Reading progress tracking method for household intelligent bookcase

Abstract

The invention belongs to the technical field of smart home, and discloses a reading progress tracking method of a household intelligent bookcase, the household intelligent bookcase is integrated with an OpenBiblio system, and data is transmitted through a communication module, and the method comprises the following steps: step 1, books are placed in a scanner area of the household intelligent bookcase, a scanner transmits scanned book images to the system, and the system is connected with the system; the system judges whether the book is opened; step 2, if the book is judged to be in an open state, running an accurate mode; and step 3, if the book is judged to be in a closed state after scanning by the scanner in the step 1, running a fuzzy mode. According to the method, the reading progress of the user can be intelligently tracked and recorded, manual input or additional equipment is not needed, manual recording or bookmarks are not needed, physical loss of the book is avoided, intelligent management of the paper book is improved, and a fuzzy mode and an accurate mode are added to adapt to different requirements of the user.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of smart home, and in particular relates to a reading progress tracking method of a household smart bookcase. BACKGROUND

[0002] With the continuous popularization of smart home, people's demand for the intelligentization of home equipment is increasing, especially in the aspect of book management. The traditional bookcase only provides a book storage function, and users usually need to manually search for specific books, which is particularly inefficient and tedious for users with a large number of books.

[0003] At present, some smart bookcases on the market have introduced electronic tags, RFID and other technologies to assist in book management, but these technologies still face some problems: existing smart bookcases often rely on RFID tags or two-dimensional codes to identify books, but due to the limited application of RFID tags, and the lack of obvious advantages in daily household use. Most smart bookcases only have the functions of book access and query, and lack intelligent analysis and dynamic management functions of book thickness, creases and other information. Even if RFID or barcodes are used, it is not possible to accurately control the book storage location or perform automatic book pushing operations. The reading progress of paper books cannot be automatically recorded like e-books, and users need to manually mark or rely on bookmarks and other tools, but these methods cannot provide accurate and convenient progress management methods, and are prone to marking errors. SUMMARY

[0004] To solve the above technical problems, the present application provides a reading progress tracking method of a household smart bookcase, which can record the reading progress without manual marking or inserting bookmarks, improve the intelligent management of paper books, and enhance the automation level, and adds fuzzy mode and precise mode to adapt to different user needs.

[0005] To achieve the above purpose, the present application is realized by the following technical scheme:

[0006] The present application is a reading progress tracking method of a household smart bookcase, the household smart bookcase is integrated with an OpenBiblio system, and the household smart bookcase and the OpenBiblio system transmit data through a communication module. The reading progress tracking method specifically includes the following steps:

[0007] Step 1, place the book in the scanner area of the household smart bookcase, the scanner transmits the scanned book image to the OpenBiblio system, and the OpenBiblio system judges whether the book is open;

[0008] Step 2, if the book is judged to be in an open state, the accurate mode is run, the scanner performs a first scan, scans the page number of the book at this time and records, when the user closes the book, the scanner performs a second scan, the OpenBiblio system judges whether the book is open, if the book is in an open state, the OpenBiblio system prompts the user to close the book, if the book is in a closed state, the scanner scans the book cover to record the book information, after the scanner completes the scan, the OpenBiblio system outputs the page number, book name and author information of the book reading up to the present;

[0009] Step 3, if the OpenBiblio system judges that the book is in a closed state after the scanner scans in step 1, the fuzzy mode is run, the high-resolution camera placed above each layer of bookshelf in the bookshelf scans the side of the book when the book is closed, records the book name and author information, the scanner extracts the crease features on the book spine, an image processing algorithm is used to identify and enhance the crease area, based on the crease features, the OpenBiblio system associates the crease depth with the reading progress to infer the approximate reading progress, after the scan is completed, the OpenBiblio system outputs the book information and the approximate reading progress.

[0010] A further improvement of the present application is that in step 1, the book is placed in the scanner area, and the specific method for judging whether the book is open is as follows: if the text coverage ratio in the image recognized by the scanner, that is, the proportion of the recognized text pixels in the total pixels, exceeds a set threshold, it is judged that the image scanned by the scanner is a large amount of text, and it is considered that the book is open, and then the accurate mode is enabled, and if the text coverage ratio is lower than the set threshold, it is judged that the image is a small amount of text, and it is considered that the book is in a closed state, and then the fuzzy mode is enabled.

