Network computer

By introducing a custom-structured algorithm operating system into a network computer, the autoencoder neural network is trained multiple times, and an AI prediction model is used for intelligent prediction. This solves the problems of flexibility and targeting in image processing in traffic light monitoring, and achieves effective compression of image data and improved monitoring results.

CN121330239APending Publication Date: 2026-01-13NANJING LUZHIRUI TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202511488658.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2025-02-20
Filing Date
2025-10-17
Publication Date
2026-01-13

AI Technical Summary

Technical Problem

In existing technologies, when network computers monitor the operating environment of traffic lights, they lack the ability to match the imaging features of key components, resulting in a lack of flexibility and specificity in image processing.

Method used

A customized network computer algorithm operating system is introduced. By learning the autoencoder neural network multiple times, relevant data of traffic lights is obtained, and an AI prediction model is used for intelligent prediction. The most matching reference frame is selected for inter-frame coding to achieve the compression ratio prediction of image data.

Benefits of technology

It improves the flexibility and targeting of image processing, provides important basic information and basis, and selects the most suitable reference frame for subsequent traffic light monitoring images, thereby improving the monitoring effect.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to a network computer which comprises a memory, a processor and a network computer algorithm operating system. According to the method, the AI pre-judgment model can be introduced to intelligently pre-judge the data volume of the coding data stream of the operating environment image obtained by performing inter-frame coding on the operating environment image by adopting the reference image, so that pre-judgment of inter-frame coding compression ratios of different numerical values obtained by selecting different reference frames is completed, and an important basis is provided for selection of the reference frames; therefore, the most matched image processing mechanism is selected for the associated data.
Need to check novelty before this filing date? Find Prior Art

Description

TECHNICAL FIELD

[0001] The present application relates to the field of computer systems, and more particularly to a network computer. BACKGROUND

[0002] A network computer is an interactive information device under the client computing model, and the typical application program behavior analysis on it is of great significance to processor design and system development. The network computer is a computer used on the network, but it removes the traditional hard disk, floppy disk and other components, and belongs to a thin PC, and programs or storage on the network are provided by a server. The network computer has its own processing capability, but other software needs to be downloaded from the network server except for the core software, which saves frequent software upgrade and maintenance and reduces the cost. The network computer is a low-cost, upgrade-free, maintenance-free, easy-to-operate, right management, strong security, high-reliability terminal client in a certain application field and network environment, and application program running and data storage are on the server, and it has PC functions. It can meet the needs of managers and the public for information processing and information access, and is an inevitable product of information application subdivision in various industries.

[0003] CN118541653A discloses a method for configuring a control network, wherein the control network comprises a central unit of a superior for controlling and monitoring a facility system having a plurality of distributed facilities and a remote unit of an inferior for directly controlling the distributed facilities, wherein for controlling the distributed facilities, planning data are stored in the remote unit, the planning data comprising facility-related objects and / or modules.

[0004] CN117955728A discloses a single-blockchain system for different networks, comprising an external network computer, an internal network computer, a physically isolated computer A and a physically isolated computer B; the external network computer uses a serial port server with an Ethernet port to exchange network data with the physically isolated computer A; the physically isolated computer A and the physically isolated computer B use a serial bus port USB for non-network data transmission; the physically isolated computer B and the internal network computer use a serial port server with a USB port for network data interaction.

[0005] CN116204128A discloses a storage system of a network computer and a storage method thereof, and relates to the technical field of network computers. The storage system of the network computer comprises a main storage group, which is configured to be connected with the network computer through a server network and store all generated data. The main storage group comprises a first storage group, a second storage group, a third storage group, a data analysis system and a backup system. The data analysis system is configured to read the data of the first storage group and divide the data into hot data, cold data and read-only data. The data analysis system divides the data into hot data, cold data and read-only data, and stores the hot data using a data replication redundancy scheme and stores the cold data and read-only data using a distributed storage scheme based on error correction code, thereby reducing the amount of data that needs to be stored by the replication redundancy scheme and avoiding the high data update cost of the distributed storage scheme based on error correction code. SUMMARY

