Method for automatically playing advertisements

Abstract

The application discloses a method for automatically playing advertisements, comprising: a processing module receiving a video stream, identifying characteristic words of pedestrians according to the video stream to form characteristic word data; matching and calculating the characteristic word data with each advertisement data to obtain a matching score of at least one characteristic word data and the advertisement data, screening out advertisement data corresponding to the maximum value in the matching score of the characteristic word data and the advertisement data and transmitting the advertisement data to a display screen for display. Through automatic matching and calculation of the advertisement data and the characteristic word data, the matching score of the advertisement data to be calculated and the characteristic data is calculated by assigning scores to each characteristic word of the characteristic word data to be calculated and summing up, which accurately reflects the adaptation degree of the advertisement data to be calculated and the pedestrians in the current scene, provides a more accurate basis for subsequent advertisement screening and delivery, thereby screening out the advertisement with the highest matching degree with the pedestrians and transmitting the advertisement to the display screen for playing, realizing accurate delivery of the advertisement and improving the conversion efficiency of the advertisement.

Description

Technical Field

[0001] This invention relates to the field of advertising media technology, and more specifically, to a method for automatically playing advertisements. Background Technology

[0002] Smart poles are now widely used along urban roads, and the display screen is a core component of the smart pole. Its main functions include providing pedestrians with convenient information such as traffic guidance and weather tips, as well as providing channels for outdoor advertising, thus becoming an important carrier for combining urban public space with commercial promotion.

[0003] Currently, smart pole displays typically use a non-discriminatory playback mode when playing advertisements. In this mode, the advertisement content is fixed and loops continuously, and the display screen keeps playing the same advertisements. It is impossible to target different users with advertisements that they are interested in. Therefore, it is essential to transform advertising from a broad-based model to a precise model.

[0004] The prior art (application number: 201210520166.7, application date: 2012.12.06) discloses an advertising delivery system, comprising: an image acquisition device for acquiring images located within a preset range in front of a multimedia playback device; a data processing device for extracting the identity type information of the advertising audience in front of the multimedia playback device from the images acquired by the image acquisition device, and selecting an advertisement that matches the identity type information of the advertising audience; and an advertising presentation device for presenting the advertisement that matches the identity type information to the advertising audience. A feature extraction module extracts the features of the advertising audience from the image when there is an advertising audience in front of the advertising presentation device. The feature extraction module includes: a facial feature recognition module, a clothing feature recognition module, or a body posture feature recognition module. When selecting an advertisement that matches the identity type information of the target audience, the advertising category selection module prioritizes age information and prioritizes gender information. The advertising category selection module can use a two-dimensional table, tags, or playlists, etc. In other words, existing technologies disclose the extraction of facial features, clothing features, or body shape features to identify the identity type of the advertising audience, and then match advertisements based on age or gender according to the identity type to achieve targeted advertising. However, the matching method of determining the advertisement category by looking up the extracted age and gender of the advertising audience in the existing technology is rather vague, and it cannot automatically calculate the advertisement that best matches the current audience's needs, resulting in limited accuracy and targeting of the matched advertisements.

[0005] Therefore, how to break through the fuzzy matching method between advertisements and audiences, automatically calculate the advertisements that best meet the current needs of the audience, and avoid falling into the broad category matching mode is an urgent problem to be solved. Summary of the Invention

[0006] In view of this, the present invention provides a method for automatically playing advertisements, comprising:

[0007] The same road includes at least two smart poles. One smart pole is equipped with a camera and a processing module, and the other smart pole is equipped with a display screen. The camera is coupled to the processing module, and the display screen is coupled to the processing module. The shooting end of the camera faces the side of the smart pole with the processing module that is away from the side of the smart pole with the display screen. The light-emitting surface of the display screen faces the same direction as the shooting end.

