Campus bullying prevention method and system based on artificial intelligence voice recognition and medium

Abstract

The invention relates to a campus anti-bullying method and system based on artificial intelligence voice recognition and a medium, and relates to the technical field of artificial intelligence, and the method comprises the steps: obtaining to-be-analyzed voice information in a preset range in a campus; based on the to-be-analyzed voice information, determining at least one corresponding to-be-analyzed keyword, and judging whether a target keyword related to a bullying behavior exists in all the to-be-analyzed keywords or not; if yes, determining whether a bullying occurrence risk exists in the preset range or not based on the target keywords; and when the bullying risk exists in the preset range, determining a target position where bullying occurs in the preset range, and sending bullying early warning information to a preset campus early warning platform based on the target position. The method has the effect.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and particularly relates to a campus anti-bullying method, system and medium based on artificial intelligence speech recognition. Background Art

[0002] Campus bullying, also known as school bullying, is an aggressive behavior that occurs inside and outside the campus with students as the participants. It includes both direct bullying and indirect bullying. Direct bullying refers to bullying in an overt and obvious way. Indirect bullying refers to bullying in a less easily detectable way, usually by means of a third party. In order for students to study and live in a safe campus environment, an effective campus anti-bullying method is urgently needed.

[0003] Currently, the commonly used methods for campus anti-bullying are: installing cameras in various areas of the campus for video monitoring, and at the same time cooperating with regular on-campus manual patrols to prevent the occurrence of campus bullying to a certain extent. However, whether it is video monitoring or manual patrol, once the personnel do not carefully and timely check the surveillance videos, or the bullying behavior avoids the period of manual patrol, it is easy for the bullying behavior to not be discovered and handled in time, resulting in poor timeliness in discovering campus bullying behavior. Summary of the Invention

[0004] In order to improve the timeliness of discovering campus bullying behavior, this application provides a campus anti-bullying method, system, storage medium and electronic device based on artificial intelligence speech recognition.

[0005] In the first aspect of this application, a campus anti-bullying method based on artificial intelligence speech recognition is provided, which specifically includes: Obtain the voice information to be analyzed within a preset range on the campus; Based on the voice information to be analyzed, determine at least one corresponding keyword to be analyzed, and determine whether there is a target keyword related to bullying behavior among all the keywords to be analyzed; If so, based on each of the target keywords, determine whether there is a risk of bullying occurring within the preset range; When there is a risk of bullying occurring within the preset range, determine the target location where bullying occurs from within the preset range, and based on the target location, send a bullying warning message to a preset campus warning platform.

[0006] By adopting the above technical solution, after obtaining the voice information to be analyzed within the preset range, the keywords to be analyzed included in the voice information to be analyzed are extracted. Then, it is determined whether there are target keywords among the keywords to be analyzed. If there are, it indicates that there are keywords related to bullying in the speech of personnel within the preset range, and there may be bullying behavior. Then, a comprehensive analysis is further performed on the existing target keywords to determine whether there is a risk of bullying occurring within the preset range. If there is a risk of bullying occurring, it indicates that there is probably bullying behavior within the preset range. Then, the specific location where bullying occurs, that is, the target location, is determined within the preset range, and for this target location, a bullying warning message is sent to the campus warning platform, so as to facilitate arranging relevant personnel to go to check in time, and further improve the timeliness of discovering campus bullying behavior.

[0007] Optionally, determining whether there are target keywords related to bullying behavior among all the keywords to be analyzed specifically includes: Obtain each historical bullying keyword that appears in historical bullying events on campus, count the first occurrence times of each historical bullying keyword, and select the first number of historical bullying keywords from each historical bullying keyword in descending order of the first occurrence times to be determined as reference keywords; Obtain the historical bullying location where bullying behavior occurs when a single reference keyword appears in historical bullying events, count the second occurrence times of each historical bullying location, select the second number of historical bullying locations from each historical bullying location in descending order of the second occurrence times, and determine them as the reference bullying locations corresponding to a single reference keyword; Determine the first weight of each reference keyword and the second weight of the reference bullying location corresponding to each reference keyword. The first weight is the ratio of the first occurrence times of each reference keyword to the sum of the first occurrence times of all reference keywords, and the second weight is the ratio of the second occurrence times of the single reference bullying location corresponding to the reference keyword to the sum of the second occurrence times of all corresponding reference bullying locations; Based on the first weight and the corresponding second weight, determine whether there are target keywords related to bullying behavior among all the keywords to be analyzed.

[0008] By adopting the above technical solution, the larger the first occurrence frequency is, the easier it is for the corresponding historical bullying keywords to appear when school bullying occurs, and then the reference keywords are determined; the larger the second occurrence frequency is, the easier it is for the reference keyword to appear when the bullying behavior occurs at the corresponding historical bullying position, and then the reference bullying position is determined. Finally, by combining the first weight of the reference keyword and the second weight of the corresponding reference bullying position, the possibility of each reference keyword appearing in the bullying behavior is analyzed and determined, and based on this, it is more accurate to judge whether there are target keywords related to bullying behavior in the keywords to be analyzed.

[0009] Optionally, the judging whether there are target keywords related to bullying behavior in all the keywords to be analyzed based on the first weight and the corresponding second weight specifically includes: Calculate the first product of the first weight of each reference keyword and the second weight of each corresponding reference bullying position; Sum up the first products to obtain the sum of the first products of the corresponding reference keyword; According to the sum of the first products, determine the comparison order of the corresponding reference keywords, and according to each comparison order, perform semantic comparison between each keyword to be analyzed and the corresponding reference keyword. The larger the sum of the first products is, the higher the corresponding comparison order is; If the keyword to be analyzed is semantically consistent with the corresponding reference keyword, it is determined that there are target keywords related to bullying behavior, and the corresponding keyword to be analyzed is determined as the target keyword.

