AR teaching material management system

Abstract

The invention discloses an AR textbook management system, which comprises a textbook picture acquisition module, a textbook picture processing module, a textbook picture feature point extraction module, a feature point matching module, a material matching calling module and an AR scene caching module, the parent-child relationship is established between the material picture and the AR scene, and the material picture and the AR scene are managed and stored through the parent-child relationship, so that the AR scene can be directly called when the teaching material picture and the material picture are successfully matched, the efficiency is improved, and when the AR scene is called, the user experience is improved. The multiple AR scenes are synchronously cached through the synchronous cache, so that the calling efficiency of the subsequent AR scenes is improved, other AR scenes are locked through the lock when the AR scenes are called, the transmission efficiency of the AR scenes is improved, and the multiple AR scenes are prevented from influencing the called AR scenes.

Description

Technical Field

[0001] The present invention relates to the field of AR technology, and in particular to an AR teaching material management system. Background Art

[0002] With the rapid development of information technology, especially the popularization of the Internet, mobile terminals and augmented reality technology, the education industry is undergoing a profound change. The traditional way of teaching material management can no longer meet the needs of modern education, and teachers and students have put forward higher requirements for the use and management of teaching materials.

[0003] AR technology is becoming more and more common in the field of teaching. Currently, stereoscopic projection (AR projection) is generally used. In actual application, there are a large number of stereoscopic projection devices, and their structural dimensions are limited by space, so only one or a few can be produced. In teaching scenes with larger areas, one or a few stereoscopic projections are obviously not enough. When making large-area teaching projections, a large number of AR projection devices and matching walls are required.

[0004] Limitations of traditional textbook management. Existing textbook management systems are usually based on paper books or static electronic books, and have the following main problems: Information update lags: Once traditional textbooks are published, the information update cycle is long and it is difficult to reflect the latest subject dynamics in a timely manner; Insufficient interactivity: Paper textbooks and static electronic books lack interactivity and cannot effectively enhance students’ learning interest and participation; Inefficient management: The process of purchasing, distributing and recycling teaching materials is cumbersome and often requires a lot of manpower and material resources.

[0005] Advantages of augmented reality technology: As an emerging way of presenting information, augmented reality technology can combine virtual information with the real environment to provide a more vivid and intuitive learning experience. Its advantages include: Dynamic information display: AR technology can realize the dynamic display of teaching materials, so that students can obtain the latest information in real time; Interactive experience: Through AR applications, students can interact with the textbook content to enhance their enthusiasm and interest in learning; Multi-dimensional information presentation: AR technology can present complex knowledge in various forms such as graphics and animations, helping students to better understand and remember.

[0006] Deficiencies of existing AR teaching material management systems: The existing AR teaching material management system has chaotic backend server management, which results in the inability to timely identify and mobilize teaching material pictures or physical pictures and AR scenes, as well as a loose connection between teaching material pictures or physical pictures and AR scenes. It is impossible to quickly obtain the AR scenes corresponding to teaching material pictures or physical pictures when identifying them, and the selection of AR scenes is rather cumbersome and inconvenient, making it inconvenient to create, display and interact with AR scenes.

[0007] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0008] In view of the problems in the related technology, the present invention proposes an AR teaching material management system to overcome the above technical problems existing in the existing related technology.

[0009] To this end, the specific technical solution adopted by the present invention is as follows: An AR teaching material management system, comprising a teaching material picture acquisition module, a teaching material picture processing module, a teaching material picture feature point extraction module, a feature point matching module, a material matching calling module, and an AR scene caching module; The teaching material picture acquisition module is used to acquire teaching material pictures taken by a mobile device; The textbook image processing module pre-processes the acquired textbook images, including textbook image enhancement, textbook image segmentation, and textbook image feature detection and description; The teaching material picture feature point extraction module is used to extract the feature point descriptors generated by the teaching material picture processing module, and store all feature point descriptors as a feature point vector set; The feature point matching module is used to match the feature point vector set generated by the teaching material picture with the feature point vector set of the material picture in the database acquired in advance, and determine that the teaching material picture is the same as the material picture in the database; The material matching and calling module is used to call material pictures that are determined to have the same content, and the matching of the teaching material pictures with the material pictures in the database adopts multiple repeated matching; The AR scene caching module is used to call the AR scene in the database through a foreign key after the teaching material picture matches the material picture. One material picture corresponds to multiple AR scenes, and multiple AR scenes are cached synchronously.