[0008] The camera captures a video stream in real time and transmits the video stream to the processing module;

[0009] The processing module pre-stores at least one piece of advertising data, which includes an advertisement, at least two keywords, and a score corresponding to each keyword. The processing module receives the video stream and determines whether a pedestrian has passed by based on the video stream. If a pedestrian is detected, the module identifies the pedestrian's characteristic words to form characteristic word data. The module then matches the characteristic word data with each piece of advertising data to obtain a matching score between at least one piece of characteristic word data and the advertising data. This includes: selecting any piece of advertising data and querying the score of each characteristic word in the characteristic word data based on the selected advertising data; selecting any characteristic word; querying whether there is a keyword in the selected advertising data that is identical to the selected characteristic word; if there is a keyword in the selected advertising data that is identical to the selected characteristic word, the keyword is... The score corresponding to the keyword is used as the score of the selected feature word; if there is no keyword in the selected advertising data that is the same as the selected feature word, the score of the selected feature word is 0; the scores of all the feature words obtained from the selected advertising data are added together to obtain the matching score between the feature word data and the selected advertising data; the advertising data corresponding to the maximum value of the matching score between the feature word data and the advertising data is filtered out; it is determined whether the number of filtered advertising data is greater than 1; if the number of filtered advertising data is less than or equal to 1, the advertising data of the filtered advertising data is transmitted to the display screen; if the number of filtered advertising data is greater than 1, the advertising data of the filtered advertising data is extracted, the playback order of the extracted advertising data is randomly generated, and the extracted advertising data and the playback order are transmitted to the display screen;

[0010] The display screen receives the filtered advertising data and plays the advertisements, or the display screen receives the extracted advertisements and the playback order, and plays the extracted advertisements according to the playback order.

[0011] Optionally, before the camera captures the video stream in real time, the following steps are included:

[0012] The processing module stores preset advertisements, and transmits at least one of the preset advertisements to the display screen; the display screen receives at least one of the preset advertisements and plays it.

[0013] Optionally, when the display screen receives the advertisement from the filtered advertising data, the display screen waits for the preset advertisement that is currently playing to finish before playing the advertisement from the filtered advertising data; or when the display screen receives the extracted advertisement and the playback order, the display screen waits for the preset advertisement that is currently playing to finish before playing the extracted advertisement according to the playback order.

[0014] Optionally, before the camera captures the video stream in real time, the following steps are included:

[0015] The processing module transmits a sleep signal to the display screen; the display screen receives the sleep signal and turns off.

[0016] Optionally, after the processing module receives the video stream, it includes:

[0017] The processing module determines whether the energy-saving mode is enabled. If the energy-saving mode is enabled, it determines whether a pedestrian has passed by based on the video stream. If the energy-saving mode is not enabled, the determination ends.

[0018] Optionally, the smart pole equipped with the processing module is also equipped with a distance sensor, which is coupled to the processing module;

[0019] After determining that a pedestrian has passed by, the process also includes:

[0020] The processing module acquires the measured distance from the distance sensor in real time;

[0021] When the processing module obtains that the measured distance of the distance sensor is less than or equal to a preset distance, the processing module transmits the filtered advertisement data to the display screen, or the processing module transmits the extracted advertisement and the playback order to the display screen.

[0022] Optionally, after the screen displays the filtered advertisement data, or after the screen displays the extracted advertisements in the playback order, the process includes:

[0023] When the processing module obtains that the measured distance of the distance sensor is greater than the energy-saving distance, the processing module transmits a sleep signal to the display screen, and the processing module stops obtaining the measured distance of the distance sensor;

[0024] The display screen receives the sleep signal and turns off.

[0025] Optionally, it also includes:

[0026] The number of keywords in any two sets of advertising data is the same, and the sum of the scores corresponding to the keywords in any two sets of advertising data is equal.

[0027] Compared with existing technologies, the automatic advertising playback method provided by this invention achieves at least the following beneficial effects:

[0028] 1. The method for automatically playing advertisements provided by this invention includes: matching feature word data with each advertisement data to obtain a matching score between at least one feature word data and the advertisement data, including: selecting any advertisement data, querying the score of each feature word in the feature word data based on the selected advertisement data, including: selecting any feature word; querying whether there is a keyword in the selected advertisement data that is the same as the selected feature word; if there is a keyword in the selected advertisement data that is the same as the selected feature word, the score corresponding to the keyword that is the same as the selected feature word is used as the score of the selected feature word; if there is no keyword in the selected advertisement data that is the same as the selected feature word, the score of the selected feature word is 0; summing the scores of all feature words obtained based on the selected advertisement data to obtain the matching score between the feature word data and the selected advertisement data. Through the automated matching calculation of advertisement data and feature word data, when calculating the matching score between any advertisement data and feature word data, the matching score between the advertisement data to be calculated and the feature word data is calculated by assigning a score to each feature word of the feature word data based on the advertisement data to be calculated and then summing the scores, which accurately reflects the degree of suitability between the advertisement data to be calculated and pedestrians in the current scenario, providing a more accurate basis for subsequent advertisement screening and placement.

