Integrated method and system for data compression and efficient transmission of multi-modal sensor network, and storage medium

Abstract

The invention discloses a multi-mode sensor network data compression and efficient transmission integration method and system and a storage medium, and relates to the field of data processing. Extracting a feature vector of each modal data in the multi-modal data sample, and generating a plurality of feature vector sets according to different modals; the data transmission end receives the plurality of feature vector sets, performs correlation analysis on feature vectors in different feature vector sets, completes feature selection and marks correlation analysis results to obtain feature subsets; compressing the feature vectors based on the feature subsets; and the data receiving end performs initial decompression on the type of the compressed modal data, recovers the compressed modal data into an initial feature vector according to a correlation analysis labeling result, and further recovers the compressed modal data into original multi-modal data. The multi-modal sensing network data compression and transmission method can stably operate in various complex application scenes, and can effectively realize compression and transmission of multi-modal sensing network data in the fields of industrial monitoring, intelligent transportation, medical health and the like.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and more particularly to an integrated method, system and storage medium for multi-modal sensor network data compression and efficient transmission. Background Art

[0002] In today's digital age, multi-modal sensor networks, with their advantage of being able to collect multiple types of data simultaneously (such as temperature, humidity, images, sounds, etc.), are widely used in many fields such as intelligent transportation, industrial monitoring, environmental monitoring, and healthcare. For example, in an intelligent transportation system, a multi-modal sensor network can combine camera image recognition, vehicle sensor data, and road condition monitoring data to achieve precise regulation of traffic flow; in the healthcare field, by integrating physiological parameter sensor data and medical imaging data, more comprehensive diagnostic basis can be provided for doctors.

[0003] However, with the continuous expansion of the scale of multi-modal sensor networks and the increasing complexity of application scenarios, the amount of data generated by them has grown explosively. On the one hand, the huge amount of data poses extremely high requirements for transmission bandwidth. Under the condition of limited network bandwidth, directly transmitting the original data will lead to network congestion and increased transmission delay, seriously affecting the real-time performance and reliability of the system. For example, in telemedicine, if high-resolution medical images are not processed during real-time transmission, it is very easy to experience lags and delay the diagnosis time. On the other hand, storing a large amount of unprocessed data also requires consuming huge storage space and computing resources.

[0004] Although traditional data compression methods can reduce the amount of data to a certain extent, they often do not fully consider the characteristics of multi-modal data and the actual requirements during the transmission process, resulting in deficiencies in aspects such as transmission efficiency, data integrity, and compatibility with transmission protocols for the compressed data. At the same time, independent data compression and transmission strategies cannot achieve the collaborative optimization of the two, and it is difficult to meet the strict requirements of multi-modal sensor networks for data processing efficiency and real-time performance. Therefore, researching an integrated method for multi-modal sensor network data compression and efficient transmission has important practical significance and urgent application needs. Summary of the Invention

[0005] In view of this, the present invention provides an integrated method, system and storage medium for multi-modal sensor network data compression and efficient transmission, which can operate stably in various complex application scenarios. Whether in the fields of industrial monitoring, intelligent transportation or healthcare, etc., it can effectively achieve the compression and transmission of multi-modal sensor network data.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] An integrated method for data compression and efficient transmission in a multimodal sensing network, comprising the following steps:

[0008] The data acquisition end acquires data from different modal sensors and performs preprocessing as multimodal data samples, extracts the feature vectors of each modal data in the multimodal data samples, and generates multiple feature vector sets according to different modalities;

[0009] The data transmission end receives multiple feature vector sets, respectively performs correlation analysis on the feature vectors in different feature vector sets; completes feature selection according to the correlation analysis results and annotates the correlation analysis results to obtain a feature subset; compresses the feature vectors based on the feature subset; selects different transmission strategies according to the transmission network and different modal data;

[0010] The data receiving end performs initial decompression on the compressed modal data type, and restores it to the initial feature vector according to the correlation analysis annotation result, and then restores it to the original multimodal data.