[0010] By adopting the above technical solution, the larger the first product is, the greater the possibility of bullying behavior when the reference keyword appears at the corresponding reference bullying position. Sum up each first product to obtain the sum of the first products corresponding to the reference keyword. The larger the sum of the first products is, the greater the overall possibility of bullying behavior when the corresponding reference keyword appears. Further, according to the sum of the first products, determine the comparison order of the corresponding reference keyword. The larger the sum of the first products is, the higher the corresponding comparison order is, and the more preferentially the keyword to be analyzed is compared with the reference keyword, so as to efficiently and accurately identify the target keyword and facilitate subsequent accurate analysis of the risk of bullying behavior.

[0011] Optionally, the determining whether there is a risk of bullying occurrence within the preset range based on each target keyword specifically includes: Determine the reference keyword that is semantically consistent with the target keyword as the important keyword. If the reference bullying position exists within the preset range, determine the corresponding reference bullying position as the key bullying position, and when the key bullying position is included in each reference bullying position corresponding to the important keyword, determine the corresponding important keyword as the key keyword; Calculate the second product of the first weight of each of the key keywords and the second weight of the corresponding key bullying position, and sum up the second products to obtain the sum of the second products; If the sum of the second products exceeds a preset first threshold, it is determined that there is a risk of bullying occurrence within the preset range.

[0012] By adopting the above technical solution, the larger the second product, the easier it is for bullying behavior to occur at the corresponding key bullying position within the preset range when the key keyword appears currently. Summing up each second product to obtain the sum of the second products, the larger the sum of the second products, the greater the overall possibility of bullying occurrence currently within the preset range. If the sum of the second products exceeds the preset first threshold, it indicates that the overall possibility of bullying occurrence within the preset range is relatively large. Then, it is determined that there is a risk of bullying occurrence within the preset range and timely warning is required.

[0013] Optionally, determining the target position where bullying occurs within the preset range specifically includes: Sum up the second products corresponding to the same key bullying position to obtain the corresponding sum of the third products; Compare the sum of the third products with a preset second threshold. If the sum of the third products exceeds the second threshold, determine the corresponding key bullying position as the target position where bullying occurs; Based on the target position, sending a bullying warning message to a preset campus warning platform specifically includes: Determine the inspection order of the corresponding target position according to the sum of the third products, and send a bullying warning message to a preset campus warning platform for the target position and the corresponding inspection order. The larger the sum of the third products, the higher the corresponding inspection order.

[0014] By adopting the above technical solution, if the sum of the third products exceeds the preset second threshold, it indicates that the possibility of bullying occurrence at the corresponding key bullying position is relatively large. Then, determine the corresponding key bullying position as the target position where bullying occurs. Finally, send a bullying warning message to the campus warning platform for this target position, so as to facilitate notifying the inspection personnel to go to the corresponding position in time for inspection and handling.

[0015] Optionally, the method further includes: Obtain the historical time periods when historical bullying incidents occurred on campus, count the first occurrence frequencies of each historical time period, and select a third number of historical time periods from each historical time period in descending order of the first occurrence frequencies as the target time periods; Obtain the bullying locations of historical bullying incidents that occurred within a single target period, count the second occurrence frequency of each bullying location, select the fourth number of bullying locations from each bullying location in descending order of the second occurrence frequency, and determine them as the target bullying locations corresponding to a single target period; Determine the third weight of each target period and the fourth weight of the target bullying location corresponding to each target period. The third weight is the ratio of the first occurrence frequency of each target period to the sum of the first occurrence frequencies of all target periods, and the fourth weight is the ratio of the second occurrence frequency of a single target bullying location corresponding to the target period to the sum of the second occurrence frequencies of all corresponding target bullying locations; Determine the appropriate path for the campus patrol personnel to go to the target location according to the third weight and the corresponding fourth weight.

[0016] By adopting the above technical solution, the larger the first occurrence frequency, the more likely bullying is to occur within the corresponding historical period, and then the target period is determined; the larger the second occurrence frequency, the more likely bullying is to occur at the corresponding bullying location within the target period, and then the target bullying location is determined. Finally, by combining the third weight of the target period and the fourth weight of the corresponding target bullying location, analyze and determine the possibility of bullying occurring at each location on the way for the campus patrol personnel to go to the target location, and then determine a more appropriate path to the target location.

[0017] Optionally, the determining the appropriate path for the campus patrol personnel to go to the target location according to the third weight and the corresponding fourth weight specifically includes: Determine at least one patrol path according to the target location and the location of the campus patrol personnel, and the moving duration of the patrol path is less than a preset duration threshold; Determine the patrol period of the corresponding patrol path according to the moving duration and the current time, and determine the target period within a single patrol period as an important period, and determine the bullying blind spot location involved in the corresponding patrol path as a key blind spot location; If there is a key blind spot location among the target bullying locations corresponding to the important period, determine the corresponding important period as a key period, and calculate the third product of the third weight of each key period and the fourth weight of the corresponding key blind spot location; Sum up the third products to obtain the sum of the fourth products corresponding to the corresponding patrol path, select the largest sum of the fourth products from the sums of the fourth products, and determine the patrol path corresponding to the largest sum of the fourth products as the appropriate path for the campus patrol personnel to go to the target location.

[0018] By adopting the above technical solution, the larger the third product is, the greater the possibility of bullying occurring at the corresponding key blind spot position during the key period. Summing up each third product to obtain the sum of the fourth products of the corresponding patrol path, the greater the sum of the fourth products, the greater the overall possibility of bullying occurring in the corresponding patrol path. Finally, select the largest sum of the fourth products from each sum of the fourth products, and determine the patrol path corresponding to this largest sum of the fourth products as the appropriate path for the campus patrol personnel to go to the target location to check. Subsequently, the campus patrol personnel go to the target location along this appropriate path, and while checking the bullying situation at the target location, they can also timely discover the bullying behavior at the blind spot position on the way, improving the prevention effect of bullying behavior on campus.