[0029] 2. The method for automatically playing advertisements provided by this invention includes: filtering out the advertisement data corresponding to the maximum value in the matching score between feature word data and advertisement data, transmitting the advertisements from the filtered advertisement data to a display screen, or extracting the advertisements from the filtered advertisement data, randomly generating the playback order of the extracted advertisements, and transmitting the extracted advertisements and the playback order to the display screen. By filtering out the advertisement data corresponding to the maximum value in the matching score between feature word data and advertisement data, that is, filtering out the advertisements with the highest matching degree with pedestrians and transmitting them to the display screen for playback, the method achieves precise advertising and improves advertising conversion efficiency.

[0030] Of course, any product implementing this invention does not necessarily need to achieve all of the technical effects described above at the same time.

[0031] Other features and advantages of the invention will become clear from the following detailed description of exemplary embodiments of the invention with reference to the accompanying drawings. Attached Figure Description

[0032] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments of the invention and, together with their description, serve to explain the principles of the invention.

[0033] Figure 1 This is a schematic diagram of the structure of two smart poles on the same road.

[0034] Figure 2 This is a flowchart illustrating a method for automatically playing advertisements provided by the present invention.

[0035] Figure 3 This is another flowchart illustrating the automatic advertising playback method provided by the present invention.

[0036] Figure 4 This is another flowchart illustrating the automatic advertising playback method provided by the present invention.

[0037] Figure 5 This is another flowchart illustrating the automatic advertising playback method provided by the present invention.

[0038] In the diagram: 1. Smart pole; 2. Camera; 3. Processing module; 4. Display screen. Detailed Implementation

[0039] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the invention.

[0040] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the invention or its application or use.

[0041] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0042] In all the examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0043] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0044] Example 1

[0045] Reference Figure 1 , Figure 2 and Figure 3 , Figure 1 This is a schematic diagram of the structure of two smart poles on the same road. Figure 2 This is a flowchart illustrating a method for automatically playing advertisements provided by the present invention. Figure 3 This is another flowchart illustrating the automatic advertising playback method provided by the present invention, to demonstrate a specific embodiment of the automatic advertising playback method provided by the present invention, including:

[0046] The same road includes at least two smart poles 1. One smart pole 1 is equipped with a camera 2 and a processing module 3, and another smart pole 1 is equipped with a display screen 4. The camera 2 is coupled to the processing module 3, and the display screen 4 is coupled to the processing module 3. The shooting end of the camera 2 faces the side of the smart pole 1 equipped with the processing module 3 away from the side of the smart pole 1 equipped with the display screen 4. The light-emitting surface of the display screen 4 faces the same direction as the shooting end.

[0047] S101: Camera 2 captures video stream in real time and transmits the video stream to processing module 3;

[0048] Processing module 3 has at least one set of advertising data stored in it. The advertising data includes the advertisement, at least two keywords, and the score corresponding to each keyword.

[0049] S102: Processing module 3 receives the video stream and determines whether a pedestrian has passed by based on the video stream;

[0050] S1021: If it is determined that a pedestrian has passed by, the characteristic words of the pedestrian are used to form characteristic word data; the characteristic word data is matched and calculated with each advertisement data to obtain a matching score between at least one characteristic word data and the advertisement data, including:

[0051] S10211: Select any ad data, and query the score of each feature word in the feature word data based on the selected ad data, including:

[0052] S102111: Select any feature word;

[0053] S102112: Query whether there are keywords in the selected advertising data that are the same as the selected feature words. If there are keywords in the selected advertising data that are the same as the selected feature words, the score corresponding to the keyword that is the same as the selected feature words will be used as the score of the selected feature words; if there are no keywords in the selected advertising data that are the same as the selected feature words, the score of the selected feature words will be 0.

[0054] S10212: Sum the scores of all feature words obtained from the selected advertising data to get the matching score between the feature word data and the selected advertising data;

[0055] S1022: Filter out the advertising data corresponding to the maximum value of the matching score between the feature word data and the advertising data; determine whether the number of filtered advertising data is greater than 1. If the number of filtered advertising data is less than or equal to 1, transmit the advertising data of the filtered advertising data to the display screen 4; if the number of filtered advertising data is greater than 1, extract the advertising data of the filtered advertising data, randomly generate the playback order of the extracted advertising data, and transmit the extracted advertising data and the playback order to the display screen 4.