[0006] To solve the technical problems in the prior art, the present application provides a network computer, by introducing a network computer algorithm operating system with a customized structure, the automatic encoder neural network is learned multiple times to obtain an automatic encoder neural network after multiple learning and as an AI prediction model output, the number of times of learning of the automatic encoder neural network is positively correlated with the number of light bodies of the traffic signal lights at the set traffic intersection, thereby realizing the structure customization of the AI prediction model, obtaining the light body shape number, the number of light bodies, the adjacent light body interval distance, and the cross-sectional area of a single light body of the traffic signal lights at the set traffic intersection, and obtaining each item of image data of the running environment image and each item of image data of the reference image used when interframe coding is performed on the running environment image, each item of image data of each image is the number of pixel columns, the number of pixel rows, the number of foreground pixel points, and the coordinate values of each foreground pixel point of the image, thereby providing valuable basic information for subsequent intelligent prediction, and the AI prediction model is used to intelligently predict the data amount of the encoded data stream of the running environment image obtained by performing interframe coding on the running environment image using the reference image, and determine the compression ratio value obtained by performing interframe coding on the running environment image using the reference image based on the data amount of the encoded data stream of the running environment image and the original data amount of the running environment image, complete the prediction of the interframe coding compression ratio of different values obtained by selecting different reference frames, provide an important basis for the selection of reference frames for subsequent specific scene monitoring pictures, thereby selecting the most matched reference frame for different traffic signal light related data.

[0007] According to the application, a network computer is provided, comprising a memory, a processor and a network computer algorithm operating system, the system comprising: a directional establishment device arranged in the system, for performing multiple learning on an auto-encoder neural network to obtain an auto-encoder neural network after multiple learning and output as an AI prediction model; a signal receiving device arranged in the system, for receiving a running environment image captured by a network capture device arranged opposite a traffic signal lamp at a set traffic intersection, from a remote network transmission component, for performing image capture operation on the running environment of the traffic signal lamp at the set traffic intersection; a data acquisition device, for acquiring a lamp body shape number, a lamp body number, a distance between adjacent lamp bodies and a cross-sectional area of a single lamp body of the traffic signal lamp at the set traffic intersection; an image analysis device connected with the signal receiving device, for acquiring image data of each image of the running environment image and image data of a reference image used when performing interframe coding on the running environment image, the image data of each image being a pixel column number, a pixel row number, a foreground pixel point number and coordinate values of each foreground pixel point of the image; a prediction execution mechanism connected with the directional establishment device, the data acquisition device and the image analysis device respectively, for intelligently predicting, by using the AI prediction model, a data amount of an encoded data stream of the running environment image obtained by performing interframe coding on the running environment image using the reference image, according to the lamp body shape number, the lamp body number, the distance between adjacent lamp bodies and the cross-sectional area of a single lamp body of the traffic signal lamp at the set traffic intersection, the image data of each image of the running environment image and the image data of the reference image used when performing interframe coding on the running environment image; a ratio identification mechanism connected with the prediction execution mechanism, for determining a compression ratio value obtained by performing interframe coding on the running environment image using the reference image, based on the data amount of the encoded data stream of the running environment image and an original data amount of the running environment image; wherein the compression ratio value obtained by performing interframe coding on the running environment image using the reference image is determined based on the data amount of the encoded data stream of the running environment image and the original data amount of the running environment image, comprising: dividing the data amount of the encoded data stream of the running environment image by the original data amount of the running environment image to obtain the compression ratio value obtained by performing interframe coding on the running environment image using the reference image; The image data of each image of the running environment image and the reference image used when interframe coding is performed on the running environment image, the image data of each image being the pixel column number, pixel row number, foreground pixel point number and coordinate value of each foreground pixel point of the image, comprises: the reference image used when interframe coding is performed on the running environment image is an image captured by the network capture device at a past time point before the capture time point of the running environment image. The directional establishment device is used for multiple learning of the auto-encoder neural network to obtain the auto-encoder neural network after multiple learning and output as an AI prediction model, and the number of times of learning of the auto-encoder neural network is positively correlated with the number of lamp bodies of the traffic signal lamp of the set traffic intersection.

