parameter set;

The correlation characteristic matrix construction module is used for constructing a correlation characteristic matrix between the quality characteristics of the human body security inspection image and the operation parameters of the security inspection equipment according to the historical operation state data of the security inspection equipment and the human body security inspection image corresponding to the historical operation state data of the security inspection equipment; the correlation characteristic matrix comprises correlation between each security inspection equipment operation parameter and each human security inspection image quality characteristic;

The image screening module is used for taking the degradation operation parameter set of the security inspection equipment as an index, and extracting a human body security inspection image quality characteristic set which corresponds to the degradation operation parameter set of the security inspection equipment and has a correlation larger than a preset correlation threshold value from the correlation characteristic matrix;

the optimized particle generation module is used for inputting the quality characteristic set of the human body security inspection image into a pre-built security inspection image equipment interference optimized particle cluster to obtain first security inspection image optimized particles corresponding to the security inspection equipment degradation operation parameter set;

the image optimization processing module is used for carrying out optimization processing on the human body security inspection image to be optimized according to the first security inspection image optimization particles to obtain a human body security inspection optimization image.

The system can timely acquire the operation state data of the security inspection equipment through the real-time data acquisition module and the equipment degradation operation parameter identification module, and automatically adjust the optimization strategy according to the data, so as to keep the self-adaptability of the system to the state change of the equipment;

The correlation characteristic matrix construction module and the image screening module are utilized to accurately screen the human body security inspection image which is related to the equipment degradation operation parameters and has excellent quality, so that the accuracy and the efficiency of the security inspection image are improved;

by means of the optimized particle generation module and the image optimization processing module, optimized particles can be generated according to actual conditions, and the images are intelligently processed, so that the image quality and the security inspection effect are effectively improved;

the system considers a plurality of factors such as the running state of equipment, the image quality characteristics, the correlation degree and the like, and comprehensively analyzes and processes the factors, so that the optimization method is more comprehensive and comprehensive, and can cope with the complex and changeable actual scene demands;

The system adopts a pre-trained model and a cluster, has higher reliability and stability, can continuously and effectively perform security inspection image optimization processing, and ensures the successful completion of security inspection tasks;

By combining the advantages, the terahertz human body security inspection image optimization system can effectively cope with challenges such as equipment state change in a real security inspection environment, improve security inspection image quality and security inspection efficiency, and ensure the accuracy and high efficiency of security inspection.

The various modifications and specific embodiments of the terahertz human body security inspection image optimization method in the first embodiment are equally applicable to the terahertz human body security inspection image optimization system of this embodiment, and by the foregoing detailed description of the terahertz human body security inspection image optimization method, those skilled in the art can clearly know the implementation method of the terahertz human body security inspection image optimization system in this embodiment, so that, for brevity of the description, details will not be described here.

In addition, the application also provides an electronic device, which comprises a bus, a transceiver, a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the transceiver, the memory and the processor are respectively connected through the bus, and when the computer program is executed by the processor, the processes of the method embodiment for controlling output data are realized, and the same technical effects can be achieved, so that repetition is avoided and redundant description is omitted.

The foregoing is merely a preferred embodiment of the present invention, and it should be noted that it will be apparent to those skilled in the art that modifications and variations can be made without departing from the technical principles of the present invention, and these modifications and variations should also be regarded as the scope of the invention.

Claims (10)

1. The terahertz human body security inspection image optimization method is characterized by comprising the following steps of:

Collecting real-time running state data information of security inspection equipment and a human body security inspection image to be optimized;

Acquiring standard running state data information of security inspection equipment;

Inputting the standard operation state data information and the real-time operation state data information of the security inspection equipment into a pre-trained equipment degradation operation parameter identification model to obtain a security inspection equipment degradation operation parameter set;

Constructing a correlation feature matrix between quality features of the human body security inspection image and operation parameters of the security inspection equipment based on the historical operation state data of the security inspection equipment and the human body security inspection image corresponding to the historical operation state data of the security inspection equipment; the correlation characteristic matrix comprises correlation between each security inspection equipment operation parameter and each human security inspection image quality characteristic;

Taking the degradation operation parameter set of the security inspection equipment as an index, and extracting a human body security inspection image quality feature set which corresponds to the degradation operation parameter set of the security inspection equipment and has a correlation larger than a preset correlation threshold value from a correlation feature matrix;

Inputting the quality feature set of the human body security inspection image into a pre-built security inspection image equipment interference optimization particle cluster to obtain first security inspection image optimization particles corresponding to the security inspection equipment degradation operation parameter set;

And carrying out optimization processing on the human body security inspection image to be optimized according to the first security inspection image optimization particles to obtain a human body security inspection optimization image.