[0019] In the second aspect of the present application, a campus anti-bullying system based on artificial intelligence speech recognition is provided, specifically including: A voice acquisition module, configured to acquire the voice information to be analyzed within a preset range on campus; A voice analysis module, configured to determine at least one keyword to be analyzed corresponding thereto based on the voice information to be analyzed, and determine whether there is a target keyword related to bullying behavior among all the keywords to be analyzed; A risk assessment module, configured to, if so, determine whether there is a risk of bullying occurring within the preset range based on each of the target keywords; A bullying warning module, configured to, when there is a risk of bullying occurring within the preset range, determine the target location where bullying occurs from within the preset range, and send a bullying warning message to a preset campus warning platform based on the target location.

[0020] By adopting the above technical solution, the voice acquisition module acquires the voice information to be analyzed within the preset range, then the voice analysis module determines whether there is a target keyword related to bullying behavior among all the keywords to be analyzed, then the risk assessment module determines whether there is a risk of bullying occurring within the preset range when there is a target keyword related to bullying behavior, and finally the bullying warning module determines the target location where bullying occurs from within the preset range when there is a risk of bullying occurring within the preset range, and sends a bullying warning message to a preset campus warning platform based on the target location.

[0021] In the third aspect of the present application, a computer-readable storage medium is provided, in which a computer program is stored. When the computer program is loaded and executed by a processor, the method steps described in any one of the first aspects are executed.

[0022] In the fourth aspect of the present application, an electronic device is provided, specifically including: A processor, a memory, and a computer program stored in the memory and capable of running on the processor, the processor being configured to load and execute the computer program stored in the memory, so that the electronic device executes the method described in any one of the first aspect.

[0023] In summary, the present application includes at least one of the following beneficial technical effects: After obtaining the voice information to be analyzed within a preset range, extract the keywords to be analyzed included in the voice information to be analyzed, and then determine whether there is a target keyword among the keywords to be analyzed. If there is, it means that there are keywords related to bullying in the speech of a person within the preset range, and there may be a bullying behavior. Then, further comprehensively analyze the existing target keywords to determine whether there is a risk of bullying occurring within the preset range. If there is a risk of bullying occurring, it means that there is probably a bullying behavior within the preset range. Then, determine the specific location where the bullying occurs, that is, the target location, within the preset range, and send a bullying warning message to the campus warning platform for this target location, so as to facilitate arranging relevant personnel to go and check in time, and further improve the timeliness of discovering campus bullying behavior. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a flowchart of a campus anti-bullying method based on artificial intelligence voice recognition provided by an embodiment of the present application; Figure 2 is a structural diagram of a campus anti-bullying system based on artificial intelligence voice recognition provided by an embodiment of the present application; Figure 3 is a structural diagram of another campus anti-bullying system based on artificial intelligence voice recognition provided by an embodiment of the present application.

[0025] Description of the reference numerals: 11, voice acquisition module; 12, voice analysis module; 13, risk assessment module; 14, bullying warning module; 15, path determination module. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] In order to enable those skilled in the art to better understand the technical solutions in this specification, the following will clearly and completely describe the technical solutions in the embodiments of this specification with reference to the accompanying drawings in the embodiments of this specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments.

[0027] In the description of the embodiments of the present application, words such as "exemplary", "for example", or "for illustration" are used to give examples, illustrations, or explanations. Any embodiment or design solution described as "exemplary", "for example", or "for illustration" in the embodiments of the present application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of words such as "exemplary", "for example", or "for illustration" is intended to present related concepts in a specific manner.

[0028] In the description of the embodiments of the present application, the term "and / or" is merely an association relationship describing associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, B exists alone, and A and B exist simultaneously. Additionally, unless otherwise specified, the meaning of the term "plural" refers to two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. Furthermore, the terms "first" and "second" are only used for descriptive purposes and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. The terms "include", "comprise", "have" and their variants all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0029] See Figure 1 , the embodiments of the present application disclose a schematic flowchart of a campus anti-bullying method based on artificial intelligence speech recognition, which can be implemented depending on a computer program or run on a campus anti-bullying system based on artificial intelligence speech recognition and based on the von Neumann architecture. This computer program can be integrated into an application or run as an independent tool-type application, and specifically includes: S101: Obtain the voice information to be analyzed within a preset range on campus.

[0030] Specifically, the campus anti-bullying method based on artificial intelligence speech recognition disclosed in the embodiments of the present application is applied to a campus anti-bullying early warning terminal. The campus anti-bullying early warning terminal can adopt an emergency alarm pillar, which is built-in with an AI speech recognition module and a microprocessor, and can monitor the voice information in the surrounding environment in real time, accurately and quickly convert human speech into text, and there are limitations in the distance of speech recognition. The microprocessor is used to process and output the data obtained by the campus anti-bullying early warning terminal. Further, at least one campus anti-bullying early warning terminal is set in the campus and distributed in different areas of the campus. The preset range is the maximum distance range for voice monitoring and recognition by the campus anti-bullying early warning terminal. The campus anti-bullying early warning terminal is wirelessly connected to the campus early warning platform, and the campus early warning platform can be a physical server. When the campus early warning platform identifies a campus bullying risk, it sends a warning message to the campus early warning platform, and then the campus early warning platform forwards the warning message to the corresponding patrol personnel, who go to check and handle it. In addition, the campus anti-bullying early warning terminal also has an echo cancellation function and can realize full-duplex calls and visual calls. Finally, a feasible way to obtain the voice information to be analyzed: through the built-in AI speech recognition module, collect and recognize the voice within the preset range of the location of the corresponding campus anti-bullying early warning terminal to obtain the voice information to be analyzed. It should be noted that the execution subject of the campus anti-bullying method based on artificial intelligence speech recognition can also be the campus early warning platform, which receives the keywords related to bullying behavior recognized by the campus anti-bullying early warning terminal, processes the data, determines whether bullying has occurred and the location of the bullying, and sends a warning message to the campus patrol personnel.

