CN211606626U - Intelligent video image processing equipment based on edge calculation - Google Patents

Abstract

The utility model belongs to the technical field of the image processing technique and specifically relates to an intelligence video image processing equipment based on edge calculation, including casing and bottom plate, the bottom plate is fixed to be set up on the inside bottom plate of casing, and the bottom plate is provided with respectively and flows processing unit, intelligent image processing unit and data transmission unit through the TCP that the pin connects on the bottom plate, and the curb plate of casing left end is inside to be fixed and is provided with image channel processing unit. The utility model discloses in, for the camera integrated high performance embedded computing power, make it have the ability of marginal computing, carry out image processing promptly at the marginal end that data produced to magnanimity video data, high-efficient processing draws key information, effectively gets rid of redundant information, satisfies video data's real-time transmission, alleviates the bandwidth pressure because of cloud center calculation that the cloud computing mode brought and storage pressure and transmission course. Meanwhile, the video data are actively processed, different intelligent algorithms are loaded aiming at different scenes, and the flexibility and the reliability of the equipment are improved.

Description

An intelligent video image processing device based on edge computing

technical field

The utility model relates to the technical field of image processing, in particular to an intelligent video image processing device based on edge computing.

Background technique

Nowadays, people pay more and more attention to the security field. As an important part of the security field, video surveillance technology has attracted more and more attention, from traditional analog cameras to today's network cameras and smart cameras. More and more surveillance cameras are deployed in all corners of life to ensure people's security, but also bring many new problems.

First of all, the massive surveillance cameras not only contain important information that people pay attention to, but also include a lot of redundant information due to the characteristics of video surveillance, which brings great impact to the existing cloud computing model of centralized computing and storage. The pressure on computing and storage increases the computing processing burden and storage resource consumption of the cloud computing video server. At the same time, the transmission of massive surveillance video data to the cloud computing center also brings huge pressure on network bandwidth, resulting in the inability to guarantee the real-time surveillance video.

Secondly, the existing video surveillance system is often a passive processing mode, which needs to manually search for surveillance video data afterwards, which increases the task burden of people for the ever-increasing mass of video surveillance data. Even though some surveillance cameras are equipped with video image processing modules with basic computing capabilities, their image processing capabilities are weak, unable to meet the requirements of massive big data processing, and lack more professional real-time analysis and processing capabilities for massive data and more complex scenes. At the same time, it can only perform video processing on specific scenes, and cannot flexibly load different algorithm applications for different scenes, nor change the usage scene. Therefore, an intelligent video image processing device based on edge computing is proposed to solve the above problems.

Utility model content

The purpose of the present invention is to provide an intelligent video image processing device based on edge computing, so as to solve the problems raised in the above background technology.

To achieve the above object, the utility model provides the following technical solutions:

An intelligent video image processing device based on edge computing, comprising a casing and a bottom plate, the bottom plate is fixedly arranged on the bottom plate inside the casing, and the bottom plate is respectively provided with a TCP stream processing unit connected to the bottom plate through pins, An intelligent image processing unit and a data transmission unit, an image channel processing unit is fixed inside the side panel at the left end of the casing, a data acquisition unit is arranged on the left side of the image channel processing unit, and the data acquisition unit includes a CSI interface camera, The data acquisition unit is connected to the image channel processing unit through a data line, and the image channel processing unit is respectively connected to the TCP stream processing unit and the intelligent image processing unit through the shunt circuit, and the TCP stream processing unit and the intelligent image processing unit are respectively connected through the line. A data transmission unit, the output line of the data transmission unit extends out of the casing, an external expansion unit is arranged on the right side of the bottom of the casing, and the data transmission unit is connected to the external expansion unit through a line.

Preferably, the image channel processing unit is a LIFMD6000-main control connection board with a model of Lattice.

Preferably, the top of the intelligent image processing unit is fitted with a heat sink, and the top of the heat sink is provided with a cooling fan, the intelligent image processing unit adopts NVIDIA TX2, and includes a CPU cluster and a neural network co-processor, wherein the CPU The cluster consists of a dual-core denver2 processor and a quad-core ARM Cortex-A57.

Preferably, the TCP stream processing unit adopts a HiSilicon h3519 chip.

Preferably, the data transmission unit adopts a TPLINK router, the specific model is TP-LINK TL-SF1005+, has five 10/100M adaptive RJ45 ports, and adopts a full-duplex/half-duplex adaptive transmission mode.

Preferably, the external expansion unit includes two USB3.0 interfaces and one HDMI interface, and the HDMI interface is a high-definition multimedia interface.

Preferably, the side wall of the left casing of the CSI interface camera of the data acquisition unit is made of transparent glass.