[0011] A further improvement of the present application is that an image processing algorithm is used to identify and enhance the crease area, which specifically includes the following steps:

[0012] Step 3.1, Gaussian blur is applied to the scanned image to smooth the image and reduce noise, Gaussian processing can eliminate fine noise and enhance the recognizability of traces, which is very useful for traces with small details or unclear edges.

[0013] Step 3.2, median filtering is performed to remove random noise such as impulse noise and retain image edge information, so that the image edge is clearer and convenient for subsequent contour detection;

[0014] Step 3.3, Sobel operator is used to detect the horizontal and vertical edges of the image, especially in the case of rich image details and textures. The edge position is found by calculating the horizontal and vertical image gradients.

[0015] Further improvement of the present application is that in step 3.3, assuming that the gray value of a certain pixel point in the image is I(x, y), then the gradient in the horizontal direction and the vertical direction at the pixel point is obtained by convolution, and then the amplitude G of the gradient in the horizontal direction and the vertical direction is calculated to represent the intensity of the edge, and the gradient direction θ is calculated to represent the direction of the edge,

[0016] Horizontal direction G x Convolution kernel:

[0017]

[0018] Vertical direction G y Convolution kernel:

[0019]

[0020] Gradient amplitude:

[0021]

[0022] For convenience of calculation, an approximate formula is used:

[0023] G approximately equal to |G x + |G y |

[0024] Gradient direction:

[0025]

[0026] Further improvement of the present application is that in step 3, the system associates the crease depth with the reading progress through a machine learning algorithm, and estimates the approximate reading progress, specifically: using a convolutional neural network to actively extract image features and classify, through the CNN model, the relationship between different types of traces and the reading progress of the book can be learned. Collect book samples with labeled reading progress as training data, use the convolutional layer of the convolutional neural network to extract the image feature map, the pooling layer to reduce the size of the image feature map, and the fully connected layer of the convolutional neural network to classify and output the reading progress. Use the region convolutional neural network to locate the specific trace area in the image feature, help to accurately extract certain areas of the book and judge whether they are related to the reading progress.

[0027] The further improvement of the present application is that the household intelligent book cabinet comprises a display screen arranged on the book cabinet, a scanner for scanning book information, a plurality of independent electric push plates controlled by stepping motors are arranged on each layer of bookshelf of the household intelligent book cabinet, the electric push plates are arranged on the rear side of the household intelligent book cabinet, and the electric push plate used for pushing out the books selected by the user is arranged on the rear side of the household intelligent book cabinet; a plurality of layers of bookshelves are arranged in the household intelligent book cabinet, a high-resolution camera is arranged above each layer of bookshelf in the household intelligent book cabinet, and the high-resolution camera is used for positioning the position and change of the books and identifying the storage state of the current books on the bookshelf, i.e. the taking and storing of the books; when the user stores the book information into the OpenBiblio system through the scanner or the high-resolution camera, the OpenBiblio system records the book name, author information and thickness data, the book name, author information and thickness data are sent to the control system of the electric push plate through the API protocol of the OpenBiblio system, and the number of the electric push plate is adjusted according to the thickness; when the user queries the books in the OpenBiblio system, the OpenBiblio system returns the position information and thickness data of the books, the control system of the electric push plate receives the data, drives the electric push plate to perform the pushing operation, pushes the target books to the area where the user can take the books, the high-resolution camera feeds back the book image to the control system of the electric push plate, ensures the execution of the action of the electric push plate, and the control system of the electric push plate confirms the completion of the action and returns the state information to the OpenBiblio system.

[0028] The further improvement of the present application is that the household intelligent book cabinet uses a dynamic recommendation algorithm to introduce the time factor and the user behavior change factor, and provides personalized and dynamic book recommendation services for the user, and the specific steps include the following steps.