[0008] Therefore, the present application has at least the following three outstanding substantial features: Substantial feature A: introducing a network computer algorithm operating system of customized structure, multiple learning of the auto-encoder neural network to obtain the auto-encoder neural network after multiple learning and output as an AI prediction model, and the number of times of learning of the auto-encoder neural network is positively correlated with the number of lamp bodies of the traffic signal lamp of the set traffic intersection, so as to realize structure customization of the AI prediction model; Substantial feature B: obtaining the lamp body shape number, number of lamp bodies, adjacent lamp body spacing distance and cross-sectional area of a single lamp body of the traffic signal lamp of the set traffic intersection, and obtaining the image data of each image of the running environment image and the image data of the reference image used when interframe coding is performed on the running environment image, the image data of each image being the pixel column number, pixel row number, foreground pixel point number and coordinate value of each foreground pixel point of the image, so as to provide valuable basic information for subsequent intelligent prediction; Substantial feature C: the AI prediction model intelligently predicts the data amount of the encoded data stream of the running environment image obtained by interframe coding of the running environment image using the reference image according to the lamp body shape number, number of lamp bodies, adjacent lamp body spacing distance and cross-sectional area of a single lamp body of the traffic signal lamp of the set traffic intersection, the image data of the running environment image and the image data of the reference image used when interframe coding is performed on the running environment image, and determines the compression ratio value obtained by interframe coding of the running environment image using the reference image based on the data amount of the encoded data stream of the running environment image and the original data amount of the running environment image, so as to complete the prediction of the interframe coding compression ratio of different values obtained by different reference frames, and provide an important basis for the selection of subsequent reference frames. DETAILED DESCRIPTION

[0009] In the prior art, network computers are often used to monitor the operating environment of traffic signal lights at a designated traffic intersection, and thus it is desirable to match the processing of the monitoring picture with the imaging features of the key components in the monitoring picture, i.e., the traffic signal lights, so as to improve the flexibility and pertinence of image processing. Apparently, there is a lack of corresponding solutions in the prior art.

[0010] The present application discloses a network computer, comprising a memory, a processor and a network computer algorithm operating system. The embodiments of the network computer algorithm operating system of the present application will be described in detail below.

