Video data analysis processing system and method

Abstract

The invention discloses a video data analysis and processing system and method, a live broadcast system comprises a data acquisition unit, a personnel classification unit, a threshold storage unit and an adaptive processing unit, relates to the technical field of live broadcast video data processing, and solves the problem that the amount of bullet screens in a video is large and cannot be reasonably processed. In order to solve the technical problem that screen refreshing continuously affects experience of anchors and audiences in the prior art, people watching live broadcast are classified, barrages sent by the classified people are analyzed, the barrages are screened and filtered, the barrages are defined as screen refreshing for filtering, and normal barrages are reserved for display, so that the barrages can be displayed in a real-time manner. Therefore, the experience feeling of a user in the watching process is improved, meanwhile, the experience feeling of an anchor watching the bullet screens during live broadcasting can be ensured, output is performed according to the bullet screens of different levels of personnel, and the bullet screens with high levels are output in priority, so that the anchor can conveniently and timely see the bullet screens, and the efficiency of live broadcasting is improved. And therefore, the real-time interaction experience feeling of the high-grade user is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of live video data processing, and in particular to a video data analysis and processing system and method. Background Art

[0002] The online video live broadcast system can respond to customer requirements and transmit the audio or video signals of the event site to the multimedia server after compression, so that it can be listened or watched on the Internet by the general public or authorized specific groups of people.

[0003] According to patent application number CN202010053848.6, the patented data processing method includes: receiving first-type data from a server and displaying the first-type data on a first display interface; responding to an update request, converting the first-type data into second-type data, and displaying the second-type data on a second display interface, wherein the second-type data uses a different communication protocol than the first-type data. This application solves the technical problem of waste of server resources caused by the data transmission method of the live broadcast system in the related art.

[0004] When some existing live broadcast systems are in use, a large number of barrages will be generated because live broadcasts involve real-time interaction with the audience. For anchors with high popularity and traffic, a large number of barrages will be generated during live broadcasts. Due to the complexity of the content of the barrages, there is also the phenomenon of screen swiping. At the same time, there are also barrages swiped by robots. Various complex barrages appear on the screen, which will affect the real-time interaction between the live broadcast and the audience on the one hand, and the phenomenon of screen swiping will also reduce the user experience. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a video data analysis and processing system and method, which solves the problem that the large amount of barrage in the video cannot be reasonably processed, and the screen is constantly swiping, affecting the experience of the anchor and the audience.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A video data analysis and processing method includes the following steps:

[0007] Step 1: The data acquisition end obtains the basic information of the target object, where the target object includes the live broadcast online person, the basic information includes the level, the number of barrages and the content of the barrages, and transmits the obtained target accumulation basic information to the data processing end;

[0008] Step 2: The data processing end obtains the basic information of the transmitted target object and classifies the target object into robots, divers, and active personnel based on the basic information of the target object. The specific classification method is based on the number of barrages and the level of the target object;

[0009] Step 3: Then obtain the bullet screen content in the basic information of the target object, and at the same time analyze and process the bullet screen content of the target object according to the obtained personnel classification information. By matching the bullet screen content, the bullet screen is divided into bullet screens to be processed and filtered bullet screens. Then obtain the number K of bullet screens corresponding to the target object and the bullet screen content of all bullet screens to be processed, and divide them into normal bullet screens and spamming bullet screens;

[0010] Step 4: Sort out the analyzed normal bullet screens. By analyzing the personnel to be analyzed corresponding to the normal bullet screens, output the normal bullet screens according to the levels of the personnel to be analyzed, and sort out the normal bullet screen information. Then transmit the normal bullet screen information to the output end;

[0011] Step 5: Sort out the analyzed spamming bullet screens. Mark the personnel to be analyzed corresponding to the spamming bullet screens, and judge the number of spamming bullet screens. When the number of spamming bullet screens is greater than the preset value, the system automatically filters the spamming bullet screens. Otherwise, do nothing, and sort out the spamming bullet screens to obtain spamming bullet screen information, and at the same time transmit it to the output end;

[0012] Step 6: The output end obtains the normal bullet screen information and the spamming bullet screen information and displays them to the operator.