[0029] Step 4.1, book attribute collection:

[0030] A book set B={B1, B2,..., B n} is recorded, the initial category C(B i ), the initial classification label of the book and the latest reading time T last (B i ) of each book are recorded, the time factor is introduced, i.e. the behavior data X k of the user is collected, the behavior data X k includes the behavior data of the reading duration D read (B i ) (unit: minute), the reading frequency F read (B i ) (times / day) and the reading time period T segment (B i ) (for example, morning / afternoon / night);

[0031] Dynamic weight initialization: adjust weights ω1, ω2, ω3 according to changes in user behavior. Weight calculation formula:

[0032]

[0033] Among them, H(X k ) is the behavioral data X k Information entropy is used to measure the importance of the data;

[0034] Step 4.2. Calculate the priority score for each book:

[0035] P(B i )=ω1·D read (B i )+ω2·F read (B i )+ω3·RecencyWeight(B i )

[0036]

[0037] Among them, Δt(B i ) is the current time and the recent reading time T last (B i ) days;

[0038] Step 4.3. Calculate the probability that each book belongs to multiple categories using the fuzzy membership function:

[0039]

[0040] in, For Book B i Belongs to category c k The membership degree, T is the temperature parameter, which is used to control the fuzziness of the classification. The update is run every night to dynamically adjust the classification. Books with values ​​greater than the threshold are reclassified;

[0041] Step 4.4, target recommendation: Comprehensive priority score P(B i ) and diversity goals, resulting in a multi-objective score:

[0042]

[0043] Among them, Diversity (B j ) is the diversity of books in user behavior, α and β control the trade-off between priority and diversity;

[0044] Sort by multi-objective scores and select the top N books for recommendation generation:

[0045] R = {B1, B2,... B N}

[0046] Wherein, Score(B1)>Score(B2)>…>Score(B N )

[0047] Step 4.5, task scheduling: set to run every night 23:59: dynamic weight update, priority calculation, classification adjustment and recommendation list generation.

[0048] The further improvement of the application is that: assuming that the user has two kinds of storage states of the bookshelf, namely, storing books and taking books, the storing books is divided into taking books using the system and directly placing books on the bookshelf, the taking books is divided into taking books using the system and directly taking books from the bookshelf, and the environment of the bookshelf is relatively stable, only when the user takes books and stores books, the environment of the bookshelf will change greatly, therefore, the high-resolution camera uses the frame difference method to judge whether the current user's behavior is taking books or storing books, compares the current frame captured by the bookshelf camera with the previous background frame, the background frame is a reference image of the bookshelf in the static state, and the background frame is automatically updated after a long time of stability, and the frame difference method is used to judge whether the current user's behavior is taking books or storing books, and the specific steps are as follows: the frame difference method detects that the environment of the bookshelf changes, judges whether the user stores or takes books using the OpenBiblio system, if the user stores or takes books using the OpenBiblio system, the control system of the electric push plate judges whether it is storing books, if it is storing books, the control system of the electric push plate binds the scanned book information and the position where the camera detects that the environment changes, and stores the book information and the position information into the OpenBiblio system, if it is not storing books, the OpenBiblio system obtains the position information of the book, and transmits the information to the control system of the electric push plate through the communication module, so that the book is pushed out by using the electric push plate, and then the system updates the book information; if it is not storing or taking books using the system, the frame difference method is used to judge whether it is storing books, if it is storing books, the control system generates a 'position-identified book' record and a number, and records the position, if it is not storing books, it is judged whether there is book information in the control system, if there is book information, the control system updates the book state to'removed' and removes the book position information, if there is no book information, the control system deletes the number and the position information together; when the user needs to take out a book, the control system automatically reads the thickness of the book obtained through the scanner or the camera and selects the corresponding number of push plates, and then the control system controls the electric push plate to run and push the book out.

[0049] The application has the following beneficial effects:

[0050] The smart home bookcase of the present invention has a reading progress tracking function by crease scanning, which can intelligently track and record the user's reading progress without relying on manual input or additional equipment;

[0051] The present invention realizes the reading progress tracking function by designing an independent book information scanning area, automatically scanning and displaying detailed information of the book and related information of the book folds. The system automatically updates the reading progress of the book without manual recording or reliance on bookmarks, avoiding physical damage to the book, improving the intelligent management of paper books, and enhancing the level of automation. Fuzzy mode and precise mode are added to adapt to the different needs of users.

[0052] The present invention introduces a book-retrieval function of a multi-module push plate, and combines sensors and cameras to automatically adjust the number and thrust of the push plates to achieve the function of pushing the books needed by the user from the numerous books in the bookcase, so as to facilitate the user to quickly get the books they need. This not only saves the material of the baffles used in dividing the space, but also increases the aesthetics and utilization rate of the bookshelf space.