[0011] Optionally, the data receiving end identifies the transmission strategy, obtains the modality of the transmitted data and the transmission network, and judges the accuracy of the data modality of the decompressed data result according to the transmission modality and the transmission network; when there is a problem with the accuracy, re-combines the transmission strategy to decompress the transmitted data to complete the initial data decompression.

[0012] Optionally, the data acquisition end further includes encoding the feature vector set to generate an encoded vector.

[0013] Optionally, the data transmission end further includes receiving the encoded vector, encrypting the encoded vector using the AES encryption algorithm, compressing the encrypted encoded vector, and completing data transmission according to the data volume of the encoded vector.

[0014] Optionally, the data receiving end decrypts the decrypted data using the key to obtain the encoded vector, decodes the encoded vector to generate a feature vector, and obtains the original multimodal data according to the mapping relationship between the feature vector and the data to be transmitted.

[0015] An integrated system for data compression and efficient transmission in a multimodal sensing network, comprising:

[0016] A data acquisition module: used to acquire data from different modal sensors and perform preprocessing as multimodal data samples, extract the feature vectors of each modal data in the multimodal data samples, and generate multiple feature vector sets according to different modalities;

[0017] Data transmission module: It is used to receive multiple sets of feature vectors, perform correlation analysis on the feature vectors in different sets of feature vectors respectively; complete feature selection according to the results of correlation analysis and label the results of correlation analysis to obtain a feature subset; compress the feature vectors based on the feature subset; select different transmission strategies according to different transmission networks and modal data.

[0018] Data receiving module: It is used to perform initial decompression on the compressed modal data type, and restore it to the initial feature vectors according to the labeling results of correlation analysis, and then restore it to the original multimodal data.

[0019] A computer storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it realizes the steps of an integrated method for multi-modal sensing network data compression and efficient transmission as described in any one of the above.

[0020] As can be seen from the above technical solutions, compared with the prior art, the present invention provides an integrated method, system and storage medium for multi-modal sensing network data compression and efficient transmission, and has the following beneficial effects:

[0021] 1. During the process of compressing data, instead of blindly reducing the amount of data, feature selection is carried out according to the results of correlation analysis, which can ensure the retention of key features and important information in multimodal data. In this way, when decompressing and restoring at the data receiving end, the original multimodal data can be restored as accurately as possible, ensuring the availability and reliability of the data.

[0022] 2. The integrated design enables this method to process sensor data of multiple different modalities simultaneously, and can be flexibly adjusted according to different transmission networks and the characteristics of modal data. By reducing the amount of data and optimizing the transmission strategy, the computational burden and transmission delay of the system are reduced, thereby improving the performance of the entire multi-modal sensing network system. At the same time, the data receiving end can accurately restore the original multi-modal data, ensuring the accuracy of system data processing and analysis, and providing strong support for subsequent decision-making and applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for describing the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0024] Figure 1 It is a schematic flowchart of the present invention;

[0025] Figure 2This is a schematic structural diagram of the present invention. Detailed implementation manners

[0026] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0027] The embodiment of the present invention discloses an integrated method for multi-modal sensor network data compression and efficient transmission, as Figure 1 shown, including the following steps:

[0028] The data acquisition end acquires data from different modal sensors and performs preprocessing as multi-modal data samples, extracts the feature vectors of each modal data in the multi-modal data samples, and generates multiple feature vector sets according to different modalities;

[0029] The data transmission end receives multiple feature vector sets, respectively performs correlation analysis on the feature vectors in different feature vector sets; completes feature selection according to the correlation analysis results and annotates the correlation analysis results to obtain a feature subset; compresses the feature vectors based on the feature subset; selects different transmission strategies according to the transmission network and the different modal data;

[0030] The data receiving end performs initial decompression on the compressed modal data type, and restores it to the initial feature vectors according to the correlation analysis annotation results, and then restores it to the original multi-modal data.