[0010] Preferably, the textbook image preprocessing is used to scale, crop and adjust the color of the textbook image; specifically, the textbook image enhancement uses an algorithm to reduce noise in the image; the textbook image segmentation divides the textbook image into different areas by setting a threshold; the textbook image feature detection and description detects feature points through the FAST algorithm, and for each feature point, calculates the local image blocks around it and generates a feature point descriptor.

[0011] Preferably, the teaching material picture enhancement is to smooth the teaching material picture by Gaussian filtering, and the calculation formula of the Gaussian filtering is as follows: ; in, Indicates that the processed textbook image is at position The pixel value of Indicates that the textbook image is at position The pixel value of the pixel is used to calculate the center pixel of its neighborhood The new value of Represents the standard deviation, which controls the width of the Gaussian function; represents the normalization constant, ensuring that the integral of the Gaussian function is 1; Represents the value of the Gaussian function, which determines the degree of influence of each pixel in the neighborhood on the central pixel; Represents pixels To center pixel The square of the distance.

[0012] Preferably, the teaching material picture segmentation divides the filtered teaching material picture into background and foreground by single threshold segmentation, and the specific calculation formula of single threshold segmentation is as follows: Set a threshold , and the threshold If the grayscale value is 120, the pixels with grayscale value less than or equal to 120 in the textbook image will be regarded as the background, and the pixels with grayscale value greater than 120 will be regarded as the foreground; ; in, is the output binary textbook picture, Is the textbook picture in position The pixel value of is the set threshold; And the separated foreground is used as feature points to extract the image.

[0013] Preferably, the steps of detecting and describing the characteristics of the teaching material pictures are as follows: Convert the feature point extraction image into a grayscale image; Use the FAST algorithm to detect feature points in the grayscale image; For each feature point, the local image patch around it is calculated and a feature point descriptor is generated.

[0014] Preferably, the calculation formula of the FAST algorithm to detect feature points in the grayscale image is as follows: For each central pixel p in the grayscale image, use an N-point neighborhood to determine whether it is a corner point, set the threshold t, and the value of t is 12. The judgment condition is: ; in, The brightness in the neighborhood is greater than The number of pixels, The brightness in the neighborhood is less than The number of pixels; is a predefined threshold value, set to 128, Represents the function of detecting corner points in a grayscale image; If the output result value is 1, it is determined to be a feature point. If the output result value is 0, it is determined not to be a feature point.

[0015] Preferably, the calculation of generating a feature point descriptor from the feature point and the local image block around it is as follows: Set a set of random pixel pairs ,in and are the pixel coordinates near the feature point: ; Among them, is the kth bit of the binary descriptor, N is the number of randomly selected pixel pairs, and all Combine to form a binary feature point descriptor of length N; All binary feature point descriptors are stored as a feature point vector set.

[0016] Preferably, the feature point vector set is matched with the feature point vector set of the material picture in the database, and the matching calculation formula is as follows: The distance metric calculates the straight-line distance between two feature vectors. ,in, Represents the distance between two feature points. and Indicates that the first feature point and the second feature point are in The coordinates on the dimension, represents the square of the coordinate difference in each dimension, It means summing the squared differences of all dimensions; Set a threshold D if the Euclidean distance between two feature points , then the two feature points are considered similar. If the Euclidean distance between the two feature points is , then the two feature points are considered to be dissimilar.

[0017] Preferably, when the feature point vector set is matched with the feature point vector set of the material picture in the database, all feature points in the feature point vector set are matched, and the matching results are counted; Set the number of similar feature points to , the number of dissimilar feature points is , and the feature points that cannot be matched are , then the correct matching rate is calculated as follows: ; like , then the grayscale image and the material image are considered to be the same image. , then the grayscale image and the material image are considered to be similar images. , then the grayscale image is considered to be completely different from the material image; In the case of repeated matches between textbook images and material images in the database, the correct match rate shall prevail. , then a match directly determines that the textbook image and the material image are the same image. , then it is necessary to repeat the match twice, and the correct matching rate of the two matches needs to be , then the textbook picture and the material picture are considered to be the same picture. , then the matching is repeated three times, and the correct matching rate of the three matches is at least once , then the teaching material pictures and material pictures are considered to be familiar pictures.

[0018] Preferably, when the AR scene caching module matches and determines that the teaching material picture and the material picture in the database are the same picture, the AR scene in the database that is associated with the material picture through a foreign key is cached, and multiple AR scenes are cached synchronously; The rule for establishing a foreign key is to add a timestamp to the storage address of the material image, and then generate a unique identifier through a hash function; And the database stores the material image in the first storage area of ​​the database; And the database stores the AR scene in the second storage area in the database.