[0011] Embodiment 1 The network computer algorithm operating system shown in Embodiment 1 according to the present application comprises: a directional establishment device arranged in the system for performing multiple learning on an auto-encoder neural network to obtain the auto-encoder neural network after completing the multiple learning and output as an AI prediction model; Specifically, the directional establishment device arranged in the system for performing multiple learning on an auto-encoder neural network to obtain the auto-encoder neural network after completing the multiple learning and output as an AI prediction model comprises: an ASIC device is used to realize the directional establishment device arranged in the system for performing multiple learning on an auto-encoder neural network to obtain the auto-encoder neural network after completing the multiple learning and output as an AI prediction model; a signal receiving device arranged in the system for receiving an operating environment image captured by a network capture device from a remote end using a network transmission component, the network capture device being arranged opposite to a traffic signal light at a designated traffic intersection for performing an image capture operation on the operating environment of the traffic signal light at the designated traffic intersection; a data acquisition device for acquiring a light body shape number, a light body quantity, an adjacent light body interval distance and a cross-sectional area of a single light body of the traffic signal light at the designated traffic intersection; an image analysis device connected with the signal receiving device for acquiring each item of image data of the operating environment image and each item of image data of a reference image used when performing interframe coding on the operating environment image, each item of image data of each image being a pixel column quantity, a pixel row quantity, a foreground pixel point quantity and a coordinate value of each foreground pixel point of the image; A pre-judgment execution mechanism is connected with the orientation establishment device, the data acquisition device and the image analysis device, and is configured to use an AI pre-judgment model to intelligently pre-judge, according to the lamp body shape number, the number of lamp bodies, the interval distance between adjacent lamp bodies, the cross-sectional area of a single lamp body of the traffic signal lamp of the set traffic intersection, the image data of the running environment image and the image data of the reference image used when performing inter-frame coding on the running environment image, the data amount of the encoded data stream of the running environment image obtained by performing inter-frame coding on the running environment image using the reference image; A ratio identification mechanism is connected with the pre-judgment execution mechanism, and is configured to determine the compression ratio obtained by performing inter-frame coding on the running environment image using the reference image based on the data amount of the encoded data stream of the running environment image and the original data amount of the running environment image; The determination of the compression ratio obtained by performing inter-frame coding on the running environment image using the reference image based on the data amount of the encoded data stream of the running environment image and the original data amount of the running environment image includes dividing the data amount of the encoded data stream of the running environment image by the original data amount of the running environment image to obtain the compression ratio obtained by performing inter-frame coding on the running environment image using the reference image. The acquisition of the image data of the running environment image and the image data of the reference image used when performing inter-frame coding on the running environment image, and the image data of each image including the number of pixel columns, the number of pixel rows, the number of foreground pixel points and the coordinate values of each foreground pixel point of the image includes that the reference image used when performing inter-frame coding on the running environment image is an image captured by the network capture device at a past time before the capture time of the running environment image. The orientation establishment device configured to perform multiple learning on the auto-encoder neural network to obtain the auto-encoder neural network after multiple learning and output as the AI pre-judgment model includes that the number of learning performed by the auto-encoder neural network is positively correlated with the number of lamp bodies of the traffic signal lamp of the set traffic intersection. The data acquisition device configured to acquire the lamp body shape number, the number of lamp bodies, the interval distance between adjacent lamp bodies and the cross-sectional area of a single lamp body of the traffic signal lamp of the set traffic intersection includes that lamp bodies of different shapes have different lamp body shape numbers. The data acquisition device configured to acquire the lamp body shape number, the number of lamp bodies, the interval distance between adjacent lamp bodies and the cross-sectional area of a single lamp body of the traffic signal lamp of the set traffic intersection further includes that the cross-sectional area of a single lamp body of the traffic signal lamp of the set traffic intersection is the cross-sectional area of the single lamp body of the traffic signal lamp of the set traffic intersection in the vertical cross section. Among them, the AI ​​prediction model intelligently predicts the data volume of the encoded data stream of the operating environment image obtained by performing inter-frame coding on the operating environment image using the reference image based on the shape number, number of lights, spacing between adjacent lights, cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame coding on the operating environment image. The internal control command is transmitted by the command control channel. The AI ​​prediction model intelligently predicts the amount of data in the encoded data stream of the operating environment image obtained by performing inter-frame encoding on the operating environment image using the reference image based on the shape number, number of lights, spacing between adjacent lights, cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame encoding on the operating environment image. The internal control instructions are transmitted by the instruction control channel, which is operated and maintained by the SOC chip. The AI ​​prediction model intelligently predicts the amount of data in the encoded data stream of the operating environment image obtained by performing inter-frame encoding on the operating environment image using the reference image based on the shape number, number of lights, spacing between adjacent lights, cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame encoding on the operating environment image. The internal control instructions are transmitted by the instruction control channel, which is operated and maintained by an ASIC chip. Among them, the intermediate data of the encoded data stream of the operating environment image obtained by using the reference image to perform inter-frame encoding on the operating environment image is transmitted to the network transmission interface based on the shape number, number of lights, spacing between adjacent lights, cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame encoding on the operating environment image using the reference image. The intermediate data of the encoded data stream of the operating environment image obtained by performing inter-frame encoding on the operating environment image using the reference image is intelligently predicted by the AI ​​prediction model based on the shape number, number of lights, spacing between adjacent lights, cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame encoding on the operating environment image. This data is then transmitted to the network transmission interface, which includes the network transmission interface performing network data transmission and reception based on a time-division duplex communication mechanism. In addition, the intermediate data of the encoded data stream of the operating environment image obtained by performing inter-frame coding on the operating environment image using the reference image is intelligently predicted by the AI ​​prediction model based on the shape number, number of light bodies, spacing between adjacent light bodies, cross-sectional area of ​​a single light body, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame coding on the operating environment image. The data is transmitted to the network transmission interface, including: the network transmission interface performs network data transmission and reception based on the frequency division duplex communication mechanism. In addition, the intermediate data of the encoded data stream of the operating environment image obtained by performing inter-frame coding on the operating environment image using the reference image is intelligently predicted by the AI ​​prediction model based on the shape number, number of light bodies, spacing between adjacent light bodies, cross-sectional area of ​​a single light body, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame coding on the operating environment image. The data is transmitted to the network transmission interface, including the network transmission interface performing network data transmission and reception based on the WIFI communication mechanism.