[0053] The present invention also uses a dynamic classification recommendation algorithm that can combine the user's reading history and preferences to generate recommendations based on the user's reading frequency and content preferences, and display books that may be of interest. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a flow chart of the fold scanning reading progress tracking system of the present invention.

[0055] Figure 2 It is a schematic diagram of judging the book depositing and retrieving behaviors of the present invention.

[0056] Figure 3 Schematic diagram of the book thickness identification method of the present invention.

[0057] Figure 4 It is a structural diagram of the household intelligent bookcase of the present invention.

[0058] Figure 5 It is a side view of the household intelligent bookcase of the present invention.

[0059] Figure 6 This is a control flow chart of the present invention for adjusting the number of electric push plates pushed out according to thickness. DETAILED DESCRIPTION

[0060] The following diagrams illustrate embodiments of the present invention. For clarity, many practical details are included in the following description. However, it should be understood that these practical details are not intended to limit the present invention. In other words, in some embodiments of the present invention, these practical details are not essential.

[0061] likeFigures 4-5 As shown, the present application provides a household intelligent bookcase, the bookcase is internally provided with several layers of bookshelves, a high-resolution camera is arranged above each layer of bookshelves in the bookcase, the high-resolution camera is used for positioning the position and change of books, identifying the storage state of the current books on the bookshelf, i.e. the taking and storing of books, the bookcase further comprises a display screen 3 arranged on the bookcase, a scanner 2 for scanning book information, a plurality of independent electric push plates 4 controlled by stepping motors are arranged on each layer of bookshelves of the household intelligent bookcase, the electric push plates are arranged at the rear side of the household intelligent bookcase, and are used for pushing out the books selected by the user. When the user stores the book information into the OpenBiblio system through the scanner or the high-resolution camera, the OpenBiblio system records the book name, author information and thickness data, the book name, author information and thickness data are sent to the control system of the electric push plate through the API protocol of the OpenBiblio system, and the number of the electric push plate pushed out is adjusted according to the thickness; when the user queries the book in the OpenBiblio system, the OpenBiblio system returns the position information and thickness data of the book, after the control system of the electric push plate receives the data, the electric push plate is driven to perform the pushing operation, the target book is pushed to the area where the user can take the book, the high-resolution camera feeds back the book image to the control system of the electric push plate, so as to ensure the execution of the action of the electric push plate, and after the book is pushed to the specified position, the control system of the electric push plate confirms the completion of the action and returns the state information to the OpenBiblio system.

[0062] As shown in the figure, Figure 6 The scanner or the high-resolution camera acquires the book image, the OpenBiblio system performs image processing and extracts the features of the book, then performs feature analysis, calculates the thickness of the book, and transmits the thickness information of the book to the control system of the electric push plate, and adjusts the number of the push plate to push out the book.

[0063] As shown in the figure, Figure 1 The present application also provides a reading progress tracking method of the household intelligent bookcase, the household intelligent bookcase of the present application is integrated with the OpenBiblio system, and the household intelligent bookcase and the OpenBiblio system transmit data through a communication module, i.e. a Wi-Fi module, and specifically, the reading progress tracking method specifically comprises the following steps:

[0064] Step 1, place the book into the scanner area of the home smart bookcase, the scanner transmits the scanned book image to the OpenBiblio system, and the OpenBiblio system determines whether the book is open. The specific method is as follows: place the book in the scanner area, if the text coverage in the image recognized by the scanner, that is, the proportion of recognized text pixels in the total pixels, exceeds the set threshold value, such as 10% or 20%, it is judged that the image scanned by the scanner is a large amount of text, and it is considered that the book is open, then the precise mode is enabled, if the text coverage is lower than the set threshold value, it is judged that the image is a small amount of text, and it is considered that the book is closed, then the fuzzy mode is enabled.

[0065] Step 2, if it is judged that the book is in an open state, the precise mode is run, the scanner performs a first scan to scan the page number of the book at this time and record, when the user closes the book, the scanner performs a second scan, the OpenBiblio system determines whether the book is open, if the book is in an open state, the OpenBiblio system prompts the user to close the book, if the book is in a closed state, the scanner scans the book cover to record the book information, after the scanner completes the scan, the OpenBiblio system outputs the page number, book name and author information of the book reading up to.