[0031] S102: Based on the voice information to be analyzed, determine at least one corresponding keyword to be analyzed, and determine whether there is a target keyword related to bullying behavior among all the keywords to be analyzed.

[0032] Specifically, after the voice information to be analyzed is determined, at least one keyword to be analyzed is extracted from the voice information to be analyzed through Natural Language Processing (NLP) technology. Then, it is judged whether there is a target keyword related to bullying behavior among all the keywords to be analyzed. A feasible judgment method is as follows: based on the historical bullying records, each historical bullying keyword that appears in the historical bullying incidents on campus is obtained. The bullying keyword or the keyword related to bullying behavior can be understood as the keyword contained in the words of the bully when bullying others, or the keyword contained in the words of the bullied when being bullied. The first occurrence times of each historical bullying keyword are counted. The more the first occurrence times are, the more likely the corresponding historical bullying keyword is to appear when campus bullying occurs. The first number of historical bullying keywords is selected from each historical bullying keyword in descending order of the first occurrence times and determined as the reference keyword, that is, the keyword that is likely to appear in bullying. Among them, the historical bullying records include, but are not limited to, the historical bullying keywords, the locations where bullying occurred, and the time periods when bullying occurred in different previous bullying incidents on campus.

[0033] Further, based on the above historical bullying records, the historical bullying location where the bullying behavior occurred when a single reference keyword appeared in the historical bullying incident is obtained. Among them, the historical bullying location is the location where the historical bullying behavior occurred. The second occurrence times of each historical bullying location are counted. The more the second occurrence times are, the more likely the reference keyword is to appear when the bullying behavior occurs at the corresponding historical bullying location. The second number of historical bullying locations is selected from each historical bullying location in descending order of the second occurrence times and determined as the reference bullying location corresponding to the reference keyword, that is, the bullying location where the reference keyword is likely to appear in bullying.

[0034] Further, the first weight of each reference keyword is determined. The first weight is the ratio of the first occurrence times of each reference keyword to the sum of the first occurrence times of all reference keywords. Then, the second weight of the reference bullying location corresponding to each reference keyword is determined. The second weight is the ratio of the second occurrence times of a single reference bullying location corresponding to the reference keyword to the sum of the second occurrence times of all corresponding reference bullying locations.

[0035] Finally, based on the determined first weight and the corresponding second weight, it is determined whether there is a target keyword in the keyword to be analyzed. One feasible implementation method is as follows: calculate the first product of the first weight of each reference keyword and the second weights of the corresponding reference bullying positions. The larger the first product, the greater the possibility of bullying behavior when the reference keyword appears at the corresponding reference bullying position. Sum up the first products to obtain the sum of the first products corresponding to the reference keyword. The larger the sum of the first products, the greater the overall possibility of bullying behavior when the corresponding reference keyword appears. Further, according to the sum of the first products, determine the comparison order of the corresponding reference keyword. The larger the sum of the first products, the higher the corresponding comparison order, and the more preferentially the keyword to be analyzed is compared with the reference keyword. In the embodiments of the present application, it is mainly semantic comparison. In other embodiments, when comparing the keyword to be analyzed with the reference keyword, the correlation degree between the two can also be evaluated. Determine the number of times the keyword to be analyzed and the reference keyword appear simultaneously in each historical bullying event and determine it as the correlation degree. The more times, the greater the correlation degree between the two. If the correlation degree exceeds the preset correlation degree threshold, then determine the corresponding keyword to be analyzed as the target keyword.

[0036] Further, according to each comparison order, sequentially perform semantic comparison between each keyword to be analyzed and the corresponding reference keyword. Specifically, determine the semantic similarity between the keyword to be analyzed and the reference keyword through a preset word embedding model. If the semantic similarity exceeds the preset similarity threshold, then determine that the keyword to be analyzed is semantically consistent with the corresponding reference keyword, determine that there is a target keyword related to bullying behavior, and determine the keyword to be analyzed as the target keyword. Compare the keyword to be analyzed with the reference keyword with a greater possibility of bullying first, so as to efficiently and accurately identify the target keyword, which is convenient for subsequent accurate analysis of the risk of bullying behavior.

[0037] S103: If so, based on each target keyword, determine whether there is a risk of bullying occurring within a preset range.

[0038] Specifically, if it is determined that there are target keywords in the keywords to be analyzed, then according to each existing target keyword, analyze whether there is a risk of bullying occurring within the preset range. One achievable implementation method is as follows: Determine the reference keywords with the same semantics as the target keywords as important keywords. If the reference bullying positions exist within the preset range, then determine the corresponding reference bullying positions as key bullying positions. Further, when such key bullying positions exist among the respective reference bullying positions corresponding to the important keywords, determine the corresponding important keywords as key words. Calculate the second product of the first weight of each key word and the second weight of the corresponding key bullying position. The larger the second product, the easier it is for bullying behavior to occur at the corresponding key bullying position within the preset range when the current key word appears. Sum up the respective second products to obtain the sum of the second products. The larger the sum of the second products, the greater the overall possibility of bullying occurring within the preset range at present. If the sum of the second products exceeds the preset first threshold, it indicates that the overall possibility of bullying occurring within the preset range is relatively large. Then determine that there is a risk of bullying occurring within the preset range and timely warning is required.

