Application of 3d-hevc fast coding in Internet of Things data intelligent decision
By applying the 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory in agricultural IoT systems, the problem of low agricultural data transmission efficiency is solved, efficient video encoding and transmission is realized, and agricultural IoT technology is promoted. application.
Patent Information
- Application Number
- CN202411214668.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-30
- Publication Date
- 2024-11-15
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The transmission efficiency in agricultural data transmission is low, which limits the promotion and application of agricultural Internet of Things technology.
A 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory was designed, combined with the intelligent agricultural IoT management platform, the video encoding process was optimized, the encoding complexity of deep map videos was reduced, and through streaming media technology and Hybrid encoding improves video transmission efficiency.
It significantly reduces the computational complexity and encoding time of 3D-HEVC, improves video transmission performance, realizes more efficient data transmission, and meets the actual needs of agricultural Internet of Things systems.
Smart Images

Figure CN118972413A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field related to the application of 3D-HEVC fast coding in intelligent decision-making of Internet of Things data, and specifically relates to the application of 3D-HEVC fast coding in intelligent decision-making of Internet of Things data. Background Art
[0002] Based on the Internet, the Internet of Things uses information sensing technology to quickly expand human-machine connections to human-object connections. It has been widely used in various fields, greatly improving production efficiency and work efficiency. The Internet of Things technology also has superior applicability to agriculture. In modern agricultural production, the Internet of Things technology is used to collect sensor data such as soil status, environmental status, and crop growth status through the devices in its sensing layer. The processed data is uploaded to the support layer of the agricultural Internet of Things platform through the network layer. The Internet of Things platform uses cloud computing technology to analyze, process and display the perceived data in real time, and conduct intelligent decision-making feedback to achieve real-time monitoring of agricultural production and automatic control of agricultural production equipment. The application of the agricultural Internet of Things has greatly improved the level of intelligent agricultural production. This is also a research hotspot that has attracted widespread attention from researchers. Remote video monitoring service is one of the most important technologies in agricultural Internet of Things technology. The development of video monitoring technology has overcome the shortcomings of traditional analog monitoring, such as large workload and limited transmission distance, and further expanded the scope and application scenarios of video monitoring. Through video monitoring, staff can break through geographical restrictions and understand the on-site situation more timely and accurately, greatly improving the efficiency of the system. In the current video monitoring technology, remote video monitoring service is one of the most important technologies in agricultural Internet of Things technology. In technical research, the widespread application of high-definition video, 4K video and 3D video has gradually increased the requirements for video coding. The purpose of video coding is to compress and reconstruct the compressed video under the condition of ensuring video quality. Scholars have always been committed to better video quality and higher compression efficiency. The previous generation of video coding standard H.264 / AVC can no longer meet the requirements of high-definition video and high frame rate video. Therefore, a new video coding standard, namely the high-efficiency video coding standard HEVC, was proposed in 2013. A large number of studies have shown that although HEVC can improve the coding compression rate, it also brings high computational complexity. 3D-HEVC technology is an extension of HEVC technology. In summary, promoting the development of agricultural Internet of Things and effectively solving the practical problem of low transmission rate in agricultural data transmission are extremely critical and practical for the intelligent improvement and problem solving in the agricultural field. Aiming at the problem of low transmission efficiency in agricultural data transmission, an intelligent decision-making management platform based on the agricultural Internet of Things system is innovatively designed. The platform mainly includes a remote control module and a video monitoring module. In addition, a fast 3D-HEVC algorithm based on Bayesian decision theory is proposed to optimize 3D-HEVC technology.
[0003] The transmission efficiency of agricultural data is low, so the market needs to further promote the application of agricultural Internet of Things to solve the current problems. Summary of the invention
[0004] The purpose of the present invention is to provide an application of 3D-HEVC fast coding in intelligent decision-making of Internet of Things data to solve the low transmission efficiency in agricultural data transmission proposed in the above background technology. Therefore, the market needs to further promote the promotion and application of agricultural Internet of Things to solve the current problems.
[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: an application of 3D-HEVC fast coding in intelligent decision-making of Internet of Things data, including research methods and experiments, and research results. The research methods and experiments include design requirements of intelligent agricultural Internet of Things management platform, agricultural intelligent Internet of Things management platform architecture, 3D-HEVC, 3D-HEVC video inter-frame fast algorithm based on Bayesian decision theory and test environment and evaluation indicators. The research results include database test results, 3D-HEVC inter-frame fast algorithm test results based on Bayesian decision theory, video forwarding service test results and algorithm advantage analysis.