[0056] S103: Display screen 4 receives and plays the filtered advertisements, or display screen 4 receives the extracted advertisements and the playback order, and plays the extracted advertisements according to the playback order.

[0057] It should be noted that, Figure 1 The distance between the two smart poles 1 is for illustrative purposes only. In actual applications, the actual distance between the two smart poles 1 can be flexibly set according to the actual size of the display screen 4 and actual needs. This embodiment does not impose specific limitations on this. Figure 1In this example, to easily indicate the relative positions of two smart poles 1 on the same road, the light-emitting surface of the display screen 4 and the camera end face different directions. In practical applications, the light-emitting surface of the display screen 4 should face the same direction as the camera end, so that pedestrians can see the light-emitting surface of the display screen 4 after being perceived by the smart pole 1 equipped with the camera 2, and watch the advertisement played on the display screen 4. The processing module 3 can call a multimodal large model, such as GPT-4o, Gemini, Doubao, or Qwen-VL. The multimodal large model has preset prompt words. The video stream is input into the multimodal large model, and the multimodal large model automatically outputs the feature words corresponding to the video stream. The processing module 3 is not limited to using a multimodal large model to extract feature words from the video stream. The processing module 3 can also be set to call a deep learning algorithm. The processing module 3 uses a deep learning algorithm to extract features from the video stream. The type and number of deep learning algorithms used by the processing module 3 can be selected according to needs. The number of feature words extracted can increase with the increase of deep learning algorithms. This embodiment does not improve the feature extraction algorithm and directly uses the deep learning algorithm provided by the manufacturer. Since video streams include image frames and audio, the selected deep learning algorithm performs feature extraction on either the image frames or the audio. For example, the deep learning algorithm in processing module 3 can use the VGGNet deep convolutional neural network model. This model first extracts features from the face image, then extracts high-dimensional feature vectors through convolutional and pooling layers. Fully connected layers and the Softax activation function are then used to classify the feature vectors by gender and age group, thus obtaining feature words representing gender and age group. Alternatively, the deep learning algorithm in processing module 3 can also use the YOLO model with multi-objective tracking. The MOT (Motion of Tracking) model detects pedestrian targets using the YOLO model and then uses a multi-target tracking model to lock onto the target in the time dimension, thereby continuously tracking the pedestrian's movement trajectory. This is used to classify the pedestrian's movement trajectory as walking, running, or cycling, thus obtaining feature words representing the movement state. The deep learning algorithm in processing module 3 can also employ an Automatic Emission Recognition (AER) model combined with Natural Language Processing (NLP). The AER model converts audio into text, and NLP technology extracts feature words representing the events the pedestrian is interested in from the text. The preferred NLP model is the BERT series. Specifically, when identifying pedestrian feature words based on the video stream, processing module 3 first determines whether the audio of the video stream is available. If available, processing module 3 performs feature extraction on both image frames and audio; if unavailable, processing module 3 only extracts features from image frames. Of course, the number of deep learning algorithms used in processing module 3 and the order in which they are processed can be set according to actual needs; this embodiment does not impose specific limitations on this.

[0058] Step S1021 also includes: if it is determined that no pedestrian has passed by, the determination ends.

[0059] It is understood that the automatic ad playback method provided in this embodiment matches feature word data with each ad data to obtain a matching score between at least one feature word and the ad data. This includes: selecting any ad data; querying the score of each feature word in the feature word data based on the selected ad data, including: selecting any feature word; querying whether there is a keyword in the selected ad data that is the same as the selected feature word; if there is a keyword in the selected ad data that is the same as the selected feature word, the score corresponding to the keyword that is the same as the selected feature word is used as the score of the selected feature word; if there is no keyword in the selected ad data that is the same as the selected feature word, the score of the selected feature word is 0; and summing the scores of all feature words obtained based on the selected ad data to obtain the matching score between the feature word data and the selected ad data. Through automated matching calculation of ad data and feature word data, when calculating the matching score between any ad data and feature word data, the matching score between the ad data and feature word data is calculated by assigning a score to each feature word in the feature word data based on the ad data to be calculated and then summing the scores. This accurately reflects the degree of fit between the ad data to be calculated and pedestrians in the current scenario, providing a more accurate basis for subsequent ad selection and placement. Then, the advertising data corresponding to the maximum value of the matching score between the feature word data and the advertising data is selected, that is, the advertising with the highest matching degree with pedestrians is transmitted to the display screen 4 for playback, thereby achieving accurate advertising and improving advertising conversion efficiency.