[0066] Step 3, if the OpenBiblio system judges that the book is in a closed state after the scanner scans in step 1, the fuzzy mode is run, the high-resolution camera placed above the right scanning area of each layer of bookshelf in the bookcase scans the side of the book when it is closed, records the book name and author information, extracts the crease features on the spine through the scanner, uses image processing algorithm, that is, edge detection to identify and enhance the crease area, the features of the depth, width and position of the crease are obtained by calculating the gray difference and area size in the image, these data reflect the frequency of flipping, based on the features of the depth, width and position of the crease, the OpenBiblio system associates the crease depth with the reading progress through the machine learning algorithm, and estimates the approximate reading progress, after the scanning is completed, the OpenBiblio system outputs the book related information and the approximate reading progress. With continuous learning and optimization in the use process, the system can gradually improve the accuracy of the estimation, and correct it through user feedback, and finally realize accurate reading progress estimation.

[0067] In this step 1, image processing algorithm is used to identify and enhance the crease area, which includes the following steps:

[0068] Step 3.1, apply Gaussian blur to the scanned image to smooth the image and reduce noise, Gaussian processing can eliminate fine noise and enhance the recognizability of traces, which is very useful for traces with small details or unclear edges.

[0069] Step 3.2: Perform median filtering to remove random noise such as impulse noise, retain image edge information, and make the image edge clearer to facilitate subsequent contour detection;

[0070] Step 3.3, use the Sobel operator to detect the horizontal and vertical edges of the image, especially when the image is rich in details and texture. By calculating the horizontal and vertical gradients of the image, the edge position is found. Assume that the gray value of a pixel in the image is I(x ， y), the horizontal and vertical gradients at the pixel point are convolved, and then the horizontal and vertical gradient amplitudes G are calculated to represent the edge strength, and the gradient direction θ is calculated to represent the direction of the edge.

[0071] Horizontal direction G x Convolution kernel:

[0072]

[0073] Vertical direction G y Convolution kernel:

[0074]

[0075] Gradient Magnitude:

[0076]

[0077] For the convenience of calculation, the approximate formula is used:

[0078] G≈|G x |+|G y |

[0079] Gradient direction:

[0080]

[0081] The thickness of the electric push plate installed on each layer of the bookshelf is 1-10 mm, the electric push plate is located at the rear side of each layer of the bookshelf, the push plate has a small width, and the push plate can be flexibly combined according to the thickness of the books to ensure appropriate pushing force. A plurality of cameras are installed above or in front of each layer of the bookshelf to ensure that the books on each bookshelf can be clearly scanned. When a user selects a certain book, the user can accurately determine the position of the book. The cameras should be installed above or in front of the bookshelf, and the angle should be reasonably set to cover the horizontal and vertical areas of the entire bookshelf. In order to avoid dead angles, multiple cameras can be used to take pictures of the same layer of the bookshelf from different angles to ensure that the boundaries and arrangement of the books are fully captured. Assuming that the user has two storage states of the bookshelf, one is storing books and the other is taking books, the storing books state is divided into using the system to take books and directly placing books on the bookshelf, and the taking books state is divided into using the system to take books and directly taking books from the bookshelf. The environment of the bookshelf is relatively stable, and only when the user takes or stores books will there be a large change. Therefore, a high-resolution camera uses frame difference method to determine whether the current user's behavior is taking or storing books. The current frame captured by the camera of the bookshelf is compared with the previous background frame, and the background frame is a reference image of the bookshelf in a stationary state, which is automatically updated after a long time of stability.

[0082] As shown in Figure 2 , the frame difference method is used to determine whether the current user's behavior is taking or storing books. Specifically, the frame difference method detects changes in the environment of the bookshelf to determine whether the user is using the OpenBiblio system to store or take books. If the user is using the OpenBiblio system to store or take books, the control system of the electric push plate determines whether it is storing books. If it is storing books, the control system of the electric push plate binds the scanned book information and the position where the camera detects the environmental change, and stores the book information and position information in the OpenBiblio system. If it is not storing books, the system obtains the position information of the book, and pushes the book out using the electric push plate, and then updates the book information. If it is not using the system to store or take books, the frame difference method is used to determine whether it is storing books. If it is storing books, the control system generates a "position recognition book" record and number, and records the position. If it is not storing books, the control system determines whether there is book information in the control system. If there is book information, the control system updates the book state to "removed" and removes the book position information. If there is no book information, the control system deletes the number and position information together. When a user needs to take a book, the user inputs the required book on the screen, the control system automatically reads the book thickness obtained by the scanner or camera and selects the corresponding number of push plates, and then the control system controls the electric push plate to run and push the book out.