Compared with the prior art, the utility model has the beneficial effects of integrating high-performance embedded computing capability for the camera, enabling it to have edge computing capability, and performing image processing on massive video data at the edge where the data is generated. Process and extract key information, effectively remove redundant information, meet real-time transmission of video data, and relieve cloud center computing and storage pressure and bandwidth pressure during transmission due to cloud computing mode. At the same time, it actively processes video data, reducing labor costs, and loading different intelligent algorithms for different scenarios, improving the flexibility and reliability of the equipment.

Description of drawings

1 is a schematic diagram of the overall structure of a kind of intelligent video image processing equipment based on edge computing of the present invention;

FIG. 2 is a schematic diagram of the connection structure of each component of an edge computing-based intelligent video image processing device of the present invention.

In the figure: 1. Image channel processing unit; 2. Heat sink; 3. Cooling fan; 4. Data transmission unit; 5. External expansion unit; 6. Housing; 7. Bottom plate; 8. TCP stream processing unit; 9. An intelligent image processing unit; 10. A data acquisition unit.

Detailed ways

The technical solutions in the embodiments of the present utility model will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present utility model. Obviously, the described embodiments are only a part of the embodiments of the present utility model, rather than all the implementations. example. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present invention.

Please refer to Figure 1-2, the utility model provides a technical solution:

An intelligent video image processing device based on edge computing, comprising a casing 6 and a bottom plate 7, the bottom plate 7 is fixedly arranged on the bottom plate inside the casing 6, and the bottom plate 7 is respectively provided with a TCP stream processing connected to the bottom plate 7 through pins. Unit 8, intelligent image processing unit 9 and data transmission unit 4, an image channel processing unit 1 is fixed inside the side panel at the left end of the housing 6, and a data acquisition unit 10 is provided on the left side of the image channel processing unit 1. The data acquisition unit 10 includes CSI interface camera, the data acquisition unit 10 is connected to the image channel processing unit 1 through the data line, and the image channel processing unit 1 is respectively connected to the TCP stream processing unit 8 and the intelligent image processing unit 9 through the shunt line, and the TCP stream processing unit 8 and the intelligent image processing unit 9 are respectively connected to the data transmission unit 4 through a line, the output line of the data transmission unit 4 extends out of the casing 6, an external expansion unit 5 is provided on the right side of the bottom of the casing 6, and the data transmission unit 4 is connected to the external expansion unit 5 through a line.

Among them, the image channel processing unit 1 is a LIFMD6000-main control connection board with a model of Lattice.

Among them, the top of the intelligent image processing unit 9 is fitted with a heat sink 2, and the top of the heat sink 2 is provided with a cooling fan 3. The intelligent image processing unit 9 adopts NVIDIA TX2, including a CPU cluster and a neural network co-processor, wherein the CPU cluster is composed of It consists of dual-core denver2 processor and quad-core ARM Cortex-A57.

Among them, the TCP stream processing unit 8 adopts the HiSilicon h3519 chip.

Among them, the data transmission unit 4 adopts a TPLINK router, the specific model is TP-LINK TL-SF1005+, has 5 10/100M adaptive RJ45 ports, and adopts the full-duplex/half-duplex adaptive transmission mode.

Among them, the external expansion unit 5 includes two USB3.0 interfaces and an HDMI interface, and the HDMI interface is a high-definition multimedia interface.

The side wall of the left casing 6 of the CSI interface camera of the data acquisition unit 10 is made of transparent glass.