[0019] The beneficial effects of the present invention are: 1. The present invention connects the material picture and the AR scene, so that after matching and identifying the material picture, the AR scene can be directly called, thereby improving the output efficiency of the AR scene. The material picture and the AR scene are connected through a parent-child relationship, and the parent-child relationship is implemented through a foreign key. The foreign key generates a unique identifier by adding a timestamp to the storage address of the material picture through a hash function, so that the AR scene can be called after the material picture is successfully matched; 2. The present invention improves the clarity of the photographed textbook pictures by processing them, and facilitates feature extraction of the textbook pictures, including textbook picture preprocessing, textbook picture enhancement, textbook picture segmentation, and textbook picture feature detection and description, so that the textbook pictures can remain smooth, eliminate the influence of noise, and improve the clarity of the pictures. The foreground and background in the textbook pictures have been distinguished, and feature points of the foreground have been extracted and described, so as to facilitate comparison of the feature points with the material pictures, determine whether the textbook pictures and the material pictures are the same picture, and facilitate calling the AR scene; 3. The present invention converts the collected and extracted feature points into descriptors, and stores all the descriptors in the teaching material picture into a feature point vector set, so as to facilitate comparison and determination with the feature point vector set of the material picture, and performs repeated matching according to the determination results, thereby improving the matching accuracy and preventing the matching error caused by the different angles of the teaching material picture and the material picture. The matching results can be improved by performing determinations through multiple matching structures; 4. When determining that the teaching material picture and the material picture are the same content picture, the present invention calls the AR scene through the foreign key parent-child relationship between the material picture and the AR scene, and caches the corresponding multiple AR scenes to improve the output efficiency of the AR scene. When selecting an AR scene, the other cached AR scenes are locked by locking, and only the selected AR scene is transmitted, thereby improving the efficiency and accuracy of transmission and preventing the cache of other AR scenes from affecting the transmission. The present invention establishes a parent-child relationship between material images and AR scenes, and manages and stores the material images and AR scenes through the parent-child relationship, so that when the teaching material images and the material images are successfully matched, the AR scene can be directly called to improve efficiency. When the AR scene is called, multiple AR scenes are synchronously cached through synchronous caching to improve the calling efficiency of subsequent AR scenes. When the AR scene is called, other AR scenes are locked through locking to improve the efficiency of AR scene transmission and prevent multiple AR scenes from affecting the called AR scene. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 is a schematic diagram of the system structure of an AR teaching material management system according to an embodiment of the present invention; Figure 2 It is a schematic diagram of the steps of detecting and describing features of teaching material images in an AR teaching material management system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in the field should be able to understand other possible implementations and advantages of the present invention. The components in the figures are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0023] According to an embodiment of the present invention, an AR teaching material management system is provided.

[0024] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1-2 As shown, an AR teaching material management system according to an embodiment of the present invention includes a teaching material picture acquisition module, a teaching material picture processing module, a teaching material picture feature point extraction module, a feature point matching module, a material matching calling module, and an AR scene caching module; The teaching material picture acquisition module is used to acquire teaching material pictures taken by a mobile device; The textbook image processing module pre-processes the acquired textbook images, including textbook image enhancement, textbook image segmentation, and textbook image feature detection and description; The teaching material picture feature point extraction module is used to extract the feature point descriptors generated by the teaching material picture processing module, and store all feature point descriptors as a feature point vector set; The feature point matching module is used to match the feature point vector set generated by the teaching material picture with the feature point vector set of the material picture in the database acquired in advance, and determine that the teaching material picture is the same as the material picture in the database; The material matching and calling module is used to call material images that are determined to have the same content, and the matching of teaching material images with material images in the database adopts multiple repeated matching; The AR scene cache module is used to call the AR scene in the database through a foreign key after the textbook image matches the material image. One material image corresponds to multiple AR scenes, and multiple AR scenes are cached synchronously.

[0025] Specifically, textbook image preprocessing is used to scale, crop and adjust the color of textbook images; specifically, textbook image enhancement uses an algorithm to reduce noise in the image; textbook image segmentation divides textbook images into different regions by setting thresholds; textbook image feature detection and description detects feature points through the FAST algorithm, and for each feature point, calculates the local image blocks around it and generates feature point descriptors; It should be noted that by preprocessing the textbook images, the clarity of the textbook images can be improved, making it easier to detect and extract feature points of the textbook images.