[0083] The present application associates the crease depth with the reading progress by a machine learning algorithm, and estimates the approximate reading progress: using a convolutional neural network to actively extract image features and classify, the CNN model can learn the relationship between different types of traces and book reading progress. Collect books with labeled reading progress as training data, use the convolutional layer of the convolutional neural network to extract the image feature map, the pooling layer reduces the size of the image feature map, the fully connected layer of the convolutional neural network classifies and outputs the reading progress, and the regional convolutional neural network locates the specific trace area in the image feature, helping to accurately extract certain areas of the book and judge whether it is related to the reading progress.

[0084] The home intelligent book cabinet uses a dynamic recommendation algorithm to introduce time factors and user behavior change factors to provide personalized and dynamic book recommendation services for users, specifically including the following steps:

[0085] Step 4.1, book attribute collection:

[0086] Book set B={B1, B2,..., B n}, record the initial category C(B i ), book initial classification label and the latest reading time T last (B i ), introduce time factor, that is, collect user behavior data X k , behavior data X k includes reading duration D read (B i ), unit: minutes, reading frequency F read (B i ), times / day, reading time period T segment (B i ), such as morning / afternoon / night behavior data;

[0087] Dynamic weight initialization: adjust the weights ω1, ω2, ω3 according to the user behavior change, the weight calculation formula:

[0088]

[0089] Among them, H(X k ) is the information entropy of behavior data X k , used to measure the importance of the data;

[0090] Step 4.2, calculate the priority score of each book:

[0091] P(B i )=ω1·D read (B i )+ω2·F read (B i)+ ω3· RecencyWeight(B i )

[0092]

[0093] where Δt(B i ) is the difference in days between the current time and the recency time T last (B i );

[0094] Step 4.3, calculate the probability of each book belonging to multiple categories, calculated by fuzzy membership function:

[0095]

[0096] where, is the membership degree of book B i to category c k , T is the temperature parameter, used to control the fuzziness of classification, updated every night to dynamically adjust the classification, books with greater than the threshold value are reclassified;

[0097] Step 4.4, target recommendation: combine the priority score P(B i ) of step 4.2 and the diversity target to get the multi-objective score:

[0098]

[0099] where Diversity(B j ) is the diversity of the book in user behavior, α, β control the trade-off between priority and diversity;

[0100] Sort by multi-objective score, select the top N books to generate recommendation list:

[0101] R = {B1, B2,... B N}

[0102] where Score(B1) > Score(B2) >... > Score(B N )

[0103] Step 4.5, task scheduling: set to run every night at 23:59: dynamic weight update, priority calculation, classification adjustment and recommendation list generation.

[0104] According to the reading behavior data such as reading frequency, browsing time, dynamically adjust the classification of books, especially in the case of changes in user interest, through the labeled directory system to effectively optimize the classification of books and improve the user's search efficiency.

[0105] The household intelligent book cabinet has a reading progress tracking function based on book side crease scanning, an auxiliary book taking function of a multi-module push plate, and intelligent classification using a dynamic classification recommendation algorithm. The book cabinet improves the convenience and intelligent level of book management through intelligent design, so that users can more efficiently manage books and track reading progress.

[0106] The above merely describes the embodiments of the present application and is not intended to limit the present application. The present application can be variously changed and modified by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of the claims of the present application.

