A security verification method and system for LED display rental

Abstract

The present invention relates to the technical field of security verification systems, and in particular to a security verification method and system for LED display rental. First, the playback content of the target display is obtained, and video features are established based on the playback content. Then, the video features are encrypted to obtain feature data. After that, the rental information of the target display is obtained, and the feature data is decrypted to obtain the video features. The rental information and the video features are matched to obtain a content matching score and a security response is made. Compared with the existing security verification methods, the present invention uses the feature data and the rental information for matching, that is, verifies the consistency between the playback content of the display and the pre-registered rental information to achieve security verification, without the need for relevant personnel to frequently perform identity verification, greatly improving the convenience of security verification, improving work efficiency, and solving the problem that the security verification of LED displays in the prior art is too cumbersome.

Description

Technical Field

[0001] The present invention relates to the technical field of security verification systems, and particularly to a security verification method and system for LED display rental. Background Art

[0002] Due to advantages such as high brightness, low energy consumption, and vivid colors, LED displays are widely used in fields such as advertising, stage performances, sports events, conferences, and exhibitions. With the growth of market demand, the LED display rental industry has developed rapidly. The rental model provides customers with flexible choices. However, the mobility and high value of the equipment also bring risks of theft and transfer. Therefore, the security verification of the display is particularly important during the rental process of LED displays.

[0003] Currently, during the rental process of LED displays, traditional security verification methods such as identity recognition are still used. However, LED displays are all modular structures, and they need to go through a series of complex processes such as transportation, assembly, and calibration during use. Existing verification methods, such as performing identity verification for each module, undoubtedly further increase the complexity of operations during the rental process of LED displays and affect the work efficiency of relevant personnel.

[0004] Therefore, there is a need for a more convenient security verification method and system that can be applied to LED display rental. Summary of the Invention

[0005] Therefore, the present invention provides a security verification method and system for LED display rental to solve the problem that the security verification of LED displays in the prior art is too cumbersome.

[0006] The present invention provides a security verification method for LED display rental, including:

[0007] Obtain the playback content of the target display and establish video features based on the playback content;

[0008] Encrypt the video features to obtain feature data;

[0009] Obtain the rental information of the target display, where the rental information includes information representing the playback content of the target display;

[0010] Obtain the feature data and decrypt the feature data to obtain video features;

[0011] Match the rental information with the video features to obtain a content matching score;

[0012] Perform a security response based on the matching score.

[0013] In a preferred embodiment: The video features include text features, color features, and dynamic features; obtaining the playback content of the target display screen and establishing video features based on the playback content, including:

[0014] Obtaining the video stream of the playback content;

[0015] Extracting the text in the video stream and obtaining text features according to the text content;

[0016] Performing color feature extraction on the video stream to obtain color features;

[0017] Statistical differences between adjacent image frames in the video stream and obtaining dynamic features based on the differences between adjacent image frames.

[0018] In a preferred embodiment: Extracting the text in the video stream and obtaining text features according to the text content, including:

[0019] Performing OCR blurring processing on all image frames in the video stream to obtain text data;

[0020] Performing word frequency statistics on the text data based on the TF-IDF vectorization algorithm to obtain a text word frequency matrix as text features.

[0021] In a preferred embodiment: Performing color feature extraction on the video stream to obtain color features, including:

[0022] Obtaining the image frames in the video and converting the image frames into the HSV color space;

[0023] Performing gamut quantization compression on the image frames in the converted color space and statistically calculating the pixel proportion of each quantization interval to generate a color histogram as color features.

[0024] In a preferred embodiment: Statistical differences between adjacent image frames in the video stream and obtaining dynamic features based on the differences between adjacent image frames, including:

[0025] Statistical differences between adjacent image frames in the video stream to obtain the difference rate of each pair of adjacent image frames;

[0026] Recording multiple consecutive image frames with a difference rate less than a preset difference threshold as an event;

[0027] Obtaining dynamic features according to the timing characteristics of each event in the video stream.