[0060] Example 2

[0061] Reference Figure 1 , Figure 4 and Figure 5 , Figure 4 This is another flowchart illustrating the automatic advertising playback method provided by the present invention. Figure 5 This is another flowchart illustrating a method for automatically playing advertisements provided by the present invention, to demonstrate another specific embodiment of the method for automatically playing advertisements provided by the present invention, including:

[0062] The same road includes at least two smart poles 1. One smart pole 1 is equipped with a camera 2 and a processing module 3, and the other smart pole 1 is equipped with a display screen 4. The camera 2 is coupled to the processing module 3, and the display screen 4 is coupled to the processing module 3. The shooting end of the camera 2 faces the side of the smart pole 1 equipped with the processing module 3 away from the side of the smart pole 1 equipped with the display screen 4. The light-emitting surface of the display screen 4 faces the same direction as the shooting end. The smart pole 1 equipped with the processing module 3 is also equipped with a distance sensor, which is coupled to the processing module 3.

[0063] S201: Processing module 3 stores preset advertisements, and processing module 3 transmits at least one preset advertisement to display screen 4; or processing module 3 transmits a sleep signal to display screen 4;

[0064] S202: The display screen 4 receives at least one preset advertisement and plays it; or the display screen 4 receives the sleep signal and turns off the screen.

[0065] S203: Camera 2 captures video stream in real time and transmits the video stream to processing module 3;

[0066] Processing module 3 pre-stores at least one piece of advertising data. The advertising data includes an advertisement, at least two keywords, and the score corresponding to each keyword. Any two pieces of advertising data have the same number of keywords, and the sum of the scores corresponding to the keywords in any two pieces of advertising data is equal.

[0067] S204: Processing module 3 receives the video stream and determines whether to enable energy-saving mode.

[0068] S2041: If energy-saving mode is enabled, determine whether a pedestrian has passed by based on the video stream; if energy-saving mode is not enabled, end the judgment.

[0069] S2042: If a pedestrian is detected, the processing module 3 acquires the measured distance from the distance sensor in real time, and the processing module 3 identifies the pedestrian's characteristic words to form characteristic word data; the characteristic word data is matched and calculated with each advertisement data to obtain a matching score between at least one characteristic word data and the advertisement data, including:

[0070] S20421: Select any ad data, and query the score of each feature word in the feature word data based on the selected ad data, including:

[0071] S204211: Select any feature word;

[0072] S204212: Query whether there are keywords in the selected advertising data that are the same as the selected feature words. If there are keywords in the selected advertising data that are the same as the selected feature words, the score corresponding to the keyword that is the same as the selected feature words will be used as the score of the selected feature words; if there are no keywords in the selected advertising data that are the same as the selected feature words, the score of the selected feature words will be 0.

[0073] S20422: Sum the scores of all feature words obtained from the selected advertising data to get the matching score between the feature word data and the selected advertising data;

[0074] S2043: Filter out the advertising data corresponding to the maximum value in the matching score between the feature word data and the advertising data; determine whether the number of filtered advertising data is greater than 1; if the number of filtered advertising data is less than or equal to 1, when the processing module 3 obtains that the measured distance of the distance sensor is less than or equal to a preset distance, transmit the advertisement of the filtered advertising data to the display screen 4; if the number of filtered advertising data is greater than 1, extract the advertisement of the filtered advertising data, randomly generate the playback order of the extracted advertisement, and when the processing module 3 obtains that the measured distance of the distance sensor is less than or equal to a preset distance, transmit the extracted advertisement and playback order to the display screen 4;

[0075] S205: When display screen 4 is playing a preset advertisement, display screen 4 receives the advertisement from the filtered advertisement data, and displays screen 4 waits for the preset advertisement to finish playing before playing the advertisement from the filtered advertisement data; or display screen 4 receives the extracted advertisements and their playback order, and displays screen 4 waits for the preset advertisement to finish playing before playing the extracted advertisements in the playback order. When display screen 4 is off, display screen 4 receives the advertisement from the filtered advertisement data and plays the advertisement from the filtered advertisement data; or display screen 4 receives the extracted advertisements and their playback order, and displays screen 4 plays the extracted advertisements in the playback order.