[0039] S104: When there is a risk of bullying occurring within the preset range, determine the target position where bullying occurs from within the preset range, and based on the target position, send a bullying warning message to the preset campus warning platform.

[0040] Specifically, if it is determined that there is a risk of bullying occurring within the preset range, then it is necessary to determine the target position where bullying is sent from within the preset range. One feasible determination method is as follows: Sum up the respective second products corresponding to the same key bullying position to obtain the corresponding sum of the third products. The larger the sum of the third products, the greater the possibility of bullying occurring at the corresponding key bullying position. If the sum of the third products exceeds the preset second threshold, it indicates that the possibility of bullying occurring at the corresponding key bullying position is relatively large. Then determine the corresponding key bullying position as the target position where bullying occurs. Finally, for this target position, send a bullying warning message to the campus warning platform, so as to facilitate notifying the patrol personnel to go to the corresponding position in time for inspection and handling. In other embodiments, after the target position is determined, according to the sum of the third products, determine the inspection order of the corresponding target position. The larger the sum of the third products, the higher the corresponding inspection order, and the more priority is given to inspecting the corresponding target position, so as to be able to timely discover the existing bullying behavior and improve the inspection efficiency of bullying behavior.

[0041] In other embodiments, at least one position combination is determined from each of the target positions. The position combination includes at least one target position, and the distances between the respective target positions in the position combination are less than a preset distance threshold. The sums of the third products corresponding to the respective target positions in a single position combination are summed to obtain a summation result. The larger the summation result, the greater the overall likelihood of bullying in the area where the corresponding position combination is located. The largest summation result is selected from each of the summation results, and the area where the position combination corresponding to this largest summation result is located is determined as the key patrol area. Then, based on the position information of each patrol personnel on campus, the patrol personnel closest to this key patrol area is determined and notified to go to the key patrol area for patrol, which can not only improve the timeliness of viewing bullying but also improve the patrol efficiency for bullying.

[0042] In yet another embodiment, based on the above historical bullying records, the historical time periods in which historical bullying incidents occurred on campus are obtained, and the first occurrence frequency of each historical time period is counted. The larger the first occurrence frequency, the more likely bullying is to occur during the corresponding historical time period. In the order from largest to smallest of the first occurrence frequencies, the third number of historical time periods is selected from each of the historical time periods and determined as the target time periods, that is, the time periods when bullying is likely to occur on campus. Then, from the above historical bullying records, the bullying positions of the historical bullying incidents that occurred within a single target time period are obtained, and the second occurrence frequency of each bullying position is counted. The larger the second occurrence frequency, the more likely the corresponding bullying position is to occur during that target time period. Then, in the order from largest to smallest of the second occurrence frequencies, the fourth number of bullying positions is selected from each of the bullying positions and determined as the target bullying positions corresponding to a single target time period, that is, the positions where bullying is likely to occur.

[0043] Furthermore, the third weight of each target time period is determined. The third weight is the ratio of the first occurrence frequency of each target time period to the sum of the first occurrence frequencies of all target time periods. Then, the fourth weight of the target bullying position corresponding to the target time period is determined. The fourth weight is the ratio of the second occurrence frequency of the single target bullying position corresponding to the target time period to the sum of the second occurrence frequencies of all the target bullying positions corresponding to that target time period.

[0044] Finally, according to the determined third weight and the corresponding fourth weight, determine the appropriate path for the campus patrol personnel to go to the target location. A feasible determination method is as follows: Determine the location of the campus patrol personnel according to the location information sent by the terminal of the campus patrol personnel, and determine at least one patrol path from the location of the campus patrol personnel to the target location through a preset navigation tool. The moving duration of the patrol path is less than a preset duration threshold, that is, the time taken for the campus patrol personnel to go to the target location is shorter, so that they can go to the target location in time to conduct a patrol against bullying. Further, according to the moving duration and the current time, determine the patrol time period corresponding to a single patrol path. For a single patrol time period, determine the target time period within the patrol time period as the important time period, and determine the bullying blind spot location involved in the corresponding patrol path as the key blind spot location, that is, the area where the campus anti-bullying warning terminal or camera on campus cannot monitor bullying. If there is a key blind spot location among the various target bullying locations corresponding to the important time period, then determine the corresponding important time period as the key time period, and then calculate the third product of the third weight of each key time period and the fourth weight of the corresponding key blind spot location. The larger the third product, the greater the possibility of bullying occurring at the corresponding key blind spot location during the key time period. Sum up the various third products to obtain the sum of the fourth products corresponding to the corresponding patrol path. The larger the sum of the fourth products, the greater the overall possibility of bullying occurring in the corresponding patrol path. Finally, select the largest sum of the fourth products from the various sums of the fourth products, and determine the patrol path corresponding to this largest sum of the fourth products as the appropriate path for the campus patrol personnel to go to the target location to check. Subsequently, the campus patrol personnel go to the target location along this appropriate path, and while checking the bullying situation at the target location, they can also timely discover bullying behaviors at the blind spot location on the way, improving the prevention effect of bullying behaviors on campus.

[0045] In one embodiment, if the campus patrol personnel do not find any bullying behavior after arriving at the target location, and no keywords related to bullying behavior are recognized by other campus anti-bullying warning terminals on campus, then the preset ranges of the campus anti-bullying warning terminal corresponding to the target location and the preset ranges of other adjacent campus anti-bullying warning terminals are used to determine at least one warning blind area of the warning terminal. Further, for a single warning blind area, if the reference bullying location corresponding to the important keyword exists in this warning blind area, then the corresponding important keyword is determined as the final keyword, and the reference bullying location existing in this warning blind area is determined as the final bullying location. Calculate the weight product of the first weight of each important keyword and the second weight of the corresponding final bullying location, and finally sum up the respective weight products to obtain the sum of the weight products corresponding to this warning blind area. The larger the sum of the weight products, the greater the possibility that the bullying behavior originally present at the target location has transferred to the corresponding warning blind area. Further, select the largest sum of the weight products from the sums of the weight products, and determine the warning blind area corresponding to this largest sum of the weight products as the secondary investigation area, so as to facilitate the campus patrol personnel to check and investigate bullying again, and improve the investigation effect of bullying.