[0013] As a further solution of the present invention: The specific classification method of the target object in Step 2 is as follows:

[0014] S1: First, label the target object and record it as i. Then obtain the number of bullet screens Di corresponding to the target object i within the time T and record it as Di, and the corresponding level and record it as Ji;

[0015] S2: Then analyze the obtained number of bullet screens Di of the target object. The specific analysis method is as follows:

[0016] S21: When Di≥YS1, the system determines that the corresponding target object is a pre-active person. Then obtain the level Ji of the pre-active person. When Ji≥YS2, the system determines that the pre-active person is an active person and highlights it. Otherwise, when Ji < YS2, the system determines that the pre-active person is a pending person and obtains the basic information of the pending person correspondingly;

[0017] S22: When Di < YS1, the system determines that the corresponding target object is a person to be analyzed. Then obtain the level Ji of the person to be analyzed. When Ji≥YS2, the system determines that the person to be analyzed is a diving person and highlights it. Otherwise, when Ji < YS2, the system determines that the person to be analyzed is a robot, and label the robot and record it as Qj, where j represents the label of the robot.

[0018] As a further solution of the present invention: The specific method for analyzing and processing the barrage content of the target object in step three is as follows:

[0019] P1: Obtain the barrage content generated in real time, and then match the barrage. The specific matching method is as follows:

[0020] Obtain the target object corresponding to the barrage. When the target object corresponding to the barrage obtained is a robot, mark the corresponding barrage as a filtered barrage, filter the corresponding barrage and store it. When the target object corresponding to the barrage obtained is an active person or a lurker, mark the corresponding barrage as a pending processing barrage;

[0021] P2: Then obtain all the pending processing barrages, and at the same time obtain the number K of barrages and the barrage content of the target object corresponding to the pending processing barrages, and analyze the pending processing barrages in combination with the obtained target object barrage generation speed and barrage content.

[0022] As a further solution of the present invention: The specific classification method for the pending processing barrages in P2 is as follows:

[0023] P21: Identify the barrage content to obtain the same barrages and different barrages among the pending processing barrages, and mark the target object corresponding to the same barrages as a pending analysis person. Then analyze the obtained same barrages in combination with the number K of barrages;

[0024] P22: Then obtain the number of the same barrages generated by the pending analysis person within t time and record it as K. Then substitute K and time t into the calculation formula: Obtain the real-time barrage speed. Then compare the calculated real-time barrage speed SDs with SDz;

[0025] P23: When SDs≥SDz, the system determines that the barrages generated by the pending analysis person are spam barrages and filters the spam barrages. On the contrary, when SDs<SDz, the system determines that the barrages generated by the pending analysis person are normal barrages.

[0026] As a further solution of the present invention: The specific method for sorting out the normal barrage information from the normal barrages in step four is as follows:

[0027] A1: Obtain all the normal barrages, and obtain the level of the corresponding pending analysis person. Then sort all the normal barrages according to the level of the pending analysis person. The specific sorting method is: Sort in the order from large to small according to the level to obtain the sorted barrages, and at the same time highlight the sorted barrages;

[0028] A2: Further determine whether the person to be analyzed is a VIP. If he / she is wearing a VIP badge, he / she is marked as a VIP. Otherwise, if he / she is not wearing a VIP badge, he / she is marked as an ordinary person. Then, the sorted comments are matched with the corresponding person to be analyzed, and the matched sorted comments are output according to the VIP badge.

[0029] As a further solution of the present invention: the specific method of sorting VIPs and ordinary personnel in A2 is:

[0030] A21: The output method for sorting VIP comments is: output from the highest to the lowest according to the VIP medal level;

[0031] A22: The output method for ranking the bullet comments of ordinary people is: output from large to small according to the level of ordinary people.

[0032] As a further solution of the present invention: the specific processing method of sorting out the screen-sweeping barrage information in step 5 is as follows:

[0033] All screen-sweeping barrages are obtained, and at the same time, the users corresponding to the screen-sweeping barrages are obtained and marked as screen-sweeping users. Then, the screen-sweeping barrages of the screen-sweeping users are analyzed. When the number of screen-sweeping barrages generated by the screen-sweeping users is greater than the preset value, the system automatically filters the barrages of the screen-sweeping users, and organizes and stores the filtered screen-sweeping barrages. When the number of screen-sweeping barrages generated by the screen-sweeping users is less than the preset value, no processing is done. Then, the information of the screen-sweeping barrages is organized to produce screen-sweeping barrage information and transmit it to the output end.

[0034] Beneficial effects

[0035] The present invention provides a video data analysis and processing system and method. Compared with the existing technology, it has the following advantages:

[0036] The present invention classifies people who watch live broadcasts, analyzes the barrages sent by the classified people, and screens and filters the barrages. The barrages are defined as screen-sweeping and filtered, and normal barrages are retained for display, thereby improving the user experience during the viewing process. At the same time, it can also ensure the experience of the anchor watching the barrages during the live broadcast. Secondly, the barrages are output according to the different levels of people, and the high-level barrages are output with priority, so that the anchor can see them in time, thereby improving the real-time interactive experience of high-level users. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a block diagram of the system principle of the present invention; Figure 2 It is a diagram of the steps of the present invention. DETAILED DESCRIPTION

[0038] 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.

[0039] Example 1. Please refer to Figure 1 , the present application provides a data processing method for a live broadcast system, and the method specifically includes the following steps:

[0040] Step 1: The data acquisition end acquires the basic information of the target object. The target object includes the live online personnel, and the basic information includes the level, the number of bullet screens, and the bullet screen content, and transmits the acquired basic information of the target object to the data processing end;

[0041] Step 2: The data processing end acquires the transmitted basic information of the target object, and classifies the target object into: robots, divers, and active personnel according to the basic information of the target object. The specific personnel classification method is as follows:

[0042] S1: First, label the target object and denote it as i. Here, i represents the label of the online live viewers, and the labeling order is carried out according to the VIP sorting order of the live broadcast room. The VIP sorting order is a mature technology in the existing live broadcast, and will not be elaborated here. Then, obtain the number of bullet screens Di corresponding to the target object i within the time T and denote it as Di, and the corresponding level and denote it as Ji. Here, the time T is a preset value. In the present application, the value of the time T is 5 days, and the number of bullet screens Di is the total number of bullet screens within 5 days;

[0043] S2: Then, analyze the obtained number of bullet screens Di of the target object. The specific analysis method is as follows:

[0044] S21: When Di≥YS1, where YS1 is a preset value, and YS1 represents the value of the number of bullet screens obtained according to past data statistics. The system determines that the corresponding target object is a pre-active person. Then, obtain the level Ji of the pre-active person. When Ji≥YS2, the system determines that the pre-active person is an active person and highlights it. On the contrary, when Ji<YS2, where YS2 represents the meaning that the system obtains the level threshold according to big data statistics, the system determines that the pre-active person is a pending person, and obtains the basic information of the pending person accordingly. Here, the pending person means: the person has a low level and sends a lot of bullet screens within the time T, but just doesn't watch the live broadcast often;

[0045] S22: When Di < YS1, the system determines that the corresponding target object is a person to be analyzed. Then, the level Ji of the person to be analyzed is obtained. When Ji ≥ YS2, the system determines that the person to be analyzed is a diver and highlights the person. On the contrary, when Ji < YS2, the system determines that the person to be analyzed is a robot, labels the robot and records it as Qj, where j represents the label of the robot.

[0046] Analyzed in combination with the actual application scenario, for the number of bullet screens counted within a certain period of time, if the number of bullet screens of the corresponding person is lower than the preset value YS1, then the person is classified as a person to be analyzed. At the same time, combined with the level of the person to be analyzed, if the level is lower than the preset value YS2, the system directly determines that it is a robot. There are many robots in existing live rooms, mostly with low levels and infrequent bullet screen postings. On the contrary, if the level is higher than the preset value YS2, the system determines that it is a diver. When some people watch the live broadcast, they just watch the live broadcast silently and rarely post bullet screens, then such people are defined as divers. At the same time, the highlighting method is as follows: mark the active people as red, mark the divers as gray, and mark the robots as yellow. Such markings facilitate the anchor to directly distinguish when watching, so as to distinguish the content of the bullet screens sent. <>

[0047] Step 3: Then, obtain the bullet screen content in the basic information of the target object, and at the same time analyze and process the bullet screen content of the target object according to the obtained personnel classification information. The specific analysis and processing methods are as follows:

[0048] P1: Obtain the bullet screen content generated in real time, and then match the bullet screens. The specific matching method is as follows:

[0049] Obtain the target object corresponding to the bullet screen. Here, the target object is the target object after personnel classification. When the target object corresponding to the bullet screen obtained is a robot, mark the corresponding bullet screen as a filtered bullet screen, filter the corresponding bullet screen and store it. When the target object corresponding to the bullet screen obtained is an active person or a diver, mark the corresponding bullet screen as a bullet screen to be processed;

[0050] P2: Then, obtain all the bullet screens to be processed, and at the same time obtain the number K of bullet screens and the bullet screen content of the target object corresponding to the bullet screen to be processed, and analyze the bullet screens to be processed in combination with the obtained bullet screen generation speed and bullet screen content of the target object. The specific analysis method is as follows:

[0051] P21: Identify the bullet screen content to obtain the bullet screens to be processed as identical bullet screens and different bullet screens, and mark the target object corresponding to the identical bullet screen as the person to be analyzed. The specific identification method is as follows: According to the number of times the same content appears in a bullet screen sent by the target object, if the number of times the same content appears in this bullet screen is greater than or equal to three, then define this bullet screen as an identical bullet screen. On the contrary, if no same content appears in this bullet screen, then define this bullet screen as a different bullet screen. Then analyze the obtained identical bullet screens in combination with the bullet screen quantity K.

[0052] P22: Then obtain the number of identical bullet screens generated by the person to be analyzed within time t and record it as K. Then substitute K and time t into the calculation formula: Get the real-time bullet screen speed. Then

[0053] Compare the obtained real-time bullet screen speed SDs with SDz. Here, SDz represents the generation speed of identical bullet screens under normal circumstances.

[0054] P23: When SDs ≥ SDz, the system determines that the bullet screens generated by this person to be analyzed are spam bullet screens and filters these spam bullet screens. On the contrary, when SDs < SDz, the system determines that the bullet screens generated by this person to be analyzed are normal bullet screens.

[0055] Analyze in combination with the actual application scenario. Here, if the number of identical bullet screens generated within a certain period of time is greater than the generation speed of identical bullet screens under normal conditions, it is determined as a spam bullet screen. Under normal circumstances, the people watching the live broadcast will not have the situation of sending spam bullet screens. Only when directly copying the previous sent bullet screen can the speed be greater than that under normal conditions. And in the first embodiment of this example, it is for the case of personal spam bullet screens. The determination of the case of collective spam is carried out in the second embodiment of this invention. On the contrary, if it is less than the generation speed of identical bullet screens under normal conditions, it is determined as a normal bullet screen.

[0056] Step Four: Then sort out the analyzed normal bullet screens and spam bullet screens respectively. Sort out the normal bullet screen information from the normal bullet screens and sort out the spam bullet screen information from the spam bullet screens. The specific sorting methods are as follows:

[0057] A1: Obtain all the normal bullet screens and obtain the levels of the corresponding people to be analyzed. Then sort all the normal bullet screens according to the levels of the people to be analyzed. The specific sorting method is as follows: Sort in the order from largest to smallest level to obtain the sorted bullet screens. At the same time, highlight the sorted bullet screens. Here, it should be noted that the highlighting method is to perform different color marking treatments according to the levels of different people to be analyzed. The higher the level, the darker the color; the lower the level, the lighter the color.

[0058] A2: Further determine whether the person to be analyzed is a VIP. If he / she wears a VIP badge, he / she is marked as a VIP. Otherwise, if he / she does not wear a VIP badge, he / she is marked as an ordinary person. It should be noted that the VIP badge is exclusive to paying users and there are different levels of VIP badges. If the person to be analyzed wears an exclusive VIP badge, he / she is marked as a VIP according to the VIP badge. Otherwise, he / she is marked as an ordinary person. Then, the sorted comments are matched with the corresponding person to be analyzed, and the matched sorted comments are output according to the VIP badge. The specific output method is as follows:

[0059] A21: The output method for sorting VIP comments is: output from the highest to the lowest according to the VIP medal level;

[0060] A22: The output method for ranking the bullet comments of ordinary people is: output from large to small according to the level of ordinary people.

[0061] Analyze in combination with actual application scenarios: When the sorted barrage of VIP personnel is obtained, the level of the VIP personnel is analyzed, and the barrage is sorted and output according to the VIP level from large to small. The VIP level here means the VIP medal level corresponding to the recharge. The higher the VIP medal level, the greater the recharge amount required to activate, and the lower the VIP medal level, the smaller the recharge amount required. In Example 1, the sorted barrage for different VIPs is output.

[0062] Step 5: Get all the screen-sweeping barrages, and at the same time get the users corresponding to the screen-sweeping barrages and mark them as screen-sweeping users, then analyze the screen-sweeping barrages of the screen-sweeping users. When the number of screen-sweeping barrages generated by the screen-sweeping users is greater than the preset value, the system automatically filters the barrages of the screen-sweeping users, and organizes and stores the filtered screen-sweeping barrages. It should be noted here that: the system will automatically retain a barrage, and when the same barrage appears subsequently, it will be filtered. The filtered screen-sweeping barrages will obtain the information of the screen-sweeping users accordingly, and then integrate and store the information of the screen-sweeping users and transmit it to the background storage unit. When the number of screen-sweeping barrages generated by the screen-sweeping users is less than the preset value, no processing will be done, and then the information of the screen-sweeping barrages will be organized to produce screen-sweeping barrage information and transmit it to the output end.

[0063] Step 6: The output end obtains normal barrage information and screen-swiping barrage information and displays it to the operator.

[0064] The second embodiment of the present invention differs from the first embodiment in that the second embodiment analyzes the same bullet comments in a group and the bullet comments of VIPs of the same level;

[0065] The specific analysis method for collectively flooding the screen with identical comments is as follows: starting from the current time, obtain the number of identical comments that appear within time t2. If the number of identical comments within time t2 is greater than the judgment value, the identical comment is defined as a flooding comment. The system automatically filters the flooding comments based on the level of the target object. If the target object's level is less than the judgment value, the target object's comment is directly filtered out; otherwise, no processing is performed.

[0066] The specific analysis method for VIP bullet comments of the same level is as follows: obtain all the sorted bullet comments, and then output them according to the level of the VIP personnel. It should be noted here that the level of the VIP personnel is not the level of the VIP medal, but the level of the VIP personnel themselves.

[0067] Example 3, as the third embodiment of the present invention, focuses on the implementation of Example 1 in combination with Example 2. The specific combination implementation method is: after the analysis of individual screen-sweeping barrages is completed in step 3, the same barrages of the group are analyzed. Then, after the analysis of the sorted barrages of the same VIP personnel is completed in step 4, the sorted barrages of different VIP personnel are analyzed.

[0068] At the same time, the contents not described in detail in this specification belong to the existing technology well known to those skilled in the art.

[0069] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A video data analysis and processing method, characterized in that: The steps include the following: Step 1: The data acquisition end acquires the basic information of the target object, where the target object includes the live online personnel, and the basic information includes the level, the number of bullet screens, and the content of the bullet screens, and transmits the acquired basic information of the target object to the data processing end; Step 2: The data processing end acquires the transmitted basic information of the target object, and divides the target object into robots, divers, and active personnel according to the basic information of the target object. The specific division method is carried out according to the number of bullet screens and the level of the target object; Step 3: Then, the content of the bullet screen in the basic information of the target object is acquired. At the same time, according to the acquired personnel division information, the bullet screen content of the target object is analyzed and processed. The bullet screen is divided into bullet screens to be processed and filtered bullet screens by matching the bullet screen content. Then, the number of bullet screens K and the content of the bullet screens corresponding to all the bullet screens to be processed for the target object are acquired, and they are divided into normal bullet screens and spam bullet screens; Step 4: Sort out the analyzed normal bullet screens. By analyzing the personnel to be analyzed corresponding to the normal bullet screens, the normal bullet screens are output according to the level of the personnel to be analyzed, and the normal bullet screen information is sorted out. Then, the normal bullet screen information is transmitted to the output end; Step 5: Sort out the analyzed spam bullet screens. Mark the personnel to be analyzed corresponding to the spam bullet screens, and judge the number of spam bullet screens. When the number of spam bullet screens is greater than the preset value, the system automatically filters the spam bullet screens. Otherwise, no processing is done, and the spam bullet screens are sorted out to obtain the spam bullet screen information, and at the same time, it is transmitted to the output end; Step 6: The output end acquires the normal bullet screen information and the spam bullet screen information, and displays them to the operator.

2. A video data analysis and processing method according to claim 1, characterized in that: The specific division method of the target object in step 2 is as follows: S1: The target object is pre-labeled and denoted as i. Then, the number of bullet screens Di corresponding to the target object i within the time T is acquired and denoted as Di, and the corresponding level is denoted as Ji; S2: Then, the acquired number of bullet screens Di of the target object is analyzed. The specific analysis method is as follows: S21: When Di≥YS1, the system determines that the corresponding target object is a pre-active personnel. Then, the level Ji of the pre-active personnel is acquired. When Ji≥YS2, the system determines that the pre-active personnel is an active personnel and highlights it. Otherwise, when Ji<YS2, the system determines that the pre-active personnel is a pending personnel, and the basic information of the pending personnel is acquired correspondingly; S22: When Di<YS1, the system determines that the corresponding target object is a personnel to be analyzed. Then, the level Ji of the personnel to be analyzed is acquired. When Ji≥YS2, the system determines that the personnel to be analyzed is a diver and highlights it. Otherwise, when Ji<YS2, the system determines that the personnel to be analyzed is a robot, and the robot is labeled and denoted as Qj, where j represents the label of the robot.

3. The video data analysis and processing method according to claim 1, characterized in that: The specific method for analyzing and processing the bullet screen content of the target object in step 3 is as follows: P1: Acquire the content of the bullet screen generated in real time, and then match the bullet screen. The specific matching method is as follows: Obtain the target object corresponding to the bullet screen. When the target object corresponding to the bullet screen is a robot, mark the corresponding bullet screen as a filtered bullet screen, filter the corresponding bullet screen and store it. When the target object corresponding to the bullet screen is an active user or a lurker, mark the corresponding bullet screen as a pending bullet screen; P2: Then obtain all the pending bullet screens. At the same time, obtain the number K of bullet screens and the bullet screen content corresponding to the target object of the pending bullet screen, and analyze the pending bullet screens in combination with the obtained bullet screen generation speed and bullet screen content of the target object.

4. The video data analysis and processing method according to claim 3, characterized in that: The specific classification method for the pending bullet screens in P2 is as follows: P21: Identify the bullet screen content to obtain the pending bullet screens as the same bullet screens and different bullet screens, and mark the target object corresponding to the same bullet screens as the pending analysis personnel. Then analyze the obtained same bullet screens in combination with the number K of bullet screens; P22: Then obtain the number of identical comments generated by the person to be analyzed within time t and record it as K. Then substitute K and time t into the calculation formula: Get the real-time bullet screen speed, and then compare the calculated real-time bullet screen speed SDs with SDz; P23: When SDs≥SDz, the system determines that the bullet screens generated by the pending analysis personnel are spam bullet screens and filters the spam bullet screens. On the contrary, when SDs<SDz, the system determines that the bullet screens generated by the pending analysis personnel are normal bullet screens.

5. The video data analysis and processing method according to claim 1, characterized in that: The specific method for sorting the normal bullet screens to obtain normal bullet screen information in step 4 is as follows: A1: Obtain all the normal bullet screens and obtain the level of the corresponding pending analysis personnel. Then sort all the normal bullet screens according to the level of the pending analysis personnel. The specific sorting method is: sort in descending order of level to obtain the sorted bullet screens, and at the same time highlight the sorted bullet screens; A2: Further determine whether the pending analysis personnel is a VIP. If a VIP medal is worn, mark it as a VIP. On the contrary, if no VIP medal is worn, mark it as an ordinary person. Then match the sorted bullet screens with the corresponding pending analysis personnel and output the sorted bullet screens after matching according to the VIP medal.

6. A video data analysis and processing method according to claim 5, characterized in that: The specific method for sorting VIPs and ordinary people in A2 is: A21: The output method for sorting bullet screens of VIPs is: output in descending order according to the level of the VIP medal; A22: The output method for sorting bullet screens of ordinary people is: output in descending order according to the level of ordinary people.

7. The video data analysis and processing method according to claim 1, characterized in that: The specific processing method for sorting the spam bullet screens to obtain spam bullet screen information in step 5 is as follows: Obtain all the spam bullet screens, and at the same time obtain the users corresponding to the spam bullet screens and mark them as spam users. Then analyze the spam bullet screens of the spam users. When the number of spam bullet screens generated by the spam user is greater than the preset value, the system automatically filters the bullet screens of the spam user, and sorts and stores the filtered spam bullet screens. When the number of spam bullet screens generated by the spam user is less than the preset value, no processing is done. Then sort the information of the spam bullet screens to generate spam bullet screen information and transmit it to the output end.

8. A video data analysis and processing system, applied to a video data analysis and processing method according to any one of claims 1 to 7, characterized in that: It includes a data acquisition unit, a personnel classification unit, a threshold storage unit, an adaptive processing unit, an information sorting unit, an information filtering unit, a background storage unit and an information output unit.

Images