Claims

1. A reading progress tracking method for a household smart bookcase, characterized by: The home smart bookcase is integrated with the OpenBiblio system and the home smart bookcase and the OpenBiblio system transmit data through a communication module. The reading progress tracking method specifically includes the following steps: Step 1: Place the book into the scanner area of ​​the home smart bookcase. The scanner transmits the scanned book image to the OpenBiblio system, which determines whether the book is open. Step 2: If the book is open, the scanner runs in precise mode and performs the first scan, scanning and recording the page number of the book at that time. When the user closes the book, the scanner performs the second scan, and the OpenBiblio system determines whether the book is open. If the book is open, the OpenBiblio system reminds the user to close the book. If the book is closed, the scanner scans the book cover and records the book information. After the scanner completes the scan, the OpenBiblio system outputs the page number, book title, and author information of the book that has been read so far. Step 3. If the OpenBiblio system determines that the book is closed after scanning with the scanner in step 1, it runs the fuzzy mode and places a high-resolution camera above each shelf inside the bookcase to scan the side of the book when it is closed, recording the book name and author information. The scanner extracts the crease features on the spine and uses image processing algorithms to identify and enhance the crease area. Based on the crease features, the OpenBiblio system associates the crease depth with the reading progress and infers the reading progress. After the scan is completed, the OpenBiblio system outputs book-related information and the approximate reading progress.

2. The reading progress tracking method for a household smart bookcase according to claim 1, characterized in that: In step 1, a book is placed in the scanner area to determine whether the book is open. The specific method is as follows: place the book in the scanner area. If the text coverage rate in the image recognized by the scanner, that is, the ratio of the recognized text pixels in the image to the total pixels, exceeds the set threshold, the image scanned by the scanner is judged to be a large amount of text, and the book is considered to be open. If the text coverage rate is lower than the set threshold, the image is judged to be a small amount of text, and the book is considered to be closed, and the fuzzy mode is enabled.

3. The reading progress tracking method for a household smart bookcase according to claim 1, characterized in that: Use image processing algorithms to identify and enhance crease areas, including the following steps: Step 3.

1. Apply Gaussian blur to the scanned image to smooth the image and reduce noise. Step 3.2: Perform median filtering to remove random noise, retain image edge information, and make the image edge clearer; Step 3.3: Use the Sobel operator to detect the horizontal and vertical edges of the image, and find the edge position by calculating the image gradient in the horizontal and vertical directions.

4. The reading progress tracking method for a household smart bookcase according to claim 3, characterized in that: In step 3.3, assuming that the grayscale value of a pixel in the image is I(x, y), the horizontal and vertical gradients at the pixel are convolved, and then the horizontal and vertical gradient amplitudes G are calculated to represent the edge strength, and the gradient direction θ is calculated to represent the direction of the edge. Horizontal direction G x Convolution kernel: Vertical direction G y Convolution kernel: Gradient Magnitude: For the convenience of calculation, the approximate formula is used: G≈|G x |+|G y | Gradient direction:

5. The reading progress tracking method for a household smart bookcase according to claim 1, characterized in that: In step 3, the system uses a machine learning algorithm to associate the fold depth with the reading progress and infer the approximate reading progress. Specifically, it collects book samples with marked reading progress as training data, uses the convolutional layer of the convolutional neural network to extract the image feature map, uses the pooling layer to reduce the size of the image feature map, uses the fully connected layer of the convolutional neural network to classify and output the reading progress, uses the regional convolutional neural network to locate the trace area in the image features, extracts the area of ​​the book and determines whether it is related to the reading progress.

6. A reading progress tracking method for a household smart bookcase according to any one of claims 1 to 5, characterized in that: The household intelligent bookcase comprises a display screen (3) arranged on the bookcase, a scanner (2) for scanning book information, Each shelf of the home smart bookcase is equipped with multiple independent electric push plates controlled by stepper motors. The electric push plates are set on the back side of the home smart bookcase and are used to push out the books selected by the user. The home smart bookcase is provided with several layers of bookshelves, and a high-resolution camera is provided above each layer of bookshelves inside the bookcase. The high-resolution camera is used to locate the position and changes of books and identify the storage status of the current books on the bookshelf, i.e., the book retrieval status and book storage status; When users store book information into the OpenBiblio system through a scanner or high-resolution camera, the OpenBiblio system records the book name, author information, and thickness data. The book name, author information, and thickness data are sent to the control system of the electric push plate through the OpenBiblio system's API protocol, and the number of electric push plates pushed out is adjusted according to the thickness. When a user searches for a book in the OpenBiblio system, the OpenBiblio system returns the location information and thickness data of the book. After receiving the data, the control system of the electric push plate drives the electric push plate to perform a pushing operation, pushing the target book to the user's accessible area. The high-resolution camera feeds back the book image to the control system of the electric push plate to ensure the execution of the electric push plate action. When the book is pushed to the designated position, the control system of the electric push plate confirms the completion of the action and returns status information to the OpenBiblio system.