[0046] The implementation principle of the campus anti-bullying method based on artificial intelligence speech recognition in the embodiments of this application is as follows: After obtaining the voice information to be analyzed within the preset range, extract the keywords to be analyzed included in the voice information to be analyzed, and then determine whether there are target keywords in the keywords to be analyzed. If so, it means that there are keywords related to bullying in the speech of the personnel within the preset range, and there may be a bullying behavior. Then, further comprehensively analyze the existing target keywords to determine whether there is a risk of bullying occurrence within the preset range. If there is a risk of bullying occurrence, it means that there is probably a bullying behavior within the preset range. Then, determine the specific location where the bullying occurs, that is, the target location, within the preset range, and send a bullying warning message to the campus warning platform for this target location, so as to facilitate arranging relevant personnel to go to check in time, and thus improve the timeliness of discovering campus bullying behavior.

[0047] The following is an embodiment of the system of this application, which can be used to execute the embodiment of the method of this application. For the details not disclosed in the embodiment of the system of this application, please refer to the embodiment of the method of this application.

[0048] Please refer to Figure 2 , which is a schematic structural diagram of the campus anti-bullying system based on artificial intelligence speech recognition provided by the embodiment of this application. The campus anti-bullying system based on artificial intelligence speech recognition can be implemented as all or part of the system through software, hardware, or a combination of both. The system includes a voice collection module 11, a voice analysis module 12, a risk assessment module 13, and a bullying warning module 14.

[0049] A voice acquisition module 11, configured to acquire voice information to be analyzed within a preset range on campus; A voice analysis module 12, configured to determine at least one keyword to be analyzed corresponding to the voice information to be analyzed, and determine whether there is a target keyword related to bullying behavior among all the keywords to be analyzed; A risk assessment module 13, configured to, if so, determine whether there is a risk of bullying occurrence within the preset range based on each target keyword; A bullying warning module 14, configured to, when there is a risk of bullying occurrence within the preset range, determine the target location where bullying occurs within the preset range, and send a bullying warning message to a preset campus warning platform based on the target location.

[0050] Optionally, the voice analysis module 12 is specifically configured to: Obtain each historical bullying keyword that appears in historical bullying events on campus, count the first occurrence times of each historical bullying keyword, and select the first number of historical bullying keywords from each historical bullying keyword in descending order of the first occurrence times to be determined as reference keywords; Obtain the historical bullying locations where bullying behaviors occur when a single reference keyword appears in historical bullying events, count the second occurrence times of each historical bullying location, select the second number of historical bullying locations from each historical bullying location in descending order of the second occurrence times, and determine them as the reference bullying locations corresponding to the single reference keyword; Determine the first weight of each reference keyword and the second weight of the reference bullying location corresponding to each reference keyword. The first weight is the ratio of the first occurrence times of each reference keyword to the sum of the first occurrence times of all reference keywords, and the second weight is the ratio of the second occurrence times of the single reference bullying location corresponding to the reference keyword to the sum of the second occurrence times of all corresponding reference bullying locations; Based on the first weight and the corresponding second weight, determine whether there is a target keyword related to bullying behavior among all the keywords to be analyzed.

[0051] Optionally, the voice analysis module 12 is specifically configured to: Calculate the first product of the first weight of each reference keyword and the second weight of the corresponding reference bullying location; Sum up each first product to obtain the sum of the first products corresponding to the reference keywords; Determine the comparison order corresponding to the reference keywords according to the sum of the first products, and compare the semantics of each keyword to be analyzed with the corresponding reference keyword according to each comparison order. The larger the sum of the first products, the higher the corresponding comparison order; If the keyword to be analyzed is semantically consistent with the corresponding reference keyword, then determine that there is a target keyword related to bullying behavior, and determine the corresponding keyword to be analyzed as the target keyword.

[0052] Optionally, the risk assessment module 13 is specifically configured to: Determine the reference keyword that is semantically consistent with the target keyword as the important keyword. If the reference bullying location exists within the preset range, then determine the corresponding reference bullying location as the key bullying location, and when the key bullying location is included in each reference bullying location corresponding to the important keyword, determine the corresponding important keyword as the key keyword; Calculate the second product of the first weight of each key keyword and the second weight of the corresponding key bullying location, and sum up the second products to obtain the sum of the second products; If the sum of the second products exceeds the preset first threshold, then determine that there is a risk of bullying occurring within the preset range.

[0053] Optionally, the bullying warning module 14 is specifically configured to: Sum up the second products corresponding to the same key bullying location to obtain the corresponding sum of the third products; Compare the sum of the third products with the preset second threshold. If the sum of the third products exceeds the second threshold, then determine the corresponding key bullying location as the target location where bullying occurs.

[0054] Optionally, the bullying warning module 14 is specifically configured to: Determine the inspection order of the corresponding target location according to the sum of the third products. For the target location and the corresponding inspection order, send a bullying warning message to the preset campus warning platform. The larger the sum of the third products, the higher the corresponding inspection order.