[0028] In a preferred embodiment: The rental information includes keywords; matching the rental information with the video features to obtain a content matching score, including:

[0029] Calculating a first matching score based on the word frequency of the keywords represented by the text word frequency matrix;

[0030] Obtain the preset color mode and preset dynamic mode corresponding to the keyword;

[0031] Match the preset color mode with the color feature to obtain a second matching score;

[0032] Match the preset dynamic mode with the dynamic feature to obtain a third matching score;

[0033] Obtain the preset weight ratio of each video feature;

[0034] Based on the preset weight ratio, perform weighted summation on the first matching score, the second matching score, and the third matching score to obtain a content matching score.

[0035] In a preferred embodiment: calculating the first matching score based on the word frequency of the keyword represented by the text word frequency matrix includes:

[0036] Obtain the word vector of each vocabulary in the text word frequency matrix;

[0037] Taking the word frequency of each vocabulary as the weight, perform weighted averaging on multiple word vectors to obtain a word frequency feature vector;

[0038] Obtain the word vector of the keyword;

[0039] Calculate the dot product of the word vector of the keyword and the word frequency feature vector to obtain the first matching score.

[0040] In a preferred embodiment: obtaining the preset proportional weight of each video feature includes:

[0041] According to the number of non-zero elements in the text word frequency matrix, obtain the text scale eigenvalue;

[0042] If the text scale eigenvalue exceeds the preset scale threshold, obtain the first ratio as the preset proportional weight;

[0043] If the text scale eigenvalue does not exceed the preset scale threshold, obtain the second ratio as the preset proportional weight;

[0044] Wherein, the proportion of the text feature in the first ratio is higher than that in the second ratio.

[0045] In a preferred embodiment: performing a security response according to the matching score includes:

[0046] Judge the security risk level according to the matching score;

[0047] If the security risk level is medium risk, send a screen control lock instruction to the target display screen through the LoRa channel;

[0048] If the security risk level is high risk, send a signal to activate the physical fuse circuit to the target display screen and trigger positioning alarm.

[0049] The present invention also provides a security verification system for LED display screen rental, including a recording module and a verification module. Among them, the recording module is located in the target display screen for rental and is used for:

[0050] Obtain the playing content of the target display screen and establish video features according to the playing content;

[0051] Encrypt the video features to obtain feature data;

[0052] The verification module is used for:

[0053] Obtain the rental information of the target display screen, where the rental information includes information describing the playing content of the target display screen;

[0054] Obtain the feature data, decrypt the feature data to obtain video features;

[0055] Match the rental information and the video features to obtain a content matching score;

[0056] Perform a security response according to the matching score.

[0057] The beneficial effects of adopting the above embodiments are:

[0058] The present invention provides a security verification method and system for LED display screen rental. First, it obtains the playing content of the target display screen, establishes video features according to the playing content, then encrypts the video features to obtain feature data. After that, it obtains the rental information of the target display screen, where the rental information includes information describing the playing content of the target display screen, obtains the feature data, decrypts the feature data to obtain video features, matches the rental information and the video features to obtain a content matching score, and finally performs a security response according to the matching score. Compared with the existing security verification methods, the present invention extracts and encrypts the features of the playing content of the target display screen, so that the video features and the rental information can be matched, that is, the security verification is realized by verifying the consistency between the display screen playing content and the pre-registered rental information, and can respond in time when it is detected that the device is stolen and transferred, without the need for relevant personnel to perform identity verification frequently, greatly improving the convenience of security verification, improving work efficiency, and solving the problem that the LED display screen security verification in the prior art is too cumbersome. Description of the Drawings

[0059] Figure 1 It is a flowchart of a method for an embodiment of the security verification method for LED display screen rental provided by the present invention;

[0060] Figure 2 For Figure 1 The specific step diagram of step S101 in

[0061] Figure 3 For Figure 1 The specific step diagram of step S105 in

[0062] Figure 4 The system structure diagram of an embodiment of the security verification system for LED display rental provided by the present invention. Specific embodiments

[0063] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0064] Combined with Figure 1 As shown, a specific embodiment of the present invention discloses a security verification method for LED display rental, including:

[0065] S101. Obtain the playback content of the target display screen and establish video features according to the playback content;

[0066] S102. Encrypt the video features to obtain feature data;

[0067] S103. Obtain the rental information of the target display screen, and the rental information includes information representing the playback content of the target display screen;

[0068] S104. Obtain the feature data and decrypt the feature data to obtain video features;

[0069] S105. Match the rental information and the video features to obtain a content matching score;

[0070] S106. Perform a security response according to the matching score.