[0006] Preferably, the method for designing the requirements of the smart agriculture Internet of Things management platform is as follows:
[0007] The main goal of the agricultural Internet of Things management platform is to achieve scalability and video functions. Scalability refers to the compatibility of equipment, that is, the ability to personalize the status information that needs to be monitored. The integration of equipment status information can improve the platform's adaptability to different types of monitoring equipment and expand the platform's application scope. Video functions mainly refer to the functions of the video surveillance management platform.
[0008] Preferably, the method and steps of the agricultural intelligent Internet of Things management platform architecture are as follows:
[0009] Step 1: Remote Control Module
[0010] The remote control module is built by the SSH integration framework + ApacheMina framework, and the MySQL database is used to manage data. Spring, as a lightweight solution for enterprise application development, is integrated with various technologies of the persistence layer. It is the core framework of the entire website architecture and runs through the presentation layer, business layer and persistence layer. Struts is a Web layer framework based on the MCV mode. It sets the distribution and processing rules of Web requests through file merging. Hibernate is a lightweight Java2EE application in the persistence layer, which is used to simplify operations. It is mainly responsible for the communication tasks between devices and Web application servers in the system.
[0011] Step 2: Video surveillance module
[0012] By using embedded terminals, videos can be collected, saved, and published, which optimizes video surveillance and can display monitoring images more intuitively and conveniently. Users only need to access the corresponding terminals to obtain video information. However, in this application scenario, users are not allowed to directly access the video data of the video management device. Therefore, it is necessary for the video surveillance module to provide users with video forwarding services with storage and transmission functions.
[0013] Preferably, the 3D-HEVC method is as follows:
[0014] The core technology required for video forwarding services is streaming technology. The streaming system includes compression coding services, streaming data, streaming servers, streaming networks and streaming players. Compression coding services can reduce the amount of data transmission as much as possible under the same visual effect to minimize the transmission of redundant information. Video coding technology can further improve the performance of streaming technology, usually including predictive coding using image transfer, transform coding using image orthogonality, and hybrid coding combining the above two technologies. Hybrid coding can eliminate different types of image redundancy and obtain a higher compression ratio than any single method. Therefore, hybrid coding technology is also the coding technology adopted by international video coding standards. The HEVC standard was proposed in 2013. As the key optimization standard of hybrid coding technology, it is the latest video coding standard. A large number of studies have shown that although HEVC improves the coding compression rate, it also brings higher computational complexity. As an extension of HEVC technology, 3D-HEVC adds depth map coding on the basis of multi-view video coding technology, further improving the complexity of coding, and introducing new coding-related content such as inter-view correlation and inter-object correlation. MDV is the main technical architecture of 3D-HEVC. MDV refers to shooting the video of the texture map and the corresponding video of the depth map as input information to the encoder from two or three viewpoints.
[0015] Preferably, the method of the 3D-HEVC video inter-frame fast algorithm based on Bayesian decision theory is as follows:
[0016] In 3D-HEVC, the complex encoding process of the depth map video with a large amount of computation does not contribute much to the encoding performance. Therefore, we use the fast algorithm of the depth map video encoding to reduce the computational complexity of 3D-HEVC, or the inter-frame fast algorithm of the 3D-HEVC video based on Bayesian decision theory used in this paper. Studies have shown that when using machine learning methods, as long as some features that meet specific connections are extracted, such connections can be learned to establish classification and prediction models. As a machine learning method, the Bayesian decision method has the advantages of low computational complexity and simple calculation process. Under the assumption of independent distribution, it has better and less calculation results. Therefore, the Bayesian decision theory is introduced in the fast algorithm to extract appropriate features for self-prediction and classification of the encoding process.
[0017] Preferably, the test environment and evaluation index method are as follows:
[0018] The performance test of the agricultural Internet of Things management platform mainly studies the database performance, access concurrency performance and data reception performance of the platform. In this test, data is collected by setting the corresponding parameters of the test environment. The specified parameter settings are as follows. In the database performance test, the number of threads is set to 100 and the maximum number of connections is set to 400. In the web access concurrency test, the number of threads is set to 500. In addition, multi-threaded programming is used to simulate the transmission of multi-device status data. In addition, 20 pieces of data are sent at a constant speed within 20 seconds.
[0019] Preferably, the method and steps of the database test results are as follows:
[0020] Step 1: Database performance testing
[0021] In the performance test of the database, the number of threads is set to 100 and the maximum number of connections is set to 400. In this case, the test simulates the completion time and success rate of 1536 data queries made by concurrent users under different connection numbers in the database.
[0022] Step 2: Web access concurrency test
[0023] In the Web access concurrency test, the number of threads was set to 500. The login time and data request completion rate under different user numbers were tested.