[0031] In the method of this embodiment, after compression and efficient transmission, multi-modal data can be more conveniently fused and analyzed. The data of different modalities complement each other, can provide more comprehensive and accurate information, and provide a better basis for various multi-modal data analysis and applications. For example, in the field of intelligent security, comprehensive analysis by combining video surveillance and audio monitoring data can more accurately identify abnormal events and behaviors. The application of this method can bring new development opportunities to various fields relying on multi-modal sensor networks. Through the efficient processing and transmission of data from multiple sensors on the production line, more accurate production process monitoring and optimization can be achieved; in the medical field, the effective utilization of multi-modal medical data helps to improve the accuracy of disease diagnosis and treatment effects.

[0032] Furthermore, the data receiving end identifies the transmission strategy to obtain the modality of the transmitted data and the transmission network, and judges the accuracy of the data modality of the decompressed data result according to the transmission modality and the transmission network; when there is a problem with the accuracy, the transmitted data is decompressed again in combination with the transmission strategy to complete the initial decompression of the data.

[0033] Further, the data acquisition end further includes encoding the feature vector set to generate an encoded vector.

[0034] Further, the data transmission end further includes receiving the encoded vector, encrypting the encoded vector using the AES encryption algorithm, compressing the encrypted encoded vector, and completing data transmission according to the data volume of the encoded vector.

[0035] Furthermore, in the data transmission end, when the data acquisition end completes the preprocessing of different modality sensor data and extracts the feature vectors of each modality data, and generates multiple feature vector sets according to different modalities, the data transmission end starts to receive these sets. In an actual multi-modal sensing network, the feature vector sets of different modalities may be transmitted through different communication links or protocols. In this embodiment, decision trees, random forests, etc. can be used to evaluate the correlation between feature vectors. These algorithms can judge the degree of association between features by calculating the importance scores of features. For example, in a multi-modal sensing network for industrial production, the random forest algorithm can analyze the influence degree of the feature vectors collected by different sensors on product quality, so as to determine the correlation between features.

[0036] After feature selection, it is necessary to annotate the correlation analysis results. The annotation information can include the association metric value between feature vectors, the basis for feature selection, etc. The purpose of annotation is to accurately restore the original feature vectors using this information when decompressing and recovering data at the data receiving end. The annotation information can be stored in the form of metadata. For example, the annotation information is stored in JSON format and transmitted to the data receiving end together with the feature subset.

[0037] Regarding the transmission strategy, different modality data has different characteristics, such as real-time requirements, data volume size, etc., and different transmission strategies also need to be selected according to these characteristics.

[0038] Modality data with high real-time requirements: For some modality data with high real-time requirements, such as video, audio, etc., a low-latency transmission strategy needs to be adopted. The UDP protocol can be used for transmission because the UDP protocol has a low transmission latency, but it does not guarantee the reliable transmission of data. Therefore, some application-layer retransmission mechanisms need to be combined to ensure the accuracy of data. For example, in a multi-modal sensing network for a real-time video conference, the video data needs to be transmitted to the terminals of each participant in real time. Adopting the UDP protocol combined with the packet loss retransmission mechanism can meet the requirements of real-time and accuracy.

[0039] Modal data with a large amount of data: For modal data with a large amount of data, such as high-resolution images, long-term monitoring data, etc., a batch transmission strategy can be adopted. The data is cached and grouped at the data transmission end, and then batch transmission is performed when the network is idle to make full use of the network bandwidth. For example, in a multi-modal sensing network for meteorological monitoring, a large amount of meteorological data collected by meteorological stations can be batch-transmitted when the network bandwidth is relatively idle at night.

[0040] Further, the data receiving end uses the key to decrypt the decrypted data to obtain the encoded vector, decodes the encoded vector, generates the feature vector, and obtains the original multi-modal data according to the mapping relationship between the feature vector and the data to be transmitted.