[0071] In the above steps, the rental information refers to the relevant information registered and input by the rental personnel when renting the display screen, which is mainly used to describe the playback content of the display screen. Steps S101 - S102 can be run by the hardware inside the target display screen, and its purpose is to record the playback content of the target display screen. Steps S103 - S106 mainly perform security verification, which can be run by the target display screen itself. For example, security verification is performed regularly during use and security responses are made in a timely manner. It can also be run by other facilities such as a computer in the warehouse. For example, security verification is performed when the display screen is returned. If an abnormality is found, the stolen display screen can be remotely controlled for security response.

[0072] Compared with the existing security verification methods, the present invention extracts and encrypts the features of the playback content of the target display screen, so that the video features and rental information can be matched. That is, by verifying the consistency between the display screen playback content and the pre - registered rental information, security verification is achieved, and a response can be made in a timely manner when it is detected that the device has been stolen and transferred. There is no need for relevant personnel to frequently perform identity verification, which greatly improves the convenience of security verification, improves work efficiency, and solves the problem that the security verification of LED display screens in the prior art is too cumbersome.

[0073] It can be imagined that directly saving the playback content of the target display screen and using the video itself as the video feature for comparison with the rental information is the most ideal situation. However, considering the privacy of users and the hardware conditions of the LED display screen itself, the above method is obviously difficult to implement.

[0074] Therefore, as shown in Figure 2 The present invention also provides a preferred embodiment. In this embodiment, the video features include text features, color features, and dynamic features. In the above step S101, obtaining the playback content of the target display screen and establishing video features according to the playback content specifically includes:

[0075] S201. Obtain the video stream of the playback content;

[0076] S202. Extract the text in the video stream and obtain text features according to the text content;

[0077] S203. Extract color features from the video stream to obtain color features;

[0078] S204. Count the differences between adjacent image frames in the video stream and obtain dynamic features according to the differences between adjacent image frames.

[0079] In the above process, the text features may be the text content information contained in the video stream. For example, it specifically includes advertising slogans, news headlines, program names, etc., and may also be the statistical word frequency, keywords, etc. By extracting and analyzing these text features, the theme and key information of the playing content can be effectively identified, so as to compare with the relevant content requirements in the rental information.

[0080] The color features mainly reflect the color information of the video picture. It may include the main color tone, color distribution, color saturation, etc. in the video picture. These are all helpful for matching the requirements of the color style of the playing content in the rental information.

[0081] The dynamic features reflect the characteristics of the video picture changing over time. It can be obtained by statistically analyzing the differences between adjacent image frames in the video stream. For example, in the rental scenario of an LED display screen for a live sports event, the dynamic features may be manifested as the rapid movement of athletes, the movement trajectory of the ball, etc. Specifically, the difference in pixel values between adjacent frames can be calculated, or the optical flow method can be used to estimate the movement speed and direction of the objects in the picture. According to the difference situations of these adjacent image frames, dynamic features such as the speed range of moving objects, the distribution of movement directions, and the frequency of picture changes can be obtained. These dynamic features can be used to determine whether the playing content meets the requirements of the dynamic effects in the rental information.

[0082] In this embodiment, through the above three dimensions, the usage situation of the LED display screen is implicitly recorded, thus avoiding problems such as privacy and data storage capacity, making the present invention more practical.

[0083] Specifically, in a preferred embodiment, the above step S202, extracting the text in the video stream and obtaining text features according to the text content, specifically includes:

[0084] Performing OCR blurring processing on all image frames in the video stream to obtain text data;

[0085] Based on the TF-IDF vectorization algorithm, performing word frequency statistics on the text data to obtain a text word frequency matrix as text features.

[0086] The above process obtains text data by performing OCR blurring on all image frames in the video stream, and can extract text information in the video as comprehensively as possible. Even in the case of certain degrees of blur, noise and other interferences in the video image, OCR blurring can improve the accuracy of text recognition and reduce feature deviations caused by inaccurate text extraction. At the same time, based on the TF-IDF vectorization algorithm, word frequency statistics are performed on the text data to obtain a text word frequency matrix as text features, which can fully represent the importance and distribution of each vocabulary in the text data. The TF-IDF algorithm can measure the importance of a word for a document set or a single document in a corpus. In the scenario of LED display rental, this means that key text information in the playback content (such as specific advertising slogans, important notice content, etc.) can be prominently reflected in the text word frequency matrix, so that the text features can more accurately reflect the text essence of the playback content, which is conducive to accurate matching with the text requirements in the rental information.