[0026] Specifically, the textbook image enhancement is to smooth the textbook image through Gaussian filtering. The calculation formula of Gaussian filtering is as follows: ; in, Indicates that the processed textbook image is at position The pixel value of Indicates that the textbook image is at position The pixel value of the pixel is used to calculate the center pixel of its neighborhood The new value of Represents the standard deviation, which controls the width of the Gaussian function; represents the normalization constant, ensuring that the integral of the Gaussian function is 1; Represents the value of the Gaussian function, which determines the degree of influence of each pixel in the neighborhood on the central pixel; Represents pixels To center pixel The square of the distance; It should be noted that the collected textbook pictures or real-object pictures are optimized through Gaussian filtering, which facilitates filtering out noise in the textbook pictures or real-object pictures, improves the clarity of the textbook pictures or real-object pictures, and facilitates the subsequent production of AR scenes for the textbook pictures or real-object pictures.

[0027] Specifically, the textbook image segmentation is to divide the filtered textbook image into background and foreground by single threshold segmentation. The specific calculation formula of single threshold segmentation is as follows: Set a threshold , and the threshold If the grayscale value is 120, the pixels with grayscale value less than or equal to 120 in the textbook image will be regarded as the background, and the pixels with grayscale value greater than 120 will be regarded as the foreground; ; in, is the output binary textbook picture, Is the textbook picture in position The pixel value of is the set threshold; And the separated foreground is used as feature points to extract the image; It should be noted that the teaching material picture is segmented into foreground and background through threshold segmentation, and the picture is extracted by taking the foreground as feature points to perform feature point extraction, thereby reducing the influence of the background on feature point extraction.

[0028] Specifically, the steps of textbook image feature detection and description are as follows: Convert the feature point extraction image into a grayscale image; Use the FAST algorithm to detect feature points in the grayscale image; For each feature point, calculate the local image block around it and generate a feature point descriptor; It should be noted that the above steps facilitate calculation of feature points in teaching material images, and convert feature point extraction images into grayscale images to facilitate calculation of feature points therein, and generate descriptors for feature points to facilitate subsequent matching.

[0029] Specifically, the calculation formula for detecting feature points in the grayscale image using the FAST algorithm is as follows: For each central pixel p in the grayscale image, use an N-point neighborhood to determine whether it is a corner point, set the threshold t, and the value of t is 12. The judgment condition is: ; in, The brightness in the neighborhood is greater than The number of pixels, The brightness in the neighborhood is less than The number of pixels; is a predefined threshold value, set to 128, Represents the function of detecting corner points in a grayscale image; If the output result value is 1, it is determined to be a feature point; if the output result value is 0, it is determined not to be a feature point; It should be noted that the feature points in the grayscale image are acquired and detected, and the feature points can be accurately acquired by calculating whether each central pixel p is a corner point and judging by a threshold value and each output result.

[0030] Specifically, the calculation of the feature point descriptor generated by the feature point and the local image block around it is as follows: Set a set of random pixel pairs ,in and are the pixel coordinates near the feature point: ; Among them, is the kth bit of the binary descriptor, N is the number of randomly selected pixel pairs, and all Combine to form a binary feature point descriptor of length N; Store all binary feature point descriptors as a feature point vector set; It should be noted that the feature description provides a unique identifier for each feature point, so that feature matching can be effectively performed between different images, thereby improving the rapid matching between teaching material images and material images.

[0031] Specifically, the feature point vector set is matched with the feature point vector set of the material image in the database, and the matching calculation formula is as follows: The distance metric calculates the straight-line distance between two feature vectors. ,in, Represents the distance between two feature points. and Indicates that the first feature point and the second feature point are in The coordinates on the dimension, represents the square of the coordinate difference in each dimension, It means summing the squared differences of all dimensions; Set a threshold D if the Euclidean distance between two feature points , then the two feature points are considered similar. If the Euclidean distance between the two feature points is , then the two feature points are considered to be dissimilar; It should be noted that by matching the feature point vector sets of the teaching material picture and the material picture to determine whether the teaching material picture and the material picture are the same picture, accurate judgment can be achieved.