[0041] associated with Figure 1 corresponding to the method shown, the present invention also discloses an integrated system for data compression and efficient transmission in a multi-modal sensing network for Figure 1 implementation of the method, the specific structure is as Figure 2 shown, including:

[0042] Data acquisition module: used to obtain data from different modal sensors and perform preprocessing as multi-modal data samples, extract the feature vectors of each modal data in the multi-modal data samples, and generate multiple feature vector sets according to different modalities;

[0043] Data transmission module: used to receive multiple feature vector sets, respectively perform correlation analysis on the feature vectors in different feature vector sets; complete feature selection according to the correlation analysis results and label the correlation analysis results to obtain a feature subset; compress the feature vectors based on the feature subset; select different transmission strategies according to the transmission network and different modal data;

[0044] Data receiving module: used to perform initial decompression on the compressed modal data type, and restore it to the initial feature vector according to the correlation analysis annotation result, and then restore it to the original multi-modal data.

[0045] This embodiment also discloses a computer storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of any one of the integrated methods for data compression and efficient transmission in a multi-modal sensing network.

[0046] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0047] The foregoing description of the disclosed embodiments enables those skilled in the art to practice or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An integrated method for multi-modal sensing network data compression and efficient transmission, characterized in that Including the following steps: The data acquisition end acquires data from different modality sensors and preprocesses it as multi-modal data samples, extracts the feature vectors of each modality data in the multi-modal data samples, and generates multiple feature vector sets according to different modalities; The data transmission end receives multiple feature vector sets, respectively performs correlation analysis on the feature vectors in different feature vector sets; completes feature selection according to the correlation analysis results and annotates the correlation analysis results to obtain a feature subset; Based on The feature subset compresses the feature vectors; different transmission strategies are selected according to the transmission network and the different modality data; The data receiving end performs initial decompression on the compressed modality data type, and restores it to the initial feature vectors according to the correlation analysis annotation results, and then restores it to the original multi-modal data.

2. An integrated method for multi-modal sensor network data compression and efficient transmission according to claim 1, characterized in that, The data receiving end identifies the transmission strategy, obtains the modality of the transmitted data and the transmission network, and judges the accuracy of the data modality of the decompressed data result according to the transmission modality and the transmission network; when there is a problem with the accuracy, the transmitted data is decompressed again in combination with the transmission strategy to complete the initial decompression of the data.

3. An integrated method for multi-modal sensor network data compression and efficient transmission according to claim 1, characterized in that, The data acquisition end further includes encoding the feature vector set to generate an encoded vector.

4. An integrated method for multi-modal sensor network data compression and efficient transmission according to claim 3, characterized in that, The data transmission end further includes receiving the encoded vector, encrypting the encoded vector using the AES encryption algorithm, compressing the encrypted encoded vector, and completing data transmission according to the data volume of the encoded vector.

5. An integrated method for multimodal sensor network data compression and efficient transmission according to claim 4, characterized in that, The data receiving end decrypts the decrypted data using the key to obtain the encoded vector, decodes the encoded vector to generate feature vectors, and obtains the original multi-modal data according to the mapping relationship between the feature vectors and the data to be transmitted.

6. An integrated system for multi-modal sensing network data compression and efficient transmission, characterized in that, Including: Data acquisition module: used to acquire data from different modality sensors and preprocess it as multi-modal data samples, extract the feature vectors of each modality data in the multi-modal data samples, and generate multiple feature vector sets according to different modalities; Data transmission module: used to receive multiple feature vector sets, respectively perform correlation analysis on the feature vectors in different feature vector sets; complete feature selection according to the correlation analysis results and annotate the correlation analysis results to obtain a feature subset; Based on The feature subset compresses the feature vectors; Different transmission strategies are selected according to the transmission network and the different modality data; Data receiving module: used to perform initial decompression on the compressed modality data type, and restore it to the initial feature vectors according to the correlation analysis annotation results, and then restore it to the original multi-modal data.

7. A computer storage medium, characterized in that, A computer program is stored on the computer storage medium, and when the computer program is executed by a processor, it implements the steps of an integrated method for multi-modal sensing network data compression and efficient transmission as described in any one of claims 1-6.

Images