[0012] In addition, in the network computer algorithm operating system, the AI ​​prediction model intelligently predicts the amount of data in the encoded data stream of the operating environment image obtained by performing inter-frame encoding on the operating environment image using the reference image based on the shape number, number, spacing between adjacent lights, cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame encoding on the operating environment image. This includes the testing and simulation of the data processing process of intelligently predicting the amount of data in the encoded data stream of the operating environment image obtained by performing inter-frame encoding on the operating environment image using the reference image based on the shape number, number, spacing between adjacent lights, cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame encoding on the operating environment image using the AI ​​prediction model based on the shape number, number, spacing between adjacent lights, cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame encoding on the operating environment image using the reference image.

[0013] The network computer algorithm operating system of this invention addresses the technical problem in existing technologies where image processing lacks flexibility and specificity due to the mismatch between the processing of monitoring images and the imaging characteristics of key components in the monitoring images, namely traffic lights. By introducing an AI prediction model, it intelligently predicts the amount of data in the encoded data stream of the operating environment image obtained by performing inter-frame coding on the operating environment image using reference images. This allows for the prediction of different inter-frame coding compression ratios obtained from selecting different reference frames, providing an important basis for reference frame selection and choosing the most suitable image processing mechanism for the associated data of different traffic lights, thereby solving the aforementioned technical problem.

[0014] It should be understood that the embodiments and examples disclosed herein are illustrative and non-limiting. The scope of the invention is defined by the appended claims rather than by the foregoing description, and all equivalent concepts and variations within that meaning are intended to be included in the claims.

Claims

1. A network computer, comprising a memory, a processor, and a network computer algorithm operating system, characterized in that, The system includes: A directional setup device is installed within the system to perform multiple learning operations on the automatic encoder neural network to obtain a fully learned automatic encoder neural network, which is then output as an AI prediction model. This includes a positive correlation between the number of learning operations performed by the automatic encoder neural network and the number of traffic lights at the set traffic intersection. A signal receiving device, installed within the system, is used to receive images of the operating environment captured by a remote network capturing device using a network transmission component. The network capturing device is installed opposite the traffic lights at a designated intersection and is used to perform image capture operations on the operating environment of the traffic lights at the designated intersection. Data acquisition equipment is used to obtain the shape number, number of lights, spacing between adjacent lights, and cross-sectional area of ​​a single light at a designated traffic intersection. An image analysis device, connected to the signal receiving device, is used to acquire various image data of the operating environment image and various image data of the reference image used when performing inter-frame coding on the operating environment image. The various image data of each image are the number of pixel columns, the number of pixel rows, the number of foreground pixels, and the coordinate values ​​of each foreground pixel. The reference image used when performing inter-frame coding on the operating environment image is an image captured by the network capture device at a previous time before the capture time of the operating environment image. The prediction execution mechanism is connected to the orientation establishment device, the data acquisition device, and the image analysis device, respectively. It is used to intelligently predict the amount of data in the encoded data stream of the operating environment image obtained by performing inter-frame encoding on the operating environment image using the reference image based on the shape number, number of lights, spacing between adjacent lights, cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used when performing inter-frame encoding on the operating environment image. A ratio determination mechanism, connected to the prediction execution mechanism, is used to determine the compression ratio obtained by performing inter-frame coding on the operating environment image using the reference image based on the data volume of the encoded data stream of the operating environment image and the original data volume of the operating environment image. The ratio determination includes: dividing the data volume of the encoded data stream of the operating environment image by the original data volume of the operating environment image to obtain the compression ratio obtained by performing inter-frame coding on the operating environment image using the reference image.