[0032] Specifically, in matching the feature point vector set with the feature point vector set of the material image in the database, all feature points in the feature point vector set are matched, and the matching results are counted; Set the number of similar feature points to , the number of dissimilar feature points is , and the feature points that cannot be matched are , then the correct matching rate is calculated as follows: ; like , then the grayscale image and the material image are considered to be the same image. , then the grayscale image and the material image are considered to be similar images. , then the grayscale image is considered to be completely different from the material image; In the case of repeated matches between textbook images and material images in the database, the correct match rate shall prevail. , then a match directly determines that the textbook image and the material image are the same image. , then it is necessary to repeat the match twice, and the correct matching rate of the two matches needs to be , then the textbook picture and the material picture are considered to be the same picture. , then the matching is repeated three times, and the correct matching rate of the three matches is at least once , then the textbook pictures and material pictures are considered to be familiar pictures; It should be noted that by matching all feature point vector sets and counting all matching results, it is determined whether the teaching material picture and the material picture are the same picture, or whether they have the same features, based on the statistical matching probability. In order to improve the matching degree, multiple matching judgments are performed to eliminate errors caused by viewing angles.

[0033] It should be noted that when the AR scene cache module matches and determines that the textbook image and the material image in the database are the same image, the AR scene in the database that is linked to the material image through a foreign key is cached, and multiple AR scenes are cached synchronously; The rule for establishing a foreign key is to add a timestamp to the storage address of the material image, and then generate a unique identifier through a hash function; And the database stores the material image in the first storage area of ​​the database; And the database stores the AR scene in the second storage area in the database.

[0034] It should be noted that after the feature points are matched, the material image is determined, and the AR scene in the second storage area is retrieved through the unique identifier to realize the synchronous linkage between the calling of the material image and the AR scene; by establishing a foreign key, the AR scene can be called when the material image is recognized, thereby increasing the output rate of the AR scene, and by setting a unique identifier, errors in the calling of the AR scene can be prevented; and a textbook image or a real-life image corresponds to multiple AR scenes, and multiple AR scenes are cached and stored synchronously; when one of the AR scenes is clicked, the back-end server uses a lock to ensure that only one thread accesses the cached data at the same time, and through the lock mechanism, the output of the AR scene is ensured to be single when the AR scene is written; multiple corresponding AR scenes are cached through synchronous caching, which facilitates the rapid output of the AR scene, reduces the calling time of the AR scene, and improves efficiency, and by determining the access of a single thread through a lock, the output of multiple AR scenes can be prevented. Caused by confusion.

[0035] In summary, the present invention establishes a parent-child relationship between material images and AR scenes, and manages and stores material images and AR scenes through the parent-child relationship, so that when the teaching material images and the material images are successfully matched, the AR scene can be directly called to improve efficiency. When the AR scene is called, multiple AR scenes are synchronously cached through synchronous caching to improve the calling efficiency of subsequent AR scenes. When the AR scene is called, other AR scenes are locked through locking to improve the efficiency of AR scene transmission and prevent multiple AR scenes from affecting the called AR scene.

[0036] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An AR teaching material management system, characterized in that: include: Textbook image acquisition module, textbook image processing module, textbook image feature point extraction module, feature point matching module, material matching calling module and AR scene caching module; The teaching material picture acquisition module is used to acquire teaching material pictures taken by a mobile device; The textbook image processing module pre-processes the acquired textbook images, including textbook image enhancement, textbook image segmentation, and textbook image feature detection and description; The teaching material picture feature point extraction module is used to extract the feature point descriptors generated by the teaching material picture processing module, and store all feature point descriptors as a feature point vector set; The feature point matching module is used to match the feature point vector set generated by the teaching material picture with the feature point vector set of the material picture in the database acquired in advance, and determine that the teaching material picture is the same as the material picture in the database; The material matching and calling module is used to call material pictures that are determined to have the same content, and the matching of the teaching material pictures with the material pictures in the database adopts multiple repeated matching; The AR scene caching module is used to call the AR scene in the database through a foreign key after the teaching material picture matches the material picture. One material picture corresponds to multiple AR scenes, and multiple AR scenes are cached synchronously.

2. The AR teaching material management system according to claim 1, characterized in that: The textbook image preprocessing is used to scale, crop and adjust the color of the textbook images; specifically, the textbook image enhancement uses an algorithm to reduce noise in the image; the textbook image segmentation divides the textbook image into different areas by setting a threshold; the textbook image feature detection and description detects feature points through the FAST algorithm, and for each feature point, calculates the local image blocks around it and generates a feature point descriptor.