[0076] S206: When the processing module 3 obtains that the measured distance of the distance sensor is greater than the energy-saving distance, the processing module 3 transmits a sleep signal to the display screen 4, and the processing module 3 stops obtaining the measured distance of the distance sensor.

[0077] S207: Display 4 receives a sleep signal and turns off the screen.

[0078] It should be noted that the score corresponding to a keyword is the keyword weight. In this embodiment, any two ad data sets have the same number of keywords, and the sum of the scores corresponding to the keywords in any two ad data sets is equal, establishing an equipotential competitive environment and eliminating the deviation caused by the number of keywords and the sum of the keyword weights in the ad data. In the ad data, the multiple keywords corresponding to the ad can be extracted by an AI model, or they can be pre-set, or they can be extracted by an AI model and then modified by human configuration. Similarly, the score corresponding to the keyword can be automatically assigned after the AI ​​model extracts the keyword, or it can be set manually, or it can be automatically assigned after the AI ​​model extracts the keyword and then modified by human configuration. This embodiment does not impose specific restrictions on this.

[0079] In step S204, a judgment on energy-saving mode is added. If energy-saving mode is enabled, it is detected whether a pedestrian is passing by. When a pedestrian passes by, the most suitable advertisement is matched and played. If energy-saving mode is not enabled, the display screen 4 is kept playing a preset advertisement or the display screen 4 is kept in the off state.

[0080] In step S2042, taking a 28-year-old male running past in sportswear as an example, the processing module 3 identifies the pedestrian's feature words as: young and middle-aged, sports, and sportswear. Taking an advertising data set of 4 groups, with 3 keywords corresponding to each ad in each group, and a total score of 6 points for all keywords in each group, the feature word data is matched and calculated with each ad data set, as shown in Table 1.

[0081] Table 1. Matching Calculation Table Between Feature Word Data and Each Ad Data

[0082]

[0083] As shown in Table 1, the maximum matching score between the feature word data and the advertising data is 6, which corresponds to the gym. The gym advertisement is then transmitted to display screen 4 for playback.

[0084] In step S2042, taking a 35-year-old mother walking with a 6-year-old boy as an example, the feature words identified by processing module 3 are: young and middle-aged, female, male, child, parent-child, and walking. Taking an advertising data set of 4 groups, with 3 keywords corresponding to each ad in each group, and a total score of 6 points for all keywords in each group, the feature word data is matched and calculated with each ad data set, as shown in Table 2.

[0085] Table 2. Matching Calculation Table Between Feature Word Data and Each Ad Data

[0086]

[0087] As shown in Table 2, the maximum matching score between the feature word data and the advertising data is 5, which corresponds to the children's playground. The children's playground advertisement will be transmitted to the display screen 4 for playback.

[0088] In step S2042, taking a well-dressed 35-year-old white-collar worker passing by as an example, the feature words identified by processing module 3 are: young and middle-aged, female, makeup, work clothes, and walking. Taking an advertising data set of 4 groups, with 3 keywords corresponding to each ad in each group, and a total score of 6 points for all keywords in each group, the feature word data is matched and calculated with each ad data set, as shown in Table 3.

[0089] Table 3. Matching Calculation Table Between Feature Word Data and Each Ad Data

[0090]

[0091] As shown in Table 3, the maximum matching score between the feature word data and the advertising data is 6, which corresponds to the cosmetics advertisement. The cosmetics advertisement will be transmitted to the display screen 4 for playback.

[0092] In step S2042, taking a 65-year-old man riding a bicycle as an example, the feature words identified by processing module 3 are: elderly, male, bicycle, and exercise. Taking a scenario where there are 4 groups of advertising data, each group contains 3 keywords corresponding to the advertisement, and the total score of all keywords in each group is 6 points, the feature word data is matched and calculated with each advertising data point, as shown in Table 4.

[0093] Table 4. Matching Calculation Table Between Feature Word Data and Each Ad Data

[0094]

[0095] As shown in Table 4, the maximum matching score between the feature word data and the advertising data is 5, which corresponds to the health product advertisement. The health product advertisement will be transmitted to the display screen 4 for playback.