7. The reading progress tracking method for a household smart bookcase according to claim 6, characterized in that: The home smart bookcase uses a dynamic recommendation algorithm to introduce time factors and user behavior change factors to provide users with personalized and dynamic book recommendation services, which specifically includes the following steps: Step 4.1: Book attribute collection: Book set B = {B1, B2, ..., B n }, record the initial category C(B i ), the initial classification label of the book and the most recent reading time T last (B i ), introducing the time factor, that is, collecting user behavior data X k , behavioral data X k Including reading time D read (B i ), reading frequency F read (B i ), reading time period T segment (B i ) behavioral data; Dynamic weight initialization: adjust weights ω1, ω2, ω3 according to changes in user behavior. Weight calculation formula: Among them, H(X k ) is the behavioral data X k Information entropy of Step 4.

2. Calculate the priority score for each book: P(B i )=ω1·D read (B i )+ω2·F read (B i )+ω3·RecencyWeight(B i ) Among them, Δt(B i ) is the current time and the recent reading time T last (B i ) days; Step 4.

3. Calculate the probability that each book belongs to multiple categories using the fuzzy membership function: in, For Book B i Belongs to category c k The membership degree, T is the temperature parameter, which is used to control the fuzziness of the classification. The update is run every night to dynamically adjust the classification. Books with values ​​greater than the threshold are reclassified; Step 4.4, target recommendation: Comprehensive priority score P(B i ) and diversity goals, resulting in a multi-objective score: Among them, Diversity (B j ) is the diversity of books in user behavior, α and β control the trade-off between priority and diversity; Sort by multi-objective scores and select the top N books for recommendation generation: R={B1,B2,...B N } Among them, Score(B1)>Score(B2)>…>Score(B N ) Step 4.5, Task Scheduling: Set fixed running time: dynamic weight update, priority calculation, classification adjustment and recommendation list generation.

8. The reading progress tracking method for a household smart bookcase according to claim 6, characterized in that: Several 1 cm thick electric push plates are installed on each layer of bookshelf. The electric push plates are located on the back side of each layer of bookshelf. The electric push plates are set on each layer of bookshelf, and several cameras are installed above or in front of each layer of bookshelf to ensure that the books on each bookshelf can be scanned clearly. Users have two storage states for bookshelves: storing books and retrieving books. The storing state is divided into using the system to retrieve books and directly placing books on the bookshelf. The retrieving state is divided into using the system to retrieve books and directly taking books from the bookshelf. Changes will occur when users retrieve and store books. The high-resolution camera uses the frame difference method to determine whether the current user's behavior is to retrieve or store books. The current frame captured by the bookshelf camera is compared with the previous background frame. The background frame is a reference image of the bookshelf in a static state. It is automatically updated after a long period of stability. The frame difference method is used to determine whether the current user's behavior is to retrieve or store books. Specifically: the frame difference method detects changes in the bookshelf environment and determines whether the user uses the OpenBiblio system to store or retrieve books. If the user uses the OpenBiblio system to store or retrieve books, When the enBiblio system is used to deposit or retrieve a book, the control system of the electric push plate determines whether the book is being deposited. If so, the control system of the electric push plate binds the scanned book information with the location where the camera detects a change in the environment, and stores the book information and location information together in the OpenBiblio system. If the book is not being deposited, the OpenBiblio system obtains the location information of the book, transmits the information to the control system of the electric push plate through the communication module, and pushes the book out using the electric push plate, and then the system updates the book information. If the system is not used to deposit or retrieve a book, the frame difference method is used to determine whether the book is being deposited. If so, the control system generates a "location identification book" record and number, and records the location. If the book is not being deposited, it determines whether there is book information at this location in the control system. If there is book information, the control system updates the book status to "removed" and removes the book location information. If there is no book information, the control system deletes the number and location information together. When the user needs to take out a book, he enters the desired book on the screen. The control system automatically reads the thickness of the book obtained by the scanner or camera and selects the corresponding number of push plates. The control system then controls the electric push plates to push the book out.

Images