[0087] The text word frequency matrix is a quantifiable form of text feature representation that converts text data into a numerical matrix, facilitating subsequent calculation and comparison operations. During the security verification process, it can be conveniently compared quantitatively with the text-related requirements in the rental information (such as specified keywords and their importance weights, etc.), and the matching degree of the text part of the playback content with the rental information can be evaluated by calculating the similarity between matrices, providing strong support for obtaining an accurate content matching score and making the security verification process more scientific, rigorous and efficient.

[0088] Further, in a preferred embodiment, the above step S203 specifically includes:

[0089] Obtain the image frames in the video and convert the image frames into the HSV color space;

[0090] Perform gamut quantization compression on the image frames with the converted color space, and count the pixel proportion of each quantization interval to generate a color histogram as color features.

[0091] In the above process, performing gamut quantization compression on the image frames with the converted color space can effectively reduce the data volume. In video images, color information is often rich and complex, and directly processing the original color data will consume a large amount of computing resources and time. Through gamut quantization compression, the continuous color space is divided into several discrete quantization intervals, and the intervals are used to replace the specific color values, which can greatly reduce the data volume to be processed. At the same time, the quantized color interval data is also more convenient for statistics and analysis, making subsequent operations such as pixel ratio statistics more efficient, which is beneficial to improving the operating efficiency of the entire security verification method, and can complete the extraction and processing of color features in a shorter time, meeting the high real-time requirements of security verification in the LED display rental scenario.

[0092] Statistical analysis of the pixel ratio of each quantization interval and generating a color histogram as a color feature can comprehensively characterize the color distribution of the video frame, reflecting the color composition of the frame as a whole. This comprehensive color feature representation method can better reflect the color style and characteristics of the playback content. When matching with the regulations on the color of the playback content in the rental information, it can more accurately determine whether it meets the requirements. For example, the rental information may stipulate that the playback content should be mainly in a certain specific color theme, and the color histogram can be used for intuitive comparison and verification, improving the accuracy and effectiveness of security verification.

[0093] Further, in a preferred embodiment, the above step S204, which is to statistically analyze the differences between adjacent image frames in the video stream and obtain dynamic features based on the differences between adjacent image frames, specifically includes:

[0094] Statistically analyze the differences between adjacent image frames in the video stream to obtain the difference rate of each pair of adjacent image frames;

[0095] Record multiple consecutive image frames with a difference rate less than a preset difference threshold as an event;

[0096] Obtain dynamic features based on the timing characteristics of each event in the video stream.

[0097] In the above process, by counting the differences between adjacent image frames in the video stream and obtaining the difference rate of each pair of adjacent image frames, the dynamic changes of the video screen can be accurately captured. This frame-by-frame comparison method can meticulously discover dynamic information such as the movement of objects in the picture and the switching of scenes. Recording multiple continuous image frames with a difference rate less than the preset difference threshold as an event can effectively filter out key events in the video. During video playback, there may be a large number of picture changes caused by some insignificant factors (such as slight light fluctuations, small background jitters, etc.), and these changes have no substantial significance for judging the dynamic characteristics of the playback content. By setting a difference threshold, those small, non-critical changes are filtered out, and only continuous image frames with relatively stable differences are treated as an event, which can highlight the main dynamic events in the video, such as a complete action process, a scene switch, etc., so that the dynamic features are more focused on the key content, improving the effectiveness and representativeness of the dynamic features, and facilitating a more accurate assessment of whether the dynamics of the playback content meet the rental requirements during the security verification process.

[0098] The temporal features of the video stream in the above process, including the occurrence time, duration, interval time between events, etc., can fully characterize the temporal characteristics of the dynamic changes of the video screen. These features describe the rhythm and regularity of the video dynamics from the time dimension, thereby generating a feature description that can fully reflect the dynamic characteristics of the playback content.