[0055] Optionally, as Figure 3 shown, the system further includes a path determination module 15, which is specifically configured to: Obtain the historical time period when the historical bullying event occurred on campus, count the first occurrence frequency of each historical time period, and select the third number of historical time periods from each historical time period in descending order of the first occurrence frequency as the target time period; Obtain the bullying locations of the historical bullying events that occurred within a single target time period, count the second occurrence frequency of each bullying location, select the fourth number of bullying locations from each bullying location in descending order of the second occurrence frequency, and determine them as the target bullying locations corresponding to a single target time period; Determine the third weight for each target time period, and determine the fourth weight for the target bullying location corresponding to each target time period. The third weight is the ratio of the first occurrence frequency of each target time period to the sum of the first occurrence frequencies of all target time periods, and the fourth weight is the ratio of the second occurrence frequency of a single target bullying location corresponding to the target time period to the sum of the second occurrence frequencies of all corresponding target bullying locations. Determine the appropriate path for the campus patrol personnel to reach the target location according to the third weight and the corresponding fourth weight.

[0056] Optionally, the path determination module 15 is specifically configured to: Determine at least one patrol path according to the target location and the location of the campus patrol personnel, and the moving duration of the patrol path is less than a preset duration threshold; Determine the patrol time period of the corresponding patrol path according to the moving duration and the current time, and determine the target time period within a single patrol time period as an important time period, and determine the bullying blind area location involved in the corresponding patrol path as a key blind area location; If there is a key blind area location among the target bullying locations corresponding to the important time period, then determine the corresponding important time period as a key time period, and calculate the third product of the third weight of each key time period and the fourth weight of the corresponding key blind area location; Sum up the third products to obtain the sum of the fourth products corresponding to the corresponding patrol path, select the largest sum of the fourth products from the sums of the fourth products, and determine the patrol path corresponding to the largest sum of the fourth products as the appropriate path for the campus patrol personnel to reach the target location.

[0057] It should be noted that when the campus anti-bullying system based on artificial intelligence voice recognition provided in the above embodiment executes the campus anti-bullying method based on artificial intelligence voice recognition, only the above-mentioned division of each functional module is used for illustration. In practical applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the campus anti-bullying system based on artificial intelligence voice recognition provided in the above embodiment and the embodiment of the campus anti-bullying method based on artificial intelligence voice recognition belong to the same concept, and the implementation process thereof is detailed in the method embodiment, which will not be repeated here.

[0058] The embodiment of the present application also discloses a computer-readable storage medium, and the computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, it adopts the campus anti-bullying method based on artificial intelligence voice recognition in the above embodiment.

[0059] Among them, the computer program can be stored in a computer-readable medium. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some middleware form, etc. The computer-readable medium includes any entity or device, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the computer-readable medium includes but is not limited to the above components.

[0060] Among them, through this computer-readable storage medium, a campus anti-bullying method based on artificial intelligence voice recognition in the above embodiment is stored in the computer-readable storage medium, and is loaded and executed on the processor to facilitate the storage and application of the above method.

[0061] The embodiment of the present application also discloses an electronic device. When a computer program stored in the computer-readable storage medium is loaded and executed by the processor, the above-mentioned campus anti-bullying method based on artificial intelligence voice recognition is adopted.

[0062] Among them, the electronic device can be a desktop computer, a laptop computer, or a cloud server, etc. And the electronic device includes but is not limited to a processor and a memory. For example, the electronic device can also include input / output devices, network access devices, and a bus, etc.

[0063] Among them, the processor can adopt a central processing unit (CPU). Of course, according to the actual usage situation, other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. can also be adopted. The general-purpose processor can adopt a microprocessor or any conventional processor, etc. The present application does not make any restrictions on this.

[0064] Among them, the memory can be an internal storage unit of the electronic device, such as the hard disk or memory of the electronic device, or an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), or flash card (FC) equipped on the electronic device. And the memory can also be a combination of the internal storage unit and the external storage device of the electronic device. The memory is used to store the computer program and other programs and data required by the electronic device. The memory can also be used to temporarily store the data that has been output or will be output. The present application does not make any restrictions on this.

[0065] Among them, through this electronic device, a campus anti-bullying method based on artificial intelligence speech recognition in the above embodiment is stored in the memory of the electronic device and is loaded and executed on the processor of the electronic device for convenient use.

[0066] The above are only exemplary embodiments of the present disclosure, and the scope of the present disclosure cannot be limited thereby. That is, all equivalent changes and modifications made according to the teachings of the present disclosure still fall within the scope covered by the present disclosure. This application aims to cover any variations, uses, or adaptive changes of the present disclosure, which follow the general principles of the present disclosure and include common general knowledge or conventional technical means in the technical field not recorded in the present disclosure. The description and the embodiments are only regarded as exemplary, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A campus anti-bullying method based on artificial intelligence speech recognition, characterized in that, The method includes: Obtaining the voice information to be analyzed within a preset range on campus; Based on the voice information to be analyzed, determining at least one keyword to be analyzed, and judging whether there is a target keyword related to bullying behavior among all the keywords to be analyzed; If so, based on each of the target keywords, determining whether there is a risk of bullying occurrence within the preset range; When there is a risk of bullying occurrence within the preset range, determining the target location where bullying occurs within the preset range, and based on the target location, sending a bullying warning message to a preset campus warning platform.

2. The campus anti-bullying method based on artificial intelligence speech recognition according to claim 1, wherein The judging whether there is a target keyword related to bullying behavior among all the keywords to be analyzed specifically includes: Obtaining each historical bullying keyword that appears in historical bullying events on campus, counting the first occurrence times of each historical bullying keyword, and selecting the first number of historical bullying keywords from each of the historical bullying keywords in descending order of the first occurrence times to be determined as reference keywords; Obtaining the historical bullying location where bullying behavior occurs when a single reference keyword appears in a historical bullying event, counting the second occurrence times of each historical bullying location, selecting the second number of historical bullying locations from each of the historical bullying locations in descending order of the second occurrence times, and determining them as the reference bullying locations corresponding to a single reference keyword; Determining the first weight of each reference keyword and the second weight of the reference bullying location corresponding to each reference keyword, where the first weight is the ratio of the first occurrence times of each reference keyword to the sum of the first occurrence times of all reference keywords, and the second weight is the ratio of the second occurrence times of a single reference bullying location corresponding to a reference keyword to the sum of the second occurrence times of all corresponding reference bullying locations; Based on the first weight and the corresponding second weight, judging whether there is a target keyword related to bullying behavior among all the keywords to be analyzed.