2. The network computer as described in claim 1, characterized in that: Data acquisition equipment is used to obtain the shape number, number of lights, spacing between adjacent lights, and cross-sectional area of ​​a single light at a designated traffic intersection. This includes light shape numbers with different values ​​for different shapes. The data acquisition device, used to obtain the shape number, number of lamps, spacing between adjacent lamps, and cross-sectional area of ​​a single lamp of a traffic light at a designated intersection, also includes: the cross-sectional area of ​​a single lamp of a traffic light at a designated intersection is the cross-sectional area of ​​a single lamp of a traffic light at a designated intersection in the vertical section.

3. The network computer as described in claim 2, characterized in that: An AI prediction model is used to intelligently predict the amount of data in the encoded data stream of the operating environment image obtained by performing inter-frame coding on the operating environment image using the reference image. This prediction is based on the shape and number of the traffic lights at the set traffic intersection, the number of lights, the distance between adjacent lights, the cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame coding on the operating environment image. The internal control instructions are transmitted through the instruction control channel.

4. The network computer as described in claim 3, characterized in that: The AI ​​prediction model intelligently predicts the amount of data in the encoded data stream of the operating environment image obtained by performing inter-frame coding on the operating environment image using the reference image, based on the shape and number of the traffic lights at the set traffic intersection, the number of lights, the distance between adjacent lights, the cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame coding on the operating environment image. The internal control instructions are transmitted by the instruction control channel, which is operated and maintained by the SOC chip.

5. The network computer as described in claim 3, characterized in that: The AI ​​prediction model intelligently predicts the amount of data in the encoded data stream of the operating environment image obtained by performing inter-frame coding on the operating environment image using the reference image, based on the shape and number of the traffic lights at the set traffic intersection, the number of lights, the distance between adjacent lights, the cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame coding on the operating environment image. The internal control instructions are transmitted by the instruction control channel, which is operated and maintained by an ASIC chip.

6. The network computer as described in claim 2, characterized in that: An AI prediction model is used to intelligently predict the amount of intermediate data in the encoded data stream of the operating environment image obtained by performing inter-frame coding on the operating environment image using the reference image. This data is transmitted to the network transmission interface based on the shape and number of the traffic lights at the set traffic intersection, the number of lights, the distance between adjacent lights, the cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame coding on the operating environment image.

7. The network computer as described in claim 6, characterized in that: An AI prediction model is used to intelligently predict the amount of intermediate data in the encoded data stream of the operating environment image obtained by performing inter-frame coding on the operating environment image using the reference image, based on the shape and number of the traffic lights at the set traffic intersection, the number of lights, the distance between adjacent lights, the cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame coding on the operating environment image. This intermediate data is transmitted to the network transmission interface, which performs network data transmission and reception based on a time-division duplex communication mechanism.

8. The network computer as described in claim 6, characterized in that: An AI prediction model is used to intelligently predict the amount of intermediate data in the encoded data stream of the operating environment image obtained by performing inter-frame coding on the operating environment image using the reference image, based on the shape and number of the traffic lights at the set traffic intersection, the number of lights, the distance between adjacent lights, the cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame coding on the operating environment image. This intermediate data is transmitted to the network transmission interface, which performs network data transmission and reception based on the frequency division duplex communication mechanism.

9. The network computer as described in claim 6, characterized in that: An AI prediction model is used to intelligently predict the amount of intermediate data of the encoded data stream of the operating environment image obtained by performing inter-frame coding on the operating environment image using the reference image, based on the shape and number of the traffic lights at the set traffic intersection, the number of lights, the distance between adjacent lights, the cross-sectional area of ​​a single light, various image data of the operating environment image, and various image data of the reference image used to perform inter-frame coding on the operating environment image. This intermediate data is transmitted to the network transmission interface, which performs network data transmission and reception based on the WIFI communication mechanism.