[0099] Similarly, similar to the problem of recording the content played by the LED display screen in the previous article, when performing security verification, if the user directly provides the video content to be displayed as the rental information, it will undoubtedly reduce the difficulty of matching and improve accuracy. However, this is only an ideal situation. In reality, the user may not be able to provide complete and accurate display content when renting, so the present invention also provides a preferred embodiment.

[0100] Combination Figure 3 As shown, in this embodiment, the rental information includes keywords, and the above step S105, matching the rental information with the video features to obtain a content matching score, specifically includes:

[0101] S301, calculating a first matching score based on the word frequency of the keyword represented by the text word frequency matrix;

[0102] S302, obtaining a preset color mode and a preset dynamic mode corresponding to the keyword;

[0103] S303, matching the preset color mode and the color feature to obtain a second matching score;

[0104] S304, matching the preset dynamic pattern with the dynamic feature to obtain a third matching score;

[0105] S305. Obtain the preset weight ratio of each video feature;

[0106] S306. Perform weighted summation on the first matching score, the second matching score, and the third matching score based on the preset weight ratio to obtain the content matching score.

[0107] This embodiment provides a method for matching keywords with the above three video features, enabling users to only provide some keywords related to the playing content when registering rental information to achieve the matching of rental information and display records. While ensuring security, it greatly improves the flexibility of use and enhances practicality.

[0108] In the above process, the preset color mode and the preset dynamic mode are respectively data samples corresponding to specific keywords stored in advance. For example, if the keyword is "wedding", the corresponding preset color mode may include a color histogram mainly composed of warm colors such as red and pink. If the keyword is "science and technology product display", then the preset color mode may be a color mode mainly composed of cold colors such as blue and silver. In practice, through the analysis of a large number of actual rental cases and relevant industry practices, a color mode library can be established to associate different keywords with the corresponding preset color modes, so as to accurately obtain the preset color mode corresponding to the keyword in the rental information during the security verification process, and thus effectively match it with the color feature in the video feature.

[0109] Similarly, the preset dynamic mode is also determined based on the characteristics of the rental scenario and keywords. Taking the rental keyword "sports event" as an example, the preset dynamic mode may include specific data such as frequency and frame length (which needs to be determined in combination with the coding method of specific dynamic features). Similarly, a dynamic mode library can be constructed to match keywords with the corresponding preset dynamic modes, so that during security verification, the preset dynamic mode can be accurately obtained according to the keyword in the rental information and compared with the dynamic feature in the video feature to evaluate whether the dynamic of the playing content meets the rental requirements.

[0110] The matching of text features is slightly different from the above two features. In a preferred embodiment, the above step S301. Calculate the first matching score based on the word frequency of the keyword represented by the text word frequency matrix, specifically including:

[0111] Obtain the word vector of each vocabulary in the text word frequency matrix;

[0112] Taking the word frequency of each vocabulary as the weight, perform weighted averaging on multiple word vectors to obtain the word frequency feature vector;

[0113] Obtain the word vector of the keyword;

[0114] Calculate the dot product of the keyword's word vector and the word frequency feature vector to obtain the first matching score.

[0115] In the above process, by obtaining the word vectors of each vocabulary in the text word frequency matrix, the semantic information of each vocabulary can be accurately represented. The word vectors can adopt the TF-IDF vectors mentioned above or can be obtained using other word embedding models. Taking the word frequency of each vocabulary as the weight, the word frequency feature vector is obtained by weighted averaging of multiple word vectors, which can highlight the semantic contribution of the key vocabulary in the text. On the premise that the word vectors can correctly represent the semantic information of words, calculating the dot product of the keyword's word vector and the word frequency feature vector to obtain the first matching score can quantify the semantic matching degree between the keyword and the text features, and obtain an objective and accurate first matching score. This method can fully consider the semantic information of vocabulary and its importance in the text, avoid the errors that may be brought by simple matching based on the surface form of vocabulary, improve the accuracy and reliability of text feature matching, and thus provide more powerful support for the entire security verification process.