[0096] The above examples all assume a single pedestrian passing by. When multiple pedestrians pass by, if the same feature words appear during pedestrian feature extraction, these feature words can be superimposed. For example, if a 35-year-old man wearing a helmet riding an electric bike, a 30-year-old man walking, and a 30-year-old well-dressed woman wearing a helmet riding an electric bike pass by simultaneously, the feature words identified by processing module 3 are: male × 2, female × 1, electric bike × 2, helmet × 2, young / middle-aged × 3, makeup × 1, and walking × 1. Taking a scenario with 4 groups of advertising data, each group containing 3 keywords, and a total score of 6 for all keywords in each group, the feature word data is matched and calculated with each advertising data point, as shown in Table 5.

[0097] Table 5. Matching Calculation Table Between Feature Word Data and Each Ad Data

[0098]

[0099] As shown in Table 5, the maximum matching score between the feature word data and the advertising data is 9, which corresponds to the gym. The gym advertisement is then transmitted to display screen 4 for playback.

[0100] In step S205, the display screen 4 plays a preset advertisement. When the advertisement that best matches the pedestrian is obtained, in order to maintain the integrity of each advertisement playback, the advertisement with the highest matching degree with the current pedestrian is played after the currently playing preset advertisement ends. Since there is a certain distance between the two smart poles 1, waiting for the currently playing preset advertisement to end will not affect the current pedestrian from seeing the advertisement with the highest matching degree, thereby ensuring the completion rate of each played advertisement.

[0101] It is understood that an energy-saving mode is introduced in this embodiment. Besides maintaining the original playback or off state of the display screen 4 when the energy-saving mode is not activated, when the energy-saving mode is activated, the display screen 4 will only light up to play the advertisement with the highest matching degree to the current pedestrian when a pedestrian is within a preset distance. The display screen 4 will turn off when the current pedestrian exceeds the preset distance or even the energy-saving distance. Typically, the energy-saving distance is greater than the preset distance; of course, the energy-saving distance can also be equal to the preset distance, and this embodiment does not impose specific limitations on this. By using distance to limit the playback of the display screen 4, the optimal viewing distance for pedestrians can be guaranteed, resource waste can be avoided, and energy saving can be achieved.

[0102] As can be seen from the above embodiments, the automatic advertising playback method provided by the present invention achieves at least the following beneficial effects:

[0103] 1. The method for automatically playing advertisements provided by this invention includes: matching feature word data with each advertisement data to obtain a matching score between at least one feature word data and the advertisement data, including: selecting any advertisement data, querying the score of each feature word in the feature word data based on the selected advertisement data, including: selecting any feature word; querying whether there is a keyword in the selected advertisement data that is the same as the selected feature word; if there is a keyword in the selected advertisement data that is the same as the selected feature word, the score corresponding to the keyword that is the same as the selected feature word is used as the score of the selected feature word; if there is no keyword in the selected advertisement data that is the same as the selected feature word, the score of the selected feature word is 0; summing the scores of all feature words obtained based on the selected advertisement data to obtain the matching score between the feature word data and the selected advertisement data. Through the automated matching calculation of advertisement data and feature word data, when calculating the matching score between any advertisement data and feature word data, the matching score between the advertisement data to be calculated and the feature word data is calculated by assigning a score to each feature word of the feature word data based on the advertisement data to be calculated and then summing the scores, which accurately reflects the degree of suitability between the advertisement data to be calculated and pedestrians in the current scenario, providing a more accurate basis for subsequent advertisement screening and placement.

[0104] 2. The method for automatically playing advertisements provided by this invention includes: filtering out the advertisement data corresponding to the maximum value in the matching score between feature word data and advertisement data, transmitting the advertisements from the filtered advertisement data to a display screen, or extracting the advertisements from the filtered advertisement data, randomly generating the playback order of the extracted advertisements, and transmitting the extracted advertisements and the playback order to the display screen. By filtering out the advertisement data corresponding to the maximum value in the matching score between feature word data and advertisement data, that is, filtering out the advertisements with the highest matching degree with pedestrians and transmitting them to the display screen for playback, the method achieves precise advertising and improves advertising conversion efficiency.

[0105] While specific embodiments of the invention have been described in detail by way of examples, those skilled in the art should understand that the examples are for illustrative purposes only and not intended to limit the scope of the invention. Those skilled in the art should understand that modifications can be made to the above embodiments without departing from the scope and spirit of the invention. The scope of the invention is defined by the appended claims.