3. The campus anti-bullying method based on artificial intelligence speech recognition according to claim 2, characterized in that, The judging whether there is a target keyword related to bullying behavior among all the keywords to be analyzed based on the first weight and the corresponding second weight specifically includes: Calculating the first product of the first weight of each reference keyword and the second weight of the corresponding reference bullying locations; Summing up each of the first products to obtain the sum of the first products of the corresponding reference keywords; According to the sum of the first products, determining the comparison order of the corresponding reference keywords, and comparing the semantics of each keyword to be analyzed with the corresponding reference keyword in accordance with each comparison order, the larger the sum of the first products, the higher the corresponding comparison order; If the keyword to be analyzed is semantically consistent with the corresponding reference keyword, determining that there is a target keyword related to bullying behavior, and determining the corresponding keyword to be analyzed as the target keyword.

4. The campus anti-bullying method based on artificial intelligence speech recognition according to claim 2, wherein, The determining whether there is a risk of bullying occurrence within the preset range based on each of the target keywords specifically includes: Determine the reference keywords that are semantically consistent with the target keyword as important keywords. If the reference bullying location exists within the preset range, determine the corresponding reference bullying location as the key bullying location, and when the key bullying location is included in each reference bullying location corresponding to the important keyword, determine the corresponding important keyword as the key keyword; Calculate the second product of the first weight of each key keyword and the second weight of the corresponding key bullying location, and sum up the second products to obtain the sum of the second products; If the sum of the second products exceeds a preset first threshold, determine that there is a risk of bullying occurrence within the preset range.

5. The campus anti-bullying method based on artificial intelligence speech recognition according to claim 4, characterized in that, The determining the target location where bullying occurs from within the preset range specifically includes: Sum up the second products corresponding to the same key bullying location to obtain the corresponding sum of the third products; Compare the sum of the third products with a preset second threshold. If the sum of the third products exceeds the second threshold, determine the corresponding key bullying location as the target location where bullying occurs; The sending the bullying warning information to a preset campus warning platform based on the target location specifically includes: Determine the inspection order of the corresponding target location according to the sum of the third products. For the target location and the corresponding inspection order, send the bullying warning information to a preset campus warning platform. The larger the sum of the third products, the higher the corresponding inspection order.

6. The campus anti-bullying method based on artificial intelligence speech recognition according to claim 1, characterized in that, The method further includes: Obtain the historical time periods when historical bullying incidents occurred on campus, count the first occurrence frequency of each historical time period, and select the third number of historical time periods from each historical time period in descending order of the first occurrence frequency as the target time periods; Obtain the bullying locations of the historical bullying incidents that occurred within a single target time period, count the second occurrence frequency of each bullying location, select the fourth number of bullying locations from each bullying location in descending order of the second occurrence frequency, and determine them as the target bullying locations corresponding to a single target time period; Determine the third weight of each target time period and the fourth weight of the target bullying location corresponding to each target time period. The third weight is the ratio of the first occurrence frequency of each target time period to the sum of the first occurrence frequencies of all target time periods, and the fourth weight is the ratio of the second occurrence frequency of the single target bullying location corresponding to the target time period to the sum of the second occurrence frequencies of all corresponding target bullying locations; Determine the appropriate path for the campus inspection personnel to go to the target location according to the third weight and the corresponding fourth weight.

7. The campus anti-bullying method based on artificial intelligence speech recognition according to claim 6, wherein The determining the appropriate path for the campus inspection personnel to go to the target location according to the third weight and the corresponding fourth weight specifically includes: Determine at least one inspection path according to the target location and the location of the campus inspection personnel, and the moving duration of the inspection path is less than a preset duration threshold; Determine the inspection period of the corresponding inspection path according to the moving duration and the current time, determine the target period within a single inspection period as the important period, and determine the bullying blind spot positions involved in the corresponding inspection path as the key blind spot positions; If there is a key blind spot position among the target bullying positions corresponding to the important period, determine the corresponding important period as the key period, and calculate the third product of the third weight of each key period and the fourth weight of the corresponding key blind spot position; Sum up the third products to obtain the sum of the fourth products of the corresponding inspection path, select the maximum sum of the fourth products from the sums of the fourth products, and determine the inspection path corresponding to the maximum sum of the fourth products as the appropriate path for the campus inspection personnel to go to the target position.

8. A campus anti-bullying system based on artificial intelligence speech recognition, characterized in that, Including: A voice acquisition module (11) for acquiring voice information to be analyzed within a preset range on campus; A voice analysis module (12) for determining at least one keyword to be analyzed corresponding to the voice information to be analyzed based on the voice information to be analyzed, and determining whether there is a target keyword related to bullying behavior among all the keywords to be analyzed; A risk assessment module (13) for, if so, determining whether there is a risk of bullying occurrence within the preset range based on the target keywords; A bullying warning module (14) for, when there is a risk of bullying occurrence within the preset range, determining the target position where bullying occurs within the preset range, and sending a bullying warning message to a preset campus warning platform based on the target position.

9. A computer-readable storage medium storing a computer program therein, characterized in that, When the computer program is loaded and executed by the processor, the method described in any one of claims 1-7 is adopted.

10. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that, When the processor loads and executes the computer program, the method described in any one of claims 1-7 is adopted.

Images