2. The terahertz human body security inspection image optimization method of claim 1, wherein the construction method of the equipment degradation operation parameter identification model comprises the following steps:

Collecting operation state data of security inspection equipment, wherein the operation state data comprise light source intensity, detector sensitivity, scanning speed, signal processing parameters and imaging algorithm parameters under different working conditions;

Labeling corresponding equipment degradation levels for each group of security inspection equipment state data to form label data required by supervised learning;

Selecting the characteristic with the most distinguishing power for the degradation state of the equipment from the original equipment state data, and carrying out standardization and normalization processing on the selected characteristic;

Selecting a deep learning model as an infrastructure of a device degradation operation parameter identification model, wherein the deep learning model comprises a logistic regression, a support vector machine, a decision tree, a random forest, a neural network, a convolutional neural network, a cyclic neural network and a long-term and short-term memory network;

dividing the processed characteristics into a training set, a verification set and a test set;

Updating parameters through a back propagation algorithm by using the training set and the corresponding degradation labels, optimizing the identification capacity of the equipment degradation operation parameter identification model on the equipment degradation state, and periodically evaluating the model performance on the verification set;

evaluating the identification accuracy, recall rate and F1 score index of the equipment degradation operation parameter identification model on an independent test set;

And packaging the trained equipment degradation operation parameter identification model into an interface which is easy to call, and integrating the interface into a system.

3. The terahertz human body security inspection image optimization method of claim 1, wherein the construction method of the correlation feature matrix comprises the following steps:

collecting state data of human body security inspection equipment in a historical operation state and corresponding generated human body security inspection images;

extracting quality characteristics of each historical security inspection image;

Measuring the association strength between the operation parameters of each security inspection device and the quality characteristics of each human security inspection image;

calculating a correlation coefficient between each set of device operating parameters and image quality features;

and organizing a correlation degree feature matrix according to the corresponding relation between the equipment operation parameters and the image quality features by the calculated correlation coefficients.

4. The terahertz human body security inspection image optimization method of claim 1, wherein the human body security inspection optimized image acquisition method comprises the following steps:

optimizing particles according to the first security inspection image, and determining an image processing strategy;

According to the first security inspection image optimization particles, parameter adjustment and optimization are carried out on the selected image processing strategy;

monitoring the quality change of the processed image in real time, and adjusting the image processing strategy according to the feedback information;

performing quality evaluation and verification on the optimized image after the processing is completed;

and outputting and recording the finally optimized human body security inspection image.

5. The terahertz human body security inspection image optimization method of claim 1, wherein the real-time operation state data information comprises light source intensity, detector sensitivity, scanning speed, signal processing strategy and imaging algorithm parameters.

6. The terahertz human body security inspection image optimization method according to claim 1, wherein the acquisition path of the security inspection equipment standard operation state data information comprises equipment manuals and technical documents, equipment calibration and authentication and experience accumulation.

7. The method for optimizing terahertz human body security inspection images according to claim 1, wherein the set influencing factors of the preset relevant threshold value include security requirements, image quality requirements, equipment operation state change range, historical data analysis and performance balance.

8. A terahertz human body security inspection image optimization system, the system comprising:

The real-time data acquisition module is used for acquiring real-time running state data information of the security inspection equipment and a human body security inspection image to be optimized;

the equipment degradation operation parameter identification module is used for acquiring the standard operation state data information of the security inspection equipment, inputting the standard operation state data information of the security inspection equipment and the real-time operation state data information of the security inspection equipment into a pre-trained equipment degradation operation parameter identification model, and acquiring a security inspection equipment degradation operation parameter set;

The correlation characteristic matrix construction module is used for constructing a correlation characteristic matrix between the quality characteristics of the human body security inspection image and the operation parameters of the security inspection equipment according to the historical operation state data of the security inspection equipment and the human body security inspection image corresponding to the historical operation state data of the security inspection equipment; the correlation characteristic matrix comprises correlation between each security inspection equipment operation parameter and each human security inspection image quality characteristic;

The image screening module is used for taking the degradation operation parameter set of the security inspection equipment as an index, and extracting a human body security inspection image quality characteristic set which corresponds to the degradation operation parameter set of the security inspection equipment and has a correlation larger than a preset correlation threshold value from the correlation characteristic matrix;

the optimized particle generation module is used for inputting the quality characteristic set of the human body security inspection image into a pre-built security inspection image equipment interference optimized particle cluster to obtain first security inspection image optimized particles corresponding to the security inspection equipment degradation operation parameter set;

the image optimization processing module is used for carrying out optimization processing on the human body security inspection image to be optimized according to the first security inspection image optimization particles to obtain a human body security inspection optimization image.

9. A terahertz human body security image optimization electronic device comprising a bus, a transceiver, a memory, a processor and a computer program stored on the memory and executable on the processor, the transceiver, the memory and the processor being connected by the bus, characterized in that the computer program when executed by the processor implements the steps of the method according to any one of claims 1-7.

10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any of claims 1-7.