Working principle: The data acquisition unit 10 contains a CSI interface camera to collect video image data in real time. After the acquisition, the image is transmitted to the rear image channel processing unit, where the video data transmitted by the data acquisition unit 10 is processed. It is respectively transmitted to the intelligent image processing unit 9 of the TCP stream processing unit 8, and the TCP stream processing unit 8 contains an image coding processing module, which provides video coding services to the video stream transmitted by the channel processing unit 4, and provides video enhancement, image splicing , video overlay and other preprocessing operations, while providing output video TCP stream service, through the media player and terminal for real-time playback, the intelligent image processing unit 9 includes image processing capabilities, real-time image channel processing unit 4 The video stream transmitted by The data information and the information transmitted by the back-end monitoring terminal are comprehensively analyzed and processed, the processing results are transmitted backwards, and the required algorithms are loaded according to the application scenario requirements for image feature analysis, image classification, image recognition and tracking, and storage and extraction of specific image data, etc. Image processing step, send the processing results back to the monitoring terminal, and output high-definition H264 code stream for video recording, including a neural network co-processor, which can run huge and complex deep neural networks, and can load different neural networks developed by users. The algorithm performs intelligent video analysis on the video stream at the edge end, and has powerful artificial intelligence computing capabilities. The algorithm includes loading corresponding algorithms in different application scenarios. In the intelligent image processing unit 9, according to requirements, the algorithm is modified and iterated, so that the intelligent video image The processing device has the versatility in various scenarios, wherein the heat sink 2 is located on the intelligent image processing unit 9, and the cooling fan 3 is located on the heat sink 2 to dissipate the heat brought by the intelligent image processing unit 9 due to high-speed operation, which increases the For the protection function and heat dissipation rate of the intelligent image processing unit 9, the data transmission unit 4 performs data transmission, forwarding and exchange processing on the video stream data from the TCP stream processing unit 8 and the processing result information from the intelligent image processing unit 9 and the back-end user, The multimedia player of the back-end device can view the video stream and processing results in real time through the access address. The external expansion unit 5 is provided with multiple data interaction interfaces, including USB and HDMI display interfaces, and has reserved assembly space to meet the needs of intelligent video image processing equipment. For functional expansion requirements, the shell 6 is an inflexible rigid shell, which is used to wrap the above structure to protect it from partial damage, including dustproof and waterproof functions, which increases the protection of the above components. The back cover is used for equipment maintenance. There are fixing holes at the bottom of the shell 6. The brackets are connected with the fixing holes so that the equipment brackets are fixedly connected to the outer ring of the shell. The heat sink 2, the cooling fan 3, the data transmission unit 4, the TCP stream processing unit 8, and the intelligent image processing unit 9 are fixedly connected. At the same time, the intelligent video image processing device can be connected to wireless and wired networks, and the external expansion unit 5 can also be connected externally. The 4G module is connected to the 4G network, and is externally stabilized through line 11. The 12V voltage of the voltage source provides power for the device. The intelligent video image processing device is especially suitable for customization and personalized scene services, and the program can be configured according to different needs.

Although the embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes and modifications can be made to these embodiments without departing from the principles and spirit of the present invention , alternatives and modifications, the scope of the present invention is defined by the appended claims and their equivalents.

Claims (7)

1. An intelligent video image processing device based on edge calculation, comprising a shell (6) and a bottom plate (7), characterized in that: the bottom plate (7) is fixedly arranged on a bottom plate inside the shell (6), the bottom plate (7) is respectively provided with a TCP stream processing unit (8), an intelligent image processing unit (9) and a data transmission unit (4) which are connected to the bottom plate (7) through pins, the inside of a side plate at the left end of the shell (6) is fixedly provided with an image channel processing unit (1), the left side of the image channel processing unit (1) is provided with a data acquisition unit (10), the data acquisition unit (10) comprises a CSI interface camera, the data acquisition unit (10) is connected with the image channel processing unit (1) through a data line, the image channel processing unit (1) is respectively connected with the TCP stream processing unit (8) and the intelligent image processing unit (9) through a shunt line, the TCP stream processing unit (8) and the intelligent image processing unit (9) are respectively connected with the data transmission unit (4) through a line, the output line of the data transmission unit (4) extends out of the shell (6), an external expansion unit (5) is arranged on the right side of the bottom of the shell (6), and the data transmission unit (4) is connected with the external expansion unit (5) through a line.

2. An intelligent video image processing device based on edge calculation according to claim 1, characterized in that: the image channel processing unit (1) is a LIFMD 6000-master control connecting plate with the model number of Lattice.

3. An intelligent video image processing device based on edge calculation according to claim 1, characterized in that: the intelligent image processing unit (9) is provided with fin (2) in the laminating of intelligence image processing unit (9) top, fin (2) top is provided with radiator fan (3), intelligence image processing unit (9) adopts NVIDIA TX2, including CPU cluster and neural network coprocessor, wherein, the CPU cluster comprises dual-core denver2 treater and four-core ARM Cortex-A57.

4. An intelligent video image processing device based on edge calculation according to claim 1, characterized in that: the TCP stream processing unit (8) adopts a Haesi h3519 chip.

5. An intelligent video image processing device based on edge calculation according to claim 1, characterized in that: the data transmission unit (4) adopts a TPLINK router, has a specific model of TP-LINK TL-SF1005+, is provided with 5 10/100M self-adaptive RJ45 ports, and adopts a full-duplex/half-duplex self-adaptive transmission mode.

6. An intelligent video image processing device based on edge calculation according to claim 1, characterized in that: the external expansion unit (5) comprises two USB3.0 interfaces and an HDMI interface, and the HDMI interface is a high-definition multimedia interface.

7. An intelligent video image processing device based on edge calculation according to claim 1, characterized in that: the side wall of the left shell (6) of the CSI interface camera of the data acquisition unit (10) is made of transparent glass.

Images