CN213338762U

CN213338762U - Edge calculation system based on animal identification

Info

Publication number: CN213338762U
Application number: CN202022596696.2U
Authority: CN
Inventors: 王楠; 程川; 赵永波
Original assignee: Chengdu Fruit Picking Technology Co ltd
Current assignee: Chengdu Fruit Picking Technology Co ltd
Priority date: 2020-11-11
Filing date: 2020-11-11
Publication date: 2021-06-01
Anticipated expiration: 2030-11-11

Abstract

The utility model discloses an edge computing system based on animal identification, which comprises an image acquisition device, an intranet switch, edge computing equipment and a cloud computing platform; the image acquisition device is connected with the intranet switch; the intranet switch is connected with the edge computing equipment; the edge computing equipment is communicated with the cloud computing platform through a mobile communication network, uploads data to the cloud computing platform to be compared, analyzed and processed, and monitors and deeply learns the zoo stadium on line. The utility model discloses a deploy marginal computing device in every zoo venue, carry out animal type discernment and quantity calculation to every cage in the beast house venue to export the result of calculation to central server in real time. The central server compares the monitoring data in real time, immediately gives an alarm when abnormal data are found, and outputs abnormal pictures, so that the safety supervision of the zoo is enhanced, the manual operation cost is reduced, and the efficiency of the zoo for processing abnormal events is improved.

Description

Edge calculation system based on animal identification

Technical Field

The utility model relates to a safety prevention and control technical field especially relates to an edge computing system based on animal identification.

Background

With the improvement of living standard of people, zoos and wild zoos with different scales are built in many large and medium-sized cities in China, the number of people visitors to the zoos is increased continuously, and the visiting activities play a positive role in enriching the amateur life of people, so that more and more people have the opportunity to watch various wild animals and a plurality of fierce beasts which can be seen only on a screen of a film and television.

Because each venue animal distributes comparatively dispersedly in the zoo, bring very big work load to the management and control of animal safety, in beast area very much, it is great to the control of animal quantity and discernment demand, and the animal house in every venue all is independent management and control simultaneously. The existing security system cannot identify the type and the number of animals in each cage in a barn, cannot give an alarm in time when abnormality occurs, and cannot enable a supervisor to know detailed information of the abnormal event at the first time.

SUMMERY OF THE UTILITY MODEL

An object of the utility model is to overcome prior art's not enough, provide an edge computing system based on animal identification, deploy edge computing equipment in every venue, carry out animal type discernment and quantity calculation to every cage in the animal house venue to export the result of calculation to central server in real time. The central server compares the monitoring data in real time, immediately gives an alarm if the monitoring data is abnormal, and outputs an abnormal picture.

The purpose of the utility model is realized through the following technical scheme:

an edge computing system based on animal identification comprises an image acquisition device, an intranet switch, edge computing equipment and a cloud computing platform; the image acquisition device is connected with the intranet switch; the intranet switch is connected with the edge computing equipment; the edge computing equipment is communicated with the cloud computing platform through a mobile communication network, uploads data to the cloud computing platform to be compared, analyzed and processed, and monitors and deeply learns the zoo stadium on line.

Specifically, the cloud computing platform comprises a central server, an image storage server and a deep learning server; the central server is used for comparing the monitoring data in real time and capturing abnormal data; the image storage server is used for storing the image data uploaded by the edge computing equipment, forming an image training data set and sending the image training data set to the deep learning server for deep learning training; and the deep learning server is used for performing deep learning training according to the image training data set.

The intelligent tour bus tracking system comprises a WIFI tracking device, is arranged on a tour bus of a zoo, and is communicated with an intranet switch through a wireless network for tracking and positioning the geographic position of the tour bus.

Specifically, still include pronunciation reputation alarm device, be connected with WIFI tracer for send pronunciation and sound reputation warning according to the instruction that intranet switch forwarded.

Specifically, the system also comprises a self-powered device which is connected with the image acquisition device; the self-powered device consists of photovoltaic power generation equipment, a wind driven generator and an electric storage device, wherein the input end of the electric storage device is respectively connected with the photovoltaic power generation device and the wind driven generator; the output end of the electric power storage device is connected with the image acquisition device.

Specifically, the edge computing device comprises a session module, a data parallel computing module, a data segmentation module and a data storage module; the session module is used for generating an empty graph to establish a session and realizing information interaction between the client and the TF system; the data parallel computing module is used for computing the acquired image data in a data parallel mode and carrying out structural processing on the computing result; the data segmentation module is used for carrying out data segmentation on the uploaded acquired image data and updating the parameters of each model; the data storage module is used for storing the structured image data.

