CN116992067A - A digital display system and method for intangible cultural heritage - Google Patents

Abstract

The invention discloses a non-heritable cultural heritage digital display system and a method, which belong to the technical field of cultural protection and comprise a user terminal, a display platform, a non-heritable data collection module, a classification marking module, a feature extraction module, a 3D reconstruction module, an intelligent correction module, a voice broadcasting module and a block storage module, wherein the user terminal is used for being in communication connection with the display platform, displaying 3D scene animation and providing an interactive browsing function; according to the invention, the randomness of the training sample can be improved, the accuracy of anomaly detection can be effectively improved, meanwhile, the 3D model with errors can be automatically corrected, the manual correction time is saved, the workload of staff is reduced, the use experience of the staff is improved, the non-heritage cultural heritage can be accurately displayed for the user, and the user's look and feel is increased.

Description

A digital display system and method for intangible cultural heritage

Technical field

The present invention relates to the technical field of cultural protection, and in particular to a digital display system and method for intangible cultural heritage.

Background technique

Intangible cultural heritage, referred to as "intangible cultural heritage", is opposite to "material cultural heritage". In China, intangible cultural heritage refers to various traditional cultural expressions that are passed down from generation to generation by people of all ethnic groups and regarded as part of their cultural heritage, as well as physical objects and places related to traditional cultural expressions. Intangible cultural heritage is the most dynamic and important part of cultural diversity. It is the crystallization of human civilization and the most precious common wealth. It carries human wisdom, civilization and glory of human history. In the process of globalization and the rapid development of modernization in the world, the introduction of the concept of intangible cultural heritage is a historical necessity and a requirement of the times. It complies with the needs of historical development and has practical significance for the effective protection of intangible cultural heritage.

After searching, Chinese patent number CN109144275A discloses a digital display system and method for intangible cultural heritage. Although this invention enables in-depth communication with visitors and makes them empathize with each other, deepening the shock and impression of intangible cultural heritage, it Only erroneous 3D models can be corrected manually, which increases the workload of the staff and takes a lot of time. It cannot accurately display the intangible cultural heritage to users and reduces the user's perception. To this end, we propose an intangible cultural heritage Heritage digital display system and method.

Contents of the invention

The purpose of the present invention is to propose a digital display system and method for intangible cultural heritage in order to solve the defects existing in the existing technology.

In order to achieve the above objects, the present invention adopts the following technical solutions:

A digital display system for intangible cultural heritage, including a user terminal, a display platform, an intangible cultural heritage data collection module, a classification labeling module, a feature extraction module, a 3D reconstruction module, an intelligent correction module, a voice broadcast module and a block storage module;

The user terminal is used to communicate with the display platform, display 3D scene animation, and provide interactive browsing functions;

The display platform is used to reproduce intangible cultural heritage in 3D and feed back the generated 3D model to the user terminal;

The intangible cultural heritage data collection module is used to collect known intangible cultural heritage data;

The classification and labeling module is used to perform category analysis and label the collected intangible cultural heritage data;

The feature extraction module is used to extract feature information from intangible cultural heritage data;

The 3D reconstruction module is used to reproduce the corresponding 3D scene and the intangible cultural heritage production process based on the characteristic information;

The intelligent correction module is used to analyze intangible cultural heritage data and detect and adjust the 3D model established by the 3D reconstruction module;

The voice broadcast module is used to introduce relevant intangible cultural heritage information to visitors;

The block storage module is used to store various collected intangible cultural heritage data.

As a further solution of the present invention, the user terminal specifically includes a smart network, a tablet computer, a desktop computer and a notebook computer.

As a further solution of the present invention, the specific steps of category analysis of the classification marking module are as follows:

Step 1: After receiving each group of intangible cultural heritage data, the classification marking module will classify each group of heritage data according to the prescribed intangible cultural heritage list system, and at the same time classify the classified heritage data according to its specific category;

Step 2: Sort each group of heritage data in the order of the initial letters A to Z of the name of each intangible cultural heritage. At the same time, classify each group of heritage data information that has been sorted according to text information, picture information, and video information.

As a further solution of the present invention, the classification standards of the intangible cultural heritage list system described in step one are specifically countries, provinces, cities and counties;

The category of intangible cultural heritage specifically includes traditional oral literature and the language as its carrier, traditional art, calligraphy, music, dance, drama, folk arts and acrobatics, traditional skills, medicine and calendar, traditional etiquette, festival folk customs, traditional sports and Entertainment and other intangible cultural heritage.