[0116] Further, in a preferred embodiment, the above step S305, obtaining the preset proportional weights of each video feature, specifically includes:

[0117] Obtain the text scale eigenvalue according to the number of non-zero elements in the text word frequency matrix;

[0118] If the text scale eigenvalue exceeds the preset scale threshold, obtain the first ratio as the preset proportional weight;

[0119] If the text scale eigenvalue does not exceed the preset scale threshold, obtain the second ratio as the preset proportional weight;

[0120] Wherein, the proportion of the text features corresponding to the first ratio is higher than that of the second ratio.

[0121] It can be understood that in practice, the LED display screen does not display text in all scenarios. Therefore, in this embodiment, by analyzing the word frequency matrix, the text scale eigenvalue is used to judge the scale of the number of words displayed by the LED, and then the proportion of text features in the security verification is dynamically adjusted based on the text quantity scale to improve the scientific rationality of the verification. For example, when the text word frequency matrix is relatively sparse, it indicates that the number of words displayed on the LED display screen is small, and the text scale eigenvalue is naturally low. Then, at this time, the weight corresponding to the text features should be reduced during the security verification to reduce its influence.

[0122] Further, in a preferred embodiment, the above step S106, performing a security response according to the matching score, specifically includes:

[0123] Judge the security risk level according to the matching score;

[0124] If the security risk level is medium risk, send a screen control lock instruction to the target display screen through the LoRa channel;

[0125] If the security risk level is high risk, send a signal to activate the physical fuse circuit to the target display screen and trigger a positioning alarm.

[0126] It can be understood that in practice, according to specific situations, other corresponding methods can also be adopted.

[0127] Combined with Figure 4 As shown, the present invention also provides a security verification system for LED display screen rental, including a recording module 410 and a verification module 420. Among them, the recording module is located in the rented target display screen and is used for:

[0128] Obtain the playing content of the target display screen, and establish video features according to the playing content;

[0129] Encrypt the video features to obtain feature data;

[0130] The verification module is used for:

[0131] Obtain the rental information of the target display screen, and the rental information includes information describing the playing content of the target display screen;

[0132] Obtain the feature data, and decrypt the feature data to obtain video features;

[0133] Match the rental information and the video features to obtain a content matching score;

[0134] Perform a security response according to the matching score.

[0135] It should be noted here that: the corresponding system provided in the above embodiments can implement the technical solutions described in the above method embodiments. The specific implementation principles of the above modules or units can be referred to the corresponding content in the above method embodiments, and will not be elaborated here.

[0136] The present invention provides a security verification method and system for LED display rental. First, the playing content of a target display is obtained, and video features are established based on the playing content. Then, the video features are encrypted to obtain feature data. After that, the rental information of the target display is obtained, where the rental information includes information describing the playing content of the target display. The feature data is obtained and decrypted to obtain the video features. The rental information and the video features are matched to obtain a content matching score. Finally, a security response is made according to the matching score. Compared with the existing security verification methods, the present invention extracts and encrypts the features of the playing content of the target display, so that the video features and the rental information can be matched, that is, the security verification is realized by verifying the consistency between the playing content of the display and the pre-registered rental information, and can respond in time when it is detected that the device is stolen and transferred, without the need for relevant personnel to frequently perform identity verification, greatly improving the convenience of security verification, improving work efficiency, and solving the problem that the security verification of LED displays in the prior art is too cumbersome.

[0137] It should be noted that the embodiments in this specification are all described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other.

[0138] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A security verification method for LED display screen rental, characterized in that: include: Obtain the playback content of the target display screen and establish video features based on the playback content; Encrypt the video features to obtain feature data; Obtaining the rental information of the target display screen, where the rental information includes information describing the playback content of the target display screen; Acquire feature data, and decrypt the feature data to obtain video features; Match the rental information and video features to obtain a content matching score; Perform security response based on matching scores; The video features include text features, color features, and dynamic features. The playback content of the target display screen is obtained, and video features are established according to the playback content, including: Get the video stream of the playback content; Perform OCR blur processing on all image frames in the video stream to obtain text data; The word frequency of text data is counted based on the TF-IDF vectorization algorithm, and the text word frequency matrix is ​​obtained as the text feature; Extract color features from the video stream to obtain color features; Count the differences between adjacent image frames in the video stream to obtain the difference rate of each pair of adjacent image frames; Recording a plurality of image frames whose difference rate is less than a preset difference threshold and are continuous as one event; According to the temporal characteristics of each event in the video stream, dynamic features are obtained; The rental information includes keywords; the rental information and video features are matched to obtain a content matching score, including: Calculate a first matching score based on the word frequency of the keyword represented by the text word frequency matrix; Acquire a preset color mode and a preset dynamic mode corresponding to the keyword, wherein the preset color mode is a pre-stored color data sample corresponding to the specific keyword, and the preset dynamic mode is a pre-stored video dynamic data sample corresponding to the specific keyword; Matching the preset color mode and the color feature to obtain a second matching score; Matching the preset dynamic pattern and the dynamic feature to obtain a third matching score; Get the preset weight ratio of each video feature; The first matching score, the second matching score and the third matching score are weighted and summed based on a preset weight ratio to obtain a content matching score.