[0024] Step 3: Data reception performance test
[0025] A multi-threaded programming method is used to simulate the transmission of multiple device status data. 20 data are sent at a constant speed within 20 seconds. The data acquisition device monitors the data port for data processing. The received status information is stored in the corresponding database table, and the data reception result is obtained by querying the data record information.
[0026] Preferably, the method for testing the results of the 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory is as follows:
[0027] Experiments were conducted on six sequences including Kendo, Newspaper, GT-Fly, Poznan-Hall2, Undo-Dancer and Shark, and ΔT1, ΔT2 and ΔB of the 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory were further studied. The experimental results are shown in Fig.11 As shown in the figure, the average values of ΔT1 in the 6 test sequences are -41.63%, -39.53%, -53.6% and -52.45% respectively, and the total average value is -46.5%. The average values of ΔT2 are -20.48%, -20.53%, -26.62%, -22.75% and -25.22% respectively, and the total average value is -22.52%. The average values of ΔB are 1.05%, 0.74%, 0.21%, 0.14%, 0.94% and 0.61% respectively. The experimental results show that compared with the traditional 3D-HEVC coding algorithm, the depth map coding time and the total coding time of the 3D-HEVC inter-frame fast algorithm based on Bayesian decision technology are reduced by 46.5% and 22.52% respectively, which proves the good application performance of the 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory.
[0028] Preferably, the video forwarding service test result is based on the above technology and performs five performance tests on the video forwarding service.
[0029] Preferably, the algorithm advantage analysis method compares the experimental results of the proposed algorithm with the experimental results of the deep texture algorithm, the early skip algorithm and the block-level algorithm to further explore the superiority of the proposed algorithm.
[0030] Compared with the prior art, the present invention provides an application of 3D-HEVC fast coding in intelligent decision-making of IoT data, which has the following beneficial effects:
[0031] Conduct targeted research on video coding in agricultural Internet of Things systems to further promote the development of agricultural Internet of Things systems and improve the application effect of video surveillance technology. First, an agricultural Internet of Things system based on an intelligent decision-making management platform is designed, which consists of a remote control module and a video surveillance module. At the same time, a three-dimensional high-efficiency video fast coding algorithm based on Bayesian decision theory is designed to optimize 3D-HEVC technology, and the performance of the intelligent management platform and the 3D-HEVC fast algorithm based on Bayesian decision theory is experimentally studied. The experimental results show that the intelligent management platform has good application effects in database performance testing, website access concurrency testing, data reception performance, etc. In addition to the good application effect, compared with the traditional 3D-HEVC algorithm, the 3D-HEVC fast algorithm proposed in this paper reduces the encoding time of the depth map and the total encoding time by 46.5% and 22.52%. At the same time, with the support of the 3D-HEVC fast algorithm, the agricultural intelligent management platform shows better video transmission performance and application effect, which meets the actual needs of this stage. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] Figure 1 This is a functional structure diagram of an agricultural intelligent Internet of Things management platform for application of 3D-HEVC fast coding in intelligent decision-making of Internet of Things data according to the present invention.
[0033] Figure 2 This is a schematic diagram of the main functional structure of an agricultural intelligent material network management platform for the application of 3D-HEVC fast coding in intelligent decision-making of Internet of Things data in the present invention.
[0034] Figure 3 The present invention is a schematic diagram of the remote control module structure for the application of 3D-HEVC fast coding in the intelligent decision-making of Internet of Things data.
[0035] Figure 4 The present invention is a schematic diagram of the data processing cycle structure of a remote control module for applying 3D-HEVC fast coding in intelligent decision-making of Internet of Things data.
[0036] Figure 5 The present invention is a schematic diagram of the framework structure of a video forwarding service for application of 3D-HEVC fast encoding in intelligent decision-making of Internet of Things data.
[0037] Figure 6 A schematic diagram of the 3D-HEVC standard architecture for the application of 3D-HEVC fast encoding in intelligent decision-making of IoT data according to the present invention.
[0038] Figure 7 The present invention is a schematic diagram of the application structure of 3D-HEVC fast coding in intelligent decision-making of Internet of Things data.
[0039] Figure 8 The present invention is a schematic diagram of the test results of the application database performance of 3D-HEVC fast coding in the intelligent decision-making of Internet of Things data.
[0040] Fig. 9 The present invention is a schematic diagram of the result structure of a Web access board concurrency test for an application of 3D-HEVC fast encoding in intelligent decision-making of Internet of Things data.
[0041] Fig.10 The present invention is a schematic diagram of the simulation result structure of data reception of a 3D-HEVC fast encoding application in the intelligent decision-making of Internet of Things data.
[0042] Fig.11 The present invention is a schematic diagram of the experimental results of a 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory for the application of 3D-HEVC fast coding in intelligent decision-making of IoT data.
[0043] Fig.12 The present invention is a schematic diagram of the test result structure of a video forwarding service using 3D-HEVC fast coding in intelligent decision-making of IoT data.
[0044] Fig.13 The present invention is a schematic diagram of the structural comparison of different algorithms for applying 3D-HEVC fast coding in the intelligent decision-making of Internet of Things data. DETAILED DESCRIPTION
[0045] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only 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 ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0046] The invention provides Figure 1-13 The application of 3D-HEVC fast coding in intelligent decision-making of Internet of Things data shown in the invention includes research methods and experiments, and research results. The research methods and experiments include design requirements of intelligent agricultural Internet of Things management platform, architecture of agricultural intelligent Internet of Things management platform, 3D-HEVC, 3D-HEVC video inter-frame fast algorithm based on Bayesian decision theory, test environment and evaluation indicators. The research results include database test results, test results of 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory, video forwarding service test results and algorithm advantage analysis.
[0047] The agricultural Internet of Things management platform is studied, with a focus on the optimization of video forwarding services in the agricultural management system. On the basis of relevant theoretical foundations, a 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory is designed. Through a series of tests, the performance verification of the agricultural Internet of Things platform and the 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory is verified. The main conclusions are as follows: First, in the agricultural Internet of Things management platform, when the number of concurrent users is less than 300, the system successful access rate is 100%, indicating that the overall performance of the platform is good. Secondly, a total of 5 repeated experiments show that the 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory is superior to the traditional 3D-HEVC algorithm in both the encoding time of the depth map and the total encoding time. Finally, the video forwarding service of the platform has good transmission performance, and the intermediate packet loss of data reception is in line with practical applications. The 3D-HEVC fast algorithm is optimized according to the specific situation, so that the overall research has certain practical application value. However, there are still some defects in this study that need to be improved. It is necessary to further reduce the complexity of the constructed algorithm and improve the performance of the algorithm to facilitate the application effect of the algorithm.
[0048] like Figure 1 As shown in the figure, the design requirements of the intelligent agricultural Internet of Things management platform are as follows. The main goal of the agricultural Internet of Things management platform is to achieve scalability and video functions. Scalability refers to the compatibility of equipment, that is, the ability to personalize the status information that needs to be monitored. The integration of equipment status information can improve the platform's adaptability to different types of monitoring equipment and expand the platform's application scope. The video function mainly refers to the function of the video monitoring management platform.
[0049] In system design, it is necessary to consider the adaptability of the management platform to different devices and realize the scalability of video surveillance. Therefore, personalized configuration is the key point in the actual design process. In actual applications, different scenarios require different basic functional information and system requirements. Therefore, it is necessary to consider the balance between common needs and personalized needs in system design. Based on the above principles, Figure 1 The following functional management platform system designs are considered.
[0050] like Figure 2 , Figure 3 , Figure 4 and Figure 5As shown in the figure, the method and steps of the architecture of the agricultural intelligent Internet of Things management platform are as follows: Step 1: Remote control module. The remote control module is built by the SSH integration framework + ApacheMina framework. The MySQL database is used to manage data. Spring, as a lightweight solution for enterprise application development, is integrated with various technologies of the persistence layer. It is the core framework of the entire website architecture, running through the presentation layer, business layer and persistence layer. Struts is a Web layer framework based on the MCV mode. It sets the distribution and processing rules of Web requests through file merging. Hibernate is a lightweight Java2EE application in the persistence layer, which is used to simplify operations. It is mainly responsible for the communication tasks between the devices in the system and the Web application server. Step 2: Video monitoring module. The embedded terminal is used to collect, save and publish videos. The video monitoring is optimized, and the monitoring screen can be displayed more intuitively and conveniently. Users only need to access the corresponding terminal to obtain video information. However, in this application scenario, users are not allowed to directly access the video data of the video management device. Therefore, it is necessary for the video monitoring module to provide users with video forwarding services with storage and transmission functions.
[0051] in Figure 2 ,The remote control module implements Web application services, mainly ,provides users with application interaction, data reception and data terminal equipment ,processing and system information management, operation and display. ,The video monitoring module provides efficient and fast monitoring video transmission and ,display optimization solutions.
[0052] The structure of the remote control module under the structure of step 1 is as follows Figure 3 As shown, Figure 3 The three layers of the remote control module are shown, namely the presentation layer, business layer and persistence layer. The presentation layer of the browser / server mode is mainly provided to users for access through a Web browser. The business layer can implement some business logic processing of the system software, including user management, device management, personnel management, log management, etc. The persistence layer is mainly used to save data, where data operations are implemented through functional references of the business layer. The remote control module has two main lines. The first main line is the device status data uploaded to the Apache Mina framework. Then, after parsing the application, the received data is saved in the database. The second main line is that the user registers or logs in to the website and manages the relevant information of the website, which is implemented based on the SSH integration framework.
[0053] Figure 4 This is the main process of data processing in this module. Figure 4First, the equipment operation status information collected by the data acquisition unit is uploaded to the system server, where the information is processed by the Apache Mina framework. Then, the processed data is saved in the database as part of the data source of system information management, and then returned to the user. Data preprocessing is crucial in this module, involving the normalization, standardization and noise processing of one-dimensional data. Commonly used data normalization methods are numerical normalization and variance normalization. Numerical normalization can eliminate the gap between data and keep the changes of each parameter within a certain range.
[0054] For step 2, Figure 5 As shown, the video forwarding service can be regarded as a transmission station between the capture device and the user. The video forwarding service mainly obtains video data, forwards video data, stores video data centrally, receives video data sent by the video acquisition device, and saves it after classification. When the client sends a request, it will find the corresponding video from the storage and send it to the client. Dual-channel technology and multi-threaded operation are two important key technologies in the video forwarding service. Dual-channel technology establishes two channels in the communication process between the client and the video forwarding service center or between the terminal device and the video forwarding service center. One channel is used to transmit control information, and the other channel is used to send video data. The client sends a video request command through the port monitored by the video forwarding server. The video forwarding server feeds back the video stream to the client from the free port randomly selected locally after receiving the request. In the whole process, the request channel and the data channel are separated. Multi-threading technology refers to a technical means to improve the overall performance of the computer by executing multiple program segments at the same time. In the video forwarding service, the main thread is used to monitor the device and the client's request, and the other thread receives the video data sent by the terminal device or sends it to the client.
[0055] like Figure 6As shown, the method of 3D-HEVC is as follows. The core technology required for video forwarding service is streaming technology. The streaming system includes compression coding service, streaming data, streaming server, streaming network and streaming player. Compression coding service can reduce the amount of data transmission as much as possible under the same visual effect to minimize the transmission of redundant information. Video coding technology can further improve the performance of streaming technology, usually including predictive coding using image transfer, transform coding using image orthogonality, and hybrid coding combining the above two technologies. Hybrid coding can eliminate different types of image redundancy and obtain a higher compression ratio than any single method. Therefore, hybrid coding technology is also an international video coding technology. The HEVC standard is the coding technology adopted by the coding standard. The HEVC standard was proposed in 2013. As a key optimization standard for hybrid coding technology, it is the latest video coding standard. A large number of studies have shown that although HEVC has improved the coding compression rate, it has also brought about higher computational complexity. As an extension of HEVC technology, 3D-HEVC adds depth map coding on the basis of multi-view video coding technology, further improving the coding complexity and introducing new coding-related content such as inter-view correlation and inter-object correlation. MDV is the main technical architecture of 3D-HEVC. MDV refers to shooting the video of the texture map and the corresponding video of the depth map as input information to the encoder from two or three viewpoints.
[0056] Architecture such as Figure 6 As shown, the encoder receives video information from multiple viewpoints, mainly including texture maps and depth maps. Figure 6 In the 3D-HEVC encoder, viewpoint 0 is called an independent viewpoint and can be directly encoded by the HEVC encoder. The encoded bit stream can be directly decoded and played as a two-dimensional video. Viewpoints 1 to N are non-independent viewpoints and can only be encoded after the 3D-HEVC encoder completes encoding of the basic view.
[0057] like Figure 7 As shown, the method of the 3D-HEVC video inter-frame fast algorithm based on Bayesian decision theory is as follows. In 3D-HEVC, the complex encoding process of the depth map video with a large amount of calculation does not contribute much to the encoding performance. Therefore, we use the fast algorithm of the depth map video encoding to reduce the computational complexity of 3D-HEVC, or use the 3D-HEVC video inter-frame fast algorithm based on Bayesian decision theory used in this article. Studies have shown that when using machine learning methods, as long as some features that meet specific connections are extracted, such connections can be learned to establish classification and prediction models. As a machine learning method, the Bayesian decision method has the advantages of low computational complexity and simple calculation process. Under the assumption of independent distribution, it has better and less calculation results. Therefore, the Bayesian decision theory is introduced in the fast algorithm to extract appropriate features to self-predict and classify the encoding process.
[0058] The key parameters of the fast algorithm are expanded to Equation 13
[0059]
[0060] Among them, x is the Cucost value when Cu does not partition NS, σ is the standard deviation, and u is the average value. The key parameters of the fast algorithm are shown in the formula. It is necessary to set the loss factor λ(k) in the Bayesian decision. The calculation of λ(k) when partition S and non-partition NS are expressed as equations 14) and 15, respectively, as follows
[0061]
[0062]
[0063] The decision expression of Bayesian risk is as follows
[0064]
[0065] The threshold selection needs to consider AZB, and the AZB of DIS and SKIP are represented as DAZB and SAZB respectively in the test.
[0066] like Figure 1 As shown, the test environment and evaluation index methods are as follows. The performance test of the agricultural Internet of Things management platform mainly studies the database performance, access concurrency performance and data reception performance of the platform. In this test, data is collected by setting the corresponding parameters of the test environment. The specified parameter settings are shown below. In the database performance test, the number of threads is set to 100 and the maximum number of connections is set to 400. In the web access concurrency test, the number of threads is set to 500. In addition, multi-threaded programming is used to simulate the transmission of multi-device status data. In addition, 20 pieces of data are sent at a constant speed within 20 seconds.
[0067] The performance test environment for this platform is shown in the following table:
[0068]
[0069] Using 3D-HEVC reference software HTM16.0 as the test platform, the performance test of the 3D-HEVC video inter-frame fast algorithm based on Bayesian decision theory was carried out. The following table shows the main integration of the test platform:
[0070]
[0071] like Figure 8 , Fig. 9 and Fig.10As shown, the method and steps of the database test results are as follows: Step 1: Database performance test. In the database performance test, the number of threads is set to 100 and the maximum number of connections is set to 400. In this case, the test simulates the completion time and success rate of 1536 data queries for concurrent users under different connection numbers in the database. Step 2: Web access concurrency test. In the Web access concurrency test, the number of threads is set to 500. The login time and completion rate of data requests under different user numbers were tested. Step 3: Data reception performance test. A multi-threaded programming method is used to simulate the transmission of multiple device status data. 20 data are sent at a constant speed within 20 seconds. The data acquisition device monitors the data port for data processing. The received status information is stored in the corresponding database table. The data reception result is obtained by querying the data record information.
[0072] For step one Figure 8 It can be seen that under the same connection number, as the number of database access users increases, the database retrieval time gradually increases, which is an important factor affecting database performance. Figure 8 a,Under the same number of connections, the reflection time database gradually increases with the growth of the number of access users, which is an important factor affecting the performance of the database. At the same time, with a certain number of access users, the larger the number of connections, the faster the database query completion time. The results of this study show that the setting of the number of connections in the database also affects the performance of the database. Therefore, the number of access users in the database can dynamically adjust the setting of the number of connections in specific scenarios, such as Figure 8 As shown in Figure 2, when the number of access users and the maximum number of connections are small, the success rate of database queries is 100%. However, as the number of access users and the maximum number of connections increase, the success rate of queries gradually decreases, and the minimum success rate is 50%.
[0073] For step 2, in Figure a, Request=0 means that only login operation is performed. When the number of requests remains unchanged, the system response time will gradually increase with the increase in the number of accessing users. In addition, when the number of concurrent users increases, the response time increases rapidly. When the number of accessing users remains unchanged, the system response time gradually increases with the increase in the number of user requests. Fig. 9 b shows that when the number of concurrent users is less than 300, the success rate of the system is 100%. However, for 400 concurrent users, the success rate of the system gradually decreases as the number of access requests increases.
[0074] For step three, Fig.10It can be seen that when the number of threads increases, the system data transmission remains basically unchanged, while the data processing time shows an increasing trend. This is mainly because the platform frst caches the received data so that it can process the data from the cache later when there are idle threads. Therefore, the data received in a certain period of time is greater than the amount of data that the system can actually process.
[0075] like Fig.11 , Fig.12 and Fig.13 As shown in the figure, the method for testing the results of the 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory is as follows. Experiments were conducted on six sequences including Kendo, Newspaper, GT-Fly, Poznan-Hall2, Undo-Dancer and Shark. ΔT1, ΔT2 and ΔB of the 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory were further studied. The experimental results are shown in Fig.11 As shown in the figure, the average values of ΔT1 in the six test sequences are -41.63%, -39.53%, -53.6% and -52.45%, respectively, with a total average value of -46.5%. The average values of ΔT2 are -20.48%, -20.53%, -26.62%, -22.75% and -25.22%, with a total average value of -22.52%. The average values of ΔB are 1.05%, 0.74%, 0.21%, 0.14%, 0.94% and 0.61%, respectively. The experimental results show that the 3D-HEVC coding algorithm is superior to the traditional 3D-HEVC coding algorithm. Compared with the above method, the depth map encoding time and the total encoding time of the 3D-HEVC inter-frame fast algorithm based on Bayesian decision technology were reduced by 46.5% and 22.52% respectively, which proved the good application performance of the 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory. The video forwarding service test results were based on the above technology and five performance tests were conducted on the video forwarding service. The algorithm advantage analysis method compared the experimental results of the proposed algorithm with the experimental results of the deep texture algorithm, early skip algorithm and block level algorithm to further explore the superiority of the proposed algorithm.
[0076] Fig.12 It shows that when the total number of transmissions is less than 50 times, there are basically no errors in the video due to video loss. However, as the total number of transmissions increases, errors in the video gradually appear. By playing the received video file, it can be found that when the error is within the range of 10KB, most of the lost data is at the beginning of the video. In addition, a blurred screen appears in the video and key frames begin to be missing. In addition, there is little packet loss in the middle of data reception. In general, the application effect of the platform's video forwarding service is consistent with the application effect in the current big data context.
[0077] according to Fig.13The results show that the proposed algorithm has obvious advantages on different data sets. Compared with the early skip algorithm and block-level algorithm, the ΔT1T2 and ΔT2 values of the proposed algorithm are the smallest, proving that the proposed algorithm can significantly reduce the data transmission time. Compared with the deep texture algorithm, the proposed algorithm can increase the image bit rate and improve the image quality.
[0078] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. Application of 3D-HEVC fast coding in intelligent decision-making of IoT data, It is characterized in that It includes research methods and experiments, and research results. The research methods and experiments include the design requirements of the intelligent agricultural Internet of Things management platform, the architecture of the agricultural intelligent Internet of Things management platform, 3D-HEVC, a 3D-HEVC video inter-frame fast algorithm based on Bayesian decision theory, a test environment and evaluation indicators. The research results include database test results, 3D-HEVC inter-frame fast algorithm test results based on Bayesian decision theory, video forwarding service test results and algorithm advantage analysis.
2. The application of 3D-HEVC fast coding in intelligent decision-making of IoT data according to claim 1, Features: The method for designing the smart agriculture IoT management platform is as follows: The main goal of the agricultural Internet of Things management platform is to achieve scalability and video functions. Scalability refers to the compatibility of equipment, that is, the ability to personalize the status information that needs to be monitored. The integration of equipment status information can improve the platform's adaptability to different types of monitoring equipment and expand the platform's application scope. Video functions mainly refer to the functions of the video surveillance management platform.
3. The application of 3D-HEVC fast coding in intelligent decision-making of IoT data according to claim 1, Features: The methods and steps of the agricultural intelligent Internet of Things management platform architecture are as follows: Step 1: Remote Control Module The remote control module is built by the SSH integration framework + ApacheMina framework, and the MySQL database is used to manage data. Spring, as a lightweight solution for enterprise application development, is integrated with various technologies of the persistence layer. It is the core framework of the entire website architecture and runs through the presentation layer, business layer and persistence layer. Struts is a Web layer framework based on the MCV mode. It sets the distribution and processing rules of Web requests through file merging. Hibernate is a lightweight Java2EE application in the persistence layer, which is used to simplify operations. It is mainly responsible for the communication tasks between devices and Web application servers in the system. Step 2: Video surveillance module By using embedded terminals, videos can be collected, saved, and published, which optimizes video surveillance and can display monitoring images more intuitively and conveniently. Users only need to access the corresponding terminals to obtain video information. However, in this application scenario, users are not allowed to directly access the video data of the video management device. Therefore, it is necessary for the video surveillance module to provide users with video forwarding services with storage and transmission functions.
4. The application of 3D-HEVC fast coding in intelligent decision-making of IoT data according to claim 1, Features: The 3D-HEVC method is as follows: The core technology required for video forwarding services is streaming technology. The streaming system includes compression coding services, streaming data, streaming servers, streaming networks and streaming players. Compression coding services can reduce the amount of data transmission as much as possible under the same visual effect to minimize the transmission of redundant information. Video coding technology can further improve the performance of streaming technology, usually including predictive coding using image transfer, transform coding using image orthogonality, and hybrid coding combining the above two technologies. Hybrid coding can eliminate different types of image redundancy and obtain a higher compression ratio than any single method. Therefore, hybrid coding technology is also the coding technology adopted by international video coding standards. The HEVC standard was proposed in 2013. As the key optimization standard of hybrid coding technology, it is the latest video coding standard. A large number of studies have shown that although HEVC improves the coding compression rate, it also brings higher computational complexity. As an extension of HEVC technology, 3D-HEVC adds depth map coding on the basis of multi-view video coding technology, further improving the complexity of coding, and introducing new coding-related content such as inter-view correlation and inter-object correlation. MDV is the main technical architecture of 3D-HEVC. MDV refers to shooting the video of the texture map and the corresponding video of the depth map as input information to the encoder from two or three viewpoints.
5. The application of 3D-HEVC fast coding in intelligent decision-making of IoT data according to claim 1, Features: The method of the 3D-HEVC video inter-frame fast algorithm based on Bayesian decision theory is as follows: In 3D-HEVC, the complex encoding process of the depth map video with a large amount of computation does not contribute much to the encoding performance. Therefore, we use the fast algorithm of the depth map video encoding to reduce the computational complexity of 3D-HEVC, or the inter-frame fast algorithm of the 3D-HEVC video based on Bayesian decision theory used in this paper. Studies have shown that when using machine learning methods, as long as some features that meet specific connections are extracted, such connections can be learned to establish classification and prediction models. As a machine learning method, the Bayesian decision method has the advantages of low computational complexity and simple calculation process. Under the assumption of independent distribution, it has better and less calculation results. Therefore, the Bayesian decision theory is introduced in the fast algorithm to extract appropriate features for self-prediction and classification of the encoding process.
6. The application of 3D-HEVC fast coding in intelligent decision-making of IoT data according to claim 1, Features: The test environment and evaluation index methods are as follows: The performance test of the agricultural Internet of Things management platform mainly studies the database performance, access concurrency performance and data reception performance of the platform. In this test, data is collected by setting the corresponding parameters of the test environment. The specified parameter settings are as follows. In the database performance test, the number of threads is set to 100 and the maximum number of connections is set to 400. In the web access concurrency test, the number of threads is set to 500. In addition, multi-threaded programming is used to simulate the transmission of multi-device status data. In addition, 20 pieces of data are sent at a constant speed within 20 seconds.
7. The application of 3D-HEVC fast coding in intelligent decision-making of IoT data according to claim 1, Features: The method and steps for database test results are as follows: Step 1: Database performance testing In the performance test of the database, the number of threads is set to 100 and the maximum number of connections is set to 400. In this case, the test simulates the completion time and success rate of 1536 data queries made by concurrent users under different connection numbers in the database. Step 2: Web access concurrency test In the Web access concurrency test, the number of threads was set to 500. The login time and data request completion rate under different user numbers were tested. Step 3: Data reception performance test A multi-threaded programming method is used to simulate the transmission of multiple device status data. 20 data are sent at a constant speed within 20 seconds. The data acquisition device monitors the data port for data processing. The received status information is stored in the corresponding database table, and the data reception result is obtained by querying the data record information.
8. The application of 3D-HEVC fast coding in intelligent decision-making of IoT data according to claim 1, Features: The method for testing the results of the 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory is as follows: Experiments were conducted on six sequences including Kendo, Newspaper, GT-Fly, Poznan-Hall2, Undo-Dancer and Shark. ΔT1, ΔT2 and ΔB of the 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory were further studied. The experimental results are shown in Figure 11. The average values of ΔT1 in the six test sequences are -41.63%, -39.53%, -53.6% and -52.45% respectively, and the total average value is -46.5%. The average values of ΔT2 are -20.48%, -20. The average values of ΔB are 1.05%, 0.74%, 0.21%, 0.14%, 0.94% and 0.61% respectively. The experimental results show that compared with the traditional 3D-HEVC coding algorithm, the depth map coding time and the total coding time of the 3D-HEVC inter-frame fast algorithm based on Bayesian decision technology are reduced by 46.5% and 22.52% respectively, which proves the good application performance of the 3D-HEVC inter-frame fast algorithm based on Bayesian decision theory.
9. The application of 3D-HEVC fast coding in intelligent decision-making of IoT data according to claim 1, Features: Video forwarding service test results Based on the above technology, five performance tests were conducted on the video forwarding service.
10. The application of 3D-HEVC fast coding in intelligent decision-making of Internet of Things data according to claim 1, Features: Method of Algorithm Advantage Analysis The experimental results of the proposed algorithm are compared with those of the deep texture algorithm, early skip algorithm and block-level algorithm to further explore the superiority of the proposed algorithm.