The utility model has the advantages that: the utility model discloses a deploy marginal computing device in every zoo venue, carry out animal type discernment and quantity calculation to every cage in the beast house venue to export the result of calculation to central server in real time. The central server compares the monitoring data in real time, immediately gives an alarm when abnormal data are found, and outputs abnormal pictures, so that the safety supervision of the zoo is enhanced, the manual operation cost is reduced, and the efficiency of the zoo for processing abnormal events is improved.

Drawings

Fig. 1 is a system architecture diagram of the present invention.

Detailed Description

In order to clearly understand the technical features, objects, and effects of the present invention, embodiments of the present invention will be described with reference to the accompanying drawings.

In this embodiment, as shown in fig. 1, an edge computing system based on animal identification includes an image acquisition device, an intranet switch, an edge computing device, and a cloud computing platform; the image acquisition device is connected with the intranet switch; the intranet switch is connected with the edge computing equipment; the edge computing equipment is communicated with the cloud computing platform through a mobile communication network, uploads data to the cloud computing platform to be compared, analyzed and processed, and monitors and deeply learns the zoo stadium on line. The image acquisition device comprises a plurality of camera groups.

The system further comprises a WIFI tracking device, a voice sound-light alarm device and a self-powered device. The WIFI tracking device is arranged on a sightseeing bus of a zoo, and is in mutual communication with the intranet switch through a wireless network, and is used for tracking and positioning the geographic position of the sightseeing bus. The voice sound-light alarm device is connected with the WIFI tracking device and used for sending out voice and sound-light alarms according to the instruction forwarded by the intranet switch. The self-powered device is connected with the image acquisition device; the self-powered device consists of photovoltaic power generation equipment, a wind driven generator and an electric storage device, wherein the input end of the electric storage device is respectively connected with the photovoltaic power generation device and the wind driven generator; the output end of the electric power storage device is connected with the image acquisition device.

The utility model discloses in, the position data of sightseeing tour bus is acquireed in real time to system's accessible wiFi tracer, and pass to the analysis of central server on through the intranet switch, find the tour bus and animal cage distance near when fixing a position through wiFi tracer, or when having the animal to flee outside the cage through image acquisition device discovery, the early warning instruction of settlement is issued to the intranet switch to central server, the intranet switch forwards the early warning instruction to wiFi tracer, the pronunciation audible-visual annunciator through connecting reminds the driver control vehicle distance of tour bus or commands the driver to stop to the cage region of escaping the animal place outward, avoid taking place animal injury people accident. Furthermore, the central server can also set the collected information of the cage area where the animal escapes or the information of the position of the area where the animal escapes in the early warning instruction.

The utility model supports automatic loading of animal identification model files and other configuration information, and the system supports remote updating of animal identification model files, can automatically download to the local and automatically update identification algorithms. The method supports that parameters such as video stream addresses, acquisition frequency (seconds), video frame rates and the like can be respectively configured for different camera groups. Different camera groups are respectively managed and controlled according to the equipment IP, and the remote control camera group is supported to acquire, analyze and set the frequency and start the disabled camera group. In addition, a plurality of recognition results are output as structured data, including recognition similarity, recognition boundary position, recognition category and the like, and the results can be actively reported to the central server in real time.

The basic principles and the main features of the invention and the advantages of the invention have been shown and described above. It will be understood by those skilled in the art that the present invention is not limited to the above embodiments, and that the foregoing embodiments and descriptions are provided only to illustrate the principles of the present invention without departing from the spirit and scope of the present invention. The scope of the invention is defined by the appended claims and equivalents thereof.

Claims

1. An edge computing system based on animal identification is characterized by comprising an image acquisition device, an intranet switch, edge computing equipment and a cloud computing platform; the image acquisition device is connected with the intranet switch; the intranet switch is connected with the edge computing equipment; the edge computing equipment is communicated with the cloud computing platform through a mobile communication network, uploads data to the cloud computing platform to be compared, analyzed and processed, and monitors and deeply learns the zoo stadium on line.

2. The animal identification-based edge computing system of claim 1, wherein the cloud computing platform comprises a central server, an image storage server, and a deep learning server; the central server is used for comparing the monitoring data in real time and capturing abnormal data; the image storage server is used for storing the image data uploaded by the edge computing equipment, forming an image training data set and sending the image training data set to the deep learning server for deep learning training; and the deep learning server is used for performing deep learning training according to the image training data set.

3. The animal identification-based edge computing system of claim 1, further comprising a WIFI tracking device, disposed on a sightseeing bus of a zoo, and communicating with the intranet switch via a wireless network for tracking and locating the geographical position of the sightseeing bus.

4. The animal identification-based edge computing system of claim 3, further comprising a voice sound and light alarm device connected to the WIFI tracking device for issuing voice and sound and light alarms according to the instruction forwarded by the intranet switch.

5. The animal identification-based edge computing system of claim 1, further comprising a self-powered device connected to the image capture device; the self-powered device consists of photovoltaic power generation equipment, a wind driven generator and an