3. The AR teaching material management system according to claim 2, characterized in that: The textbook image enhancement is performed by smoothing the textbook image through Gaussian filtering, and the calculation formula of the Gaussian filtering is as follows: ； in, Indicates that the processed textbook image is at position The pixel value of Indicates that the textbook image is at position The pixel value of the pixel is used to calculate the center pixel of its neighborhood The new value of Represents the standard deviation, which controls the width of the Gaussian function; represents the normalization constant, ensuring that the integral of the Gaussian function is 1; Represents the value of the Gaussian function, which determines the degree of influence of each pixel in the neighborhood on the central pixel; Represents pixels To center pixel The square of the distance.

4. The AR teaching material management system according to claim 3, characterized in that: The textbook image segmentation divides the filtered textbook image into background and foreground by single threshold segmentation. The specific calculation formula of single threshold segmentation is as follows: Set a threshold , and the threshold If the grayscale value is 120, the pixels with grayscale value less than or equal to 120 in the textbook image will be regarded as the background, and the pixels with grayscale value greater than 120 will be regarded as the foreground; ； in, Is the textbook picture in position The pixel value of is the set threshold; And the separated foreground is used as feature points to extract the image.

5. The AR teaching material management system according to claim 4, characterized in that: The steps of detecting and describing the characteristics of the teaching material pictures are as follows: Convert the feature point extraction image into a grayscale image; Use the FAST algorithm to detect feature points in the grayscale image; For each feature point, the local image patch around it is calculated and a feature point descriptor is generated.

6. The AR teaching material management system according to claim 5, characterized in that: The calculation formula for detecting feature points in the grayscale image using the FAST algorithm is as follows: For each central pixel p in the grayscale image, use an N-point neighborhood to determine whether it is a corner point, set the threshold t, and the value of t is 12. The judgment condition is: ； in, The brightness in the neighborhood is greater than The number of pixels, The brightness in the neighborhood is less than The number of pixels; is a predefined threshold value, set to 128, Represents the function of detecting corner points in a grayscale image; If the output result value is 1, it is determined to be a feature point. If the output result value is 0, it is determined not to be a feature point.

7. The AR teaching material management system according to claim 6, characterized in that: The calculation of generating feature point descriptors for the feature point and the local image blocks around it is as follows: Set a set of random pixel pairs ,in and are the pixel coordinates near the feature point: ； in, is the kth bit of the binary descriptor, N is the number of randomly selected pixel pairs, and all Combine to form a binary feature point descriptor of length N; All binary feature point descriptors are stored as a feature point vector set.

8. The AR teaching material management system according to claim 7, characterized in that: The feature point vector set is matched with the feature point vector set of the material image in the database, and the matching calculation formula is as follows: The distance metric calculates the straight-line distance between two feature vectors. ,in, Represents the distance between two feature points. and Indicates that the first feature point and the second feature point are in The coordinates on the dimension, represents the square of the coordinate difference in each dimension, It means summing the squared differences of all dimensions; Set a threshold D if the Euclidean distance between two feature points , then the two feature points are considered similar. If the Euclidean distance between the two feature points is , then the two feature points are considered to be dissimilar.

9. The AR teaching material management system according to claim 8, characterized in that: In matching the feature point vector set with the feature point vector set of the material picture in the database, all feature points in the feature point vector set are matched, and the matching results are counted; Set the number of similar feature points to , the number of dissimilar feature points is , and the feature points that cannot be matched are , then the correct matching rate is calculated as follows: like , then the grayscale image and the material image are considered to be the same image. , then the grayscale image and the material image are considered to be similar images. , then the grayscale image is considered to be completely different from the material image; In the case of repeated matches between textbook images and material images in the database, the correct match rate shall prevail. , then a match directly determines that the textbook image and the material image are the same image. , then it is necessary to repeat the match twice, and the correct matching rate of the two matches needs to be , then the textbook picture and the material picture are considered to be the same picture. , then the matching is repeated three times, and the correct matching rate of the three matches is at least once , then the teaching material pictures and material pictures are considered to be familiar pictures.

10. An AR teaching material management system according to claim 9, characterized in that: When the AR scene caching module matches and determines that the teaching material picture and the material picture in the database are the same picture, the AR scene in the database that is associated with the material picture through a foreign key is cached, and multiple AR scenes are cached synchronously; The rule for establishing a foreign key is to add a timestamp to the storage address of the material image, and then generate a unique identifier through a hash function; And the database stores the material image in the first storage area of ​​the database; And the database stores the AR scene in the second storage area in the database.

Images