As a further solution of the present invention, the specific extraction steps of feature information of the feature extraction module are as follows:

Step ①: The feature extraction module collects the video information and picture information in each heritage data, and then processes the video information frame by frame to obtain multiple groups of picture information, and then performs block processing according to the display ratio of each group of picture information, and then performs block processing The high-frequency components of the image information are removed through Fourier transform;

Step ②: Set the parameters of the Gaussian smoothing filter, and calculate the weight of each filter, and then smooth the image information through each weighted heterogeneous filter, and perform nonlinear transformation on the processed image. The sum of the results obtained by nonlinear transformation is weighted to obtain the final picture information;

Step ③: Select a window that meets the conditions and move it in each group of picture information. Calculate the gray-level co-occurrence matrix under the window each time it moves, and calculate the texture features in the relevant picture information from the gray-level co-occurrence matrix. Then, based on the calculated Texture features separate the target from the background, normalize the target image, and extract feature data of the target image;

Step 4: ShuffleBlock multiple times through the collection network to obtain the global posture features of each group of image information, and then use the deconvolution operation to return the global posture features to the key point feature map, and then decode the key point feature map and collect The two-dimensional key points of the human body generated after decoding.

As a further solution of the present invention, the specific steps for detection and adjustment of the intelligent correction module are as follows:

Step Ⅰ: The intelligent correction module receives each set of feature data and key point information to construct a detection data set, then calculates the standard deviation of the detection data set, and filters out abnormal data in the detection data set based on the calculated standard deviation;

Step Ⅱ: Input the detection data set into the convolutional neural network. First perform a convolution operation to obtain a feature map that meets the requirements. Repeat the structure of two sets of convolution layers and a set of max pooling layers to process the feature map, and then expand it. The deconvolution operation is performed in the channel to halve the dimension of the feature map, and then the feature map is reorganized into a feature map with 2 times the dimension, and then two sets of convolution layers are used, and the structure is repeated, and then the previous layer is added to the final output layer. The obtained feature map is mapped into a 6-dimensional output feature map, the output feature map obtained through forward propagation is collected, and then the linear prediction values of all targets are converted into probability values through the softmax function;

Step III: Use the loss function to calculate the loss value between the real data and the detection probability, then update the parameters in the convolutional neural network layer by layer, and calculate the corresponding loss value. When the loss value reaches a certain threshold, stop training and change the parameter as optimal parameters and output the modified model;

Step IV: Input the 3D model into the correction model, set the detection label on the correction model, and then perform convolution, pooling and full connection processing on the 3D model to confirm the location of the key points of the abnormal model and correct it.

A digital display method of intangible cultural heritage. The display method is as follows:

(1) Collect intangible cultural heritage data and classify it in accordance with regulations;

(2) Extract characteristic information of various types of intangible cultural heritage and build 3D models;

(3) Generate text descriptions of various types of intangible cultural heritage and match them with 3D models;

(4) Correct the 3D model and provide feedback to the staff on the model correction results;

(5) The user logs in to the display platform and selects the relevant 3D model to view;

(6) Store 3D models of each group of intangible cultural heritage and optimize the performance of the display platform.

As a further solution of the present invention, the specific steps for optimizing the performance of the display platform described in step (6) are as follows:

Step Ⅰ: Generate a startup linked list for each group of functional interfaces of the display platform, and according to the order of the LRU linked list, further link each group of startup linked lists according to the number of visits from least to most;

Step Ⅱ: Update the data of each group of pages in each group of startup link lists in real time based on the interaction information of each group of functional interfaces, and select the functional interface startup link list with the least number of visits from the head of the LRU chain list to select the victim page. Stop until enough victim pages are recovered;

Step III: Combine the selected victim pages into a block and mark it, then wake up a compression driver to parse the marked block, obtain the physical page belonging to the block, and then copy the physical page to the buffer. Then the compression algorithm is called to compress the physical page in the buffer into a compressed block, and the compressed block is stored in the compression area of the platform optimization module.