Claims

1. A method for automatically playing advertisements, characterized in that, include: The same road includes at least two smart poles. One smart pole is equipped with a camera and a processing module, and the other smart pole is equipped with a display screen. The camera is coupled to the processing module, and the display screen is coupled to the processing module. The shooting end of the camera faces the side of the smart pole with the processing module that is away from the side of the smart pole with the display screen. The light-emitting surface of the display screen faces the same direction as the shooting end. The camera captures a video stream in real time and transmits the video stream to the processing module; The processing module pre-stores at least one piece of advertising data, which includes an advertisement, at least two keywords, and a score corresponding to each keyword; the processing module receives the video stream and determines whether a pedestrian has passed by based on the video stream; if a pedestrian is determined to have passed by, the processing module identifies the pedestrian's characteristic words to form characteristic word data; Matching the feature word data with each of the advertising data to obtain a matching score for at least one feature word data and the advertising data includes: selecting any of the advertising data, and querying the score of each feature word in the feature word data based on the selected advertising data, including: selecting any of the feature words; querying whether there is a keyword in the selected advertising data that is the same as the selected feature word; if there is a keyword in the selected advertising data that is the same as the selected feature word, the score corresponding to the keyword that is the same as the selected feature word is taken as the score of the selected feature word; if there is no keyword in the selected advertising data that is the same as the selected feature word... The selected feature word has a score of 0; the scores of all the feature words obtained from the selected advertising data are added together to obtain the matching score between the feature word data and the selected advertising data; the advertising data corresponding to the maximum value of the matching score between the feature word data and the advertising data is filtered out; it is determined whether the number of filtered advertising data is greater than 1. If the number of filtered advertising data is less than or equal to 1, the advertising data of the filtered advertising data is transmitted to the display screen; if the number of filtered advertising data is greater than 1, the advertising data of the filtered advertising data is extracted, the playback order of the extracted advertising data is randomly generated, and the extracted advertising data and the playback order are transmitted to the display screen. The display screen receives the filtered advertising data and plays the advertisements, or the display screen receives the extracted advertisements and the playback order, and plays the extracted advertisements according to the playback order.

2. The method for automatically playing advertisements according to claim 1, characterized in that, Before the camera captures the video stream in real time, the following is included: The processing module stores preset advertisements, and transmits at least one of the preset advertisements to the display screen; the display screen receives at least one of the preset advertisements and plays it.

3. The method for automatically playing advertisements according to claim 2, characterized in that, When the display screen receives the filtered advertisement data, the display screen waits for the currently playing preset advertisement to finish before playing the filtered advertisement data; or when the display screen receives the extracted advertisement and the playback order, the display screen waits for the currently playing preset advertisement to finish before playing the extracted advertisement according to the playback order.

4. The method for automatically playing advertisements according to claim 1, characterized in that, Before the camera captures the video stream in real time, the following is included: The processing module transmits a sleep signal to the display screen; the display screen receives the sleep signal and turns off.

5. The method for automatically playing advertisements according to claim 2 or 4, characterized in that, After the processing module receives the video stream, it includes: The processing module determines whether the energy-saving mode is enabled. If the energy-saving mode is enabled, it determines whether a pedestrian has passed by based on the video stream. If the energy-saving mode is not enabled, the determination ends.

6. The method for automatically playing advertisements according to claim 5, characterized in that, The smart pole, which is equipped with the processing module, is also equipped with a distance sensor, which is coupled to the processing module. After determining that a pedestrian has passed by, the process also includes: The processing module acquires the measured distance from the distance sensor in real time; When the processing module obtains that the measured distance of the distance sensor is less than or equal to a preset distance, the processing module transmits the filtered advertisement data to the display screen, or the processing module transmits the extracted advertisement and the playback order to the display screen.

7. The method for automatically playing advertisements according to claim 6, characterized in that, After the screen displays the filtered advertisements, or after the screen displays the extracted advertisements in the specified playback order, the process includes: When the processing module obtains that the measured distance of the distance sensor is greater than the energy-saving distance, the processing module transmits a sleep signal to the display screen and stops obtaining the measured distance of the distance sensor. The display screen receives the sleep signal and turns off.

8. The method for automatically playing advertisements according to claim 1, characterized in that, Also includes: The number of keywords in any two sets of advertising data is the same, and the sum of the scores corresponding to the keywords in any two sets of advertising data is equal.

Images