2. The security verification method for LED display screen leasing according to claim 1 is characterized in that: Extract color features from the video stream to obtain color features, including: Get the image frame in the video and convert it to HSV color space; The image frame in the converted color space is compressed by color gamut quantization, and the pixel ratio of each quantization interval is counted to generate a color histogram as a color feature.

3. The security verification method for LED display screen leasing according to claim 1 is characterized in that: Based on the word frequency of the keyword represented by the text word frequency matrix, the first matching score is calculated, including: Get the word vector of each word in the text word frequency matrix; Taking the frequency of each word as the weight, multiple word vectors are weighted averaged to obtain the word frequency feature vector; Get the word vector of the keyword; Calculate the dot product of the keyword's word vector and the word frequency feature vector to get the first matching score.

4. The security verification method for LED display screen leasing according to claim 1 is characterized in that: Get the preset weight ratio of each video feature, including: According to the number of non-zero elements in the text word frequency matrix, the text scale feature value is obtained; If the text scale feature value exceeds the preset scale threshold, obtaining the first ratio as the preset weight ratio; If the text scale feature value does not exceed the preset scale threshold, obtaining the second ratio as the preset weight ratio; Among them, the proportion corresponding to the text features in the first ratio is higher than that in the second ratio.

5. The security verification method for LED display screen leasing according to claim 1 is characterized in that: Security responses based on match scores include: Determine the security risk level based on the matching score; If the security risk level is medium risk, a screen control lock command is sent to the target display screen through the LoRa channel; If the security risk level is high, a signal to activate the physical fuse circuit is sent to the target display screen and a positioning alarm is triggered.

6. A security verification system for LED display screen rental, characterized in that: It includes a recording module and a verification module, wherein the recording module is located in the rented target display screen and is used for: Obtain the playback content of the target display screen and establish video features based on the playback content; Encrypt the video features to obtain feature data; Authentication module, used to: Obtaining the rental information of the target display screen, where the rental information includes information describing the playback content of the target display screen; Acquire feature data, and decrypt the feature data to obtain video features; Match the rental information and video features to obtain a content matching score; Perform security response based on matching scores; The video features include text features, color features, and dynamic features. The playback content of the target display screen is obtained, and video features are established according to the playback content, including: Get the video stream of the playback content; Perform OCR blur processing on all image frames in the video stream to obtain text data; The word frequency of text data is counted based on the TF-IDF vectorization algorithm, and the text word frequency matrix is ​​obtained as the text feature; Extract color features from the video stream to obtain color features; Count the differences between adjacent image frames in the video stream to obtain the difference rate of each pair of adjacent image frames; Recording a plurality of image frames whose difference rate is less than a preset difference threshold and are continuous as one event; According to the temporal characteristics of each event in the video stream, dynamic features are obtained; The rental information includes keywords; the rental information and video features are matched to obtain a content matching score, including: Calculate a first matching score based on the word frequency of the keyword represented by the text word frequency matrix; Acquire a preset color mode and a preset dynamic mode corresponding to the keyword, wherein the preset color mode is a pre-stored color data sample corresponding to the specific keyword, and the preset dynamic mode is a pre-stored video dynamic data sample corresponding to the specific keyword; Matching the preset color mode and the color feature to obtain a second matching score; Matching the preset dynamic pattern and the dynamic feature to obtain a third matching score; Get the preset weight ratio of each video feature; The first matching score, the second matching score and the third matching score are weighted and summed based on a preset weight ratio to obtain a content matching score.

Images