Compared with the existing technology, the beneficial effects of the present invention are:

1. The present invention receives each set of feature data and key point information through the intelligent correction module to construct a detection data set. After preprocessing the detection data set, it is input into the convolutional neural network for convolution, pooling and After deconvolution, the feature map is output, and the output feature map obtained by forward propagation is collected, and then the linear prediction values of all targets are converted into probability values through the softmax function, and the loss function is used to calculate the loss between the real data and the detection probability. value, then update the parameters in the convolutional neural network layer by layer, and calculate the corresponding loss value. When the loss value reaches a certain threshold, the training will stop. At the same time, the parameter will be regarded as the optimal parameter and the correction model will be output. The 3D model will be input into the correction model. , correct the model to set the detection label, and then perform convolution, pooling and full connection processing on the 3D model to confirm the location of the key points of the abnormal model and correct it, which can improve the randomness of the training samples and effectively improve the accuracy of anomaly detection. , and at the same time, it can correct 3D models with errors by itself, saving manual correction time, reducing staff workload, improving staff experience, accurately displaying intangible cultural heritage to users, and increasing user perception.

Description of drawings

The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

Figure 1 is a system block diagram of a digital display system for intangible cultural heritage proposed by the present invention;

Figure 2 is a flow chart of a digital display method for intangible cultural heritage proposed by the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, rather than all the embodiments.

Example 1

Referring to Figure 1, a digital display system for intangible cultural heritage includes a user terminal, a display platform, an intangible cultural heritage data collection module, a classification labeling module, a feature extraction module, a 3D reconstruction module, an intelligent correction module, a voice broadcast module and a block storage module.

The user terminal is used to communicate with the display platform, display 3D scene animations, and provide interactive browsing functions; the display platform is used to reproduce intangible cultural heritage in 3D and feed back the generated 3D model to the user terminal.

It should be further explained that user terminals specifically include smart networks, tablet computers, desktop computers and laptop computers.

The intangible cultural heritage data collection module is used to collect known intangible cultural heritage data; the classification labeling module is used to perform category analysis and label the collected intangible cultural heritage data.

Specifically, after receiving each group of intangible cultural heritage data, the classification marking module classifies each group of heritage data according to the prescribed intangible cultural heritage list system. At the same time, it classifies the classified heritage data according to its specific category, and classifies each group of heritage data according to its specific category. The heritage data is sorted according to the order of the first letters of each intangible cultural heritage name, A to Z, and at the same time, each group of heritage data information that has been sorted is classified according to text information, picture information, and video information.

In this embodiment, the classification standards of the intangible cultural heritage list system are specifically countries, provinces, cities, and counties; the categories of intangible cultural heritage specifically include traditional oral literature and its carrier language, traditional art, calligraphy, music, dance, and drama , folk arts and acrobatics, traditional skills, medicine and calendar, traditional etiquette, festival folk customs, traditional sports and entertainment and other intangible cultural heritage.

The feature extraction module is used to extract feature information from intangible cultural heritage data.

Specifically, the feature extraction module collects video information and picture information in each heritage data, and then processes the video information frame by frame to obtain multiple groups of picture information, and then performs block processing according to the display ratio of each group of picture information, and then performs block processing on The high-frequency components of the picture information are removed through Fourier transform, the Gaussian smoothing filter parameters are set, and the weights of each filter are calculated, and then the picture information is smoothed through each weighted heterogeneous filter. The processed image is subjected to non-linear transformation, and the sum of the results obtained by the non-linear transformation is weighted to obtain the final picture information. A window that meets the conditions is selected to move in each group of picture information, and the gray value under the window at this time is calculated each time it is moved. degree co-occurrence matrix, and calculate the texture features in the relevant image information from the gray-level co-occurrence matrix, and then separate the target and the background based on the calculated texture features, normalize the target image, and extract the target image features at the same time The data is collected through the network ShuffleBlock multiple times to obtain the global posture features of each group of picture information, and then the global posture features are returned to the key point feature map through the deconvolution operation, and then the key point feature map is decoded and collected. The two-dimensional key points of the human body are then generated.

The 3D reconstruction module is used to reproduce the corresponding 3D scene and the intangible cultural heritage production process based on the characteristic information; the intelligent correction module is used to parse the intangible cultural heritage data and detect and adjust the 3D model established by the 3D reconstruction module.

Specifically, the intelligent correction module receives each set of feature data and key point information to construct a detection data set, and then calculates the standard deviation of the detection data set, and filters out abnormal data in the detection data set based on the calculated standard deviation, and converts the detection data set into When input into the convolutional neural network, a convolution operation is first performed to obtain a feature map that meets the requirements. The structure of two sets of convolutional layers and a set of maximum pooling layers is repeatedly used to process the feature map, and then a deconvolution operation is performed in the expansion channel. Halve the dimension of the feature map, then re-form the feature map with 2 times the dimension, then use two sets of convolutional layers, and repeat the structure, and then map the feature map obtained from the previous layer into 6 in the final output layer dimensional output feature map, collect the output feature map obtained through forward propagation, and then convert the linear prediction values of all targets into probability values through the softmax function, use the loss function to calculate the loss value between the real data and the detection probability, and then gradually The layer updates the parameters in the convolutional neural network and calculates the corresponding loss value. When the loss value reaches a certain threshold, the training stops. At the same time, the parameter is used as the optimal parameter and the corrected model is output. The 3D model is input into the corrected model and the model settings are corrected. Detect labels, then perform convolution, pooling and full connection processing on the 3D model to confirm the location of key points of the abnormal model and correct them.

The voice broadcast module is used to introduce relevant intangible cultural heritage information to visitors; the block storage module is used to store the collected intangible cultural heritage data.

Example 2

Referring to Figure 2, a digital display method of intangible cultural heritage is shown. The display method is as follows:

Collect intangible cultural heritage data and classify it in accordance with regulations.

Extract characteristic information of various types of intangible cultural heritage and build 3D models.

Generate various types of intangible cultural heritage text descriptions and match them with 3D models.

Correct the 3D model and report the model correction results to the staff.

Users log in to the display platform and select relevant 3D models to view.

Store 3D models of each group of intangible cultural heritage and optimize the performance of the display platform.

Specifically, a startup linked list is generated for each group of functional interfaces of the display platform, and according to the order of the LRU chain list, each group of startup linked lists is further linked according to the number of visits from least to most, and real-time matching is performed based on the interactive information of each group of functional interfaces. Each group starts each group of pages in the linked list to update data, and selects the functional interface with the least number of visits from the head of the LRU linked list to start the linked list to select victim pages. It stops after recycling enough victim pages, and the selected victim pages are Merge into a block and mark it, then wake up a compression driver to parse the marked block and obtain the physical page belonging to the block, copy the physical page to the buffer, and then call the compression algorithm to The physical page is compressed into compressed blocks, and the compressed blocks are stored in the compression area of the platform optimization module.

Claims (8)

1. The non-heritable heritage digital display system is characterized by comprising a user terminal, a display platform, a non-heritable data collection module, a classification marking module, a feature extraction module, a 3D reconstruction module, an intelligent correction module, a voice broadcasting module and a block storage module;

the user terminal is used for being in communication connection with the display platform, displaying 3D scene animation and providing an interactive browsing function;

the display platform is used for 3D reproduction of non-heritable heritage and feeding back the generated 3D model to the user terminal;

the non-genetic data collection module is used for collecting known non-genetic cultural heritage data;

the classification marking module is used for carrying out classification analysis on the collected non-heritable cultural heritage data and marking the collected non-heritable cultural heritage data;

the feature extraction module is used for extracting feature information in the non-heritable cultural heritage data;

the 3D reconstruction module is used for reproducing the corresponding 3D scene and the non-genetic culture making process according to the characteristic information;

the intelligent correction module is used for analyzing the non-heritable cultural heritage data and detecting and adjusting the 3D model established by the 3D reconstruction module;

the voice broadcasting module is used for introducing various information of related non-heritable cultural heritage to visitors;

the block storage module is used for storing various collected non-heritable cultural heritage data.

2. The non-heritable heritage digital display system of claim 1, wherein said user terminal specifically comprises an intelligent collection, a tablet computer, a desktop computer, and a notebook computer.

3. The non-heritable heritage digital display system of claim 1, wherein said classification marking module category analysis comprises the specific steps of:

step one: after receiving each group of non-heritable cultural heritage data, the classifying and marking module classifies each group of heritage data according to a specified non-material cultural heritage directory system, and classifies the classified heritage data according to specific categories of the heritage data;

step two: sorting the groups of heritage data according to the order of the first letters A to Z of the non-heritage names, and classifying the sorted groups of heritage data information according to the text information, the picture information and the video information.

4. The system of claim 3, wherein the non-material cultural heritage directory system classification criteria of step one are country, province, city and county;

the non-heritable cultural heritage category specifically comprises traditional spoken literature and languages, traditional art, handwriting, music, dance, drama, melody and acrobatics, traditional skills, medicine and calendar, traditional etiquette, festival celebration folks, traditional sports and recreation and other non-matter cultural heritages as carriers thereof.

5. The non-heritable heritage digital display system of claim 3, wherein said feature extraction module feature information specifically extracts the following steps:

step (1): the feature extraction module collects video information and picture information in each heritage data, then processes the video information frame by frame to obtain a plurality of groups of picture information, then performs blocking processing according to the display proportion of each group of picture information, and then removes high-frequency components in the blocked picture information through Fourier transformation;

step (2): setting Gaussian smoothing filter parameters, calculating weight occupied by each filter, smoothing the picture information through each weighted anisotropic filter, carrying out nonlinear transformation on the processed image, and carrying out weighting treatment on the sum of results obtained by the nonlinear transformation to obtain final picture information;

step (3): selecting windows meeting the conditions to move in each group of image information, calculating gray level co-occurrence matrixes under the windows at each time, calculating texture features in related image information from the gray level co-occurrence matrixes, separating a target from a background according to the calculated texture features, carrying out normalization processing on a target image, and extracting feature data of the target image;

step (4): the global attitude features of each group of image information are obtained through a plurality of times of ShuffleBlock of an acquisition network, then the global attitude features are returned to the key point feature map through deconvolution operation, then the key point feature map is decoded, and two-dimensional key points of the human body generated after decoding are collected.

6. The non-heritable heritage digital display system of claim 5, wherein said intelligent correction module detects and adjusts the specific steps as follows:

step I: the intelligent correction module receives the characteristic data and the key point information of each group to construct a detection data set, calculates the standard deviation of the detection data set, and screens out abnormal data in the detection data set according to the calculated standard deviation;

step II: inputting the detection data set into a convolutional neural network, performing convolutional operation to obtain a feature map meeting requirements, repeatedly adopting two groups of convolutional layers and a group of maximum pooling layers to process the feature map, performing deconvolution operation in an expansion channel to halve the dimension of the feature map, re-forming a feature map with 2 times of the dimension, adopting two groups of convolutional layers, repeating the structure, mapping the feature map obtained in the last layer into a 6-dimensional output feature map at a final output layer, collecting the output feature map obtained through forward propagation, and converting linear prediction values of all targets into probability values through a softmax function;

step III: calculating a loss value between real data and detection probability by using a loss function, updating parameters in the convolutional neural network layer by layer, calculating corresponding loss values, stopping training when the loss value reaches a certain threshold value, taking the parameters as optimal parameters, and outputting a correction model;

step IV: inputting the 3D model into a correction model, setting a detection label by the correction model, convoluting, pooling and fully connecting the 3D model, confirming the key point positions of the abnormal model, and correcting the abnormal model.

7. The non-heritable heritage digital display method is characterized by comprising the following steps of:

(1) Collecting non-heritable heritage data and classifying according to the regulations;

(2) Extracting various non-heritable heritage characteristic information and constructing a 3D model;

(3) Generating various non-heritable heritage literal descriptions and matching with the 3D model;

(4) Correcting the 3D model and feeding back a correction result of the staff model;

(5) The user logs in the display platform and selects a related 3D model for viewing;

(6) And storing each group of non-heritable heritage 3D models and optimizing and displaying the performance of the platform.

8. The method for digitally displaying non-heritable heritage according to claim 7, wherein the specific step of optimizing the performance of the display platform in the step (6) is as follows:

step I: generating a starting chain table for each group of functional interfaces of the display platform, and further linking each group of starting chain tables according to the sequence of the LRU chain tables from less to more according to the number of times each group is accessed;

step II: according to the interactive information of each group of functional interfaces, updating data of each group of pages in each group of starting linked lists in real time, sequentially selecting the functional interface starting linked list with the least accessed times from the head of the LRU linked list to select the victim page, and stopping until enough victim pages are recovered;

and III, step III: combining the selected victim page into a block and marking, waking up a compression driver program to analyze the marked block, obtaining a physical page belonging to the block, copying the physical page into a buffer area, then calling a compression algorithm to compress the physical page in the buffer area into a compression block, and storing the compression block into a compression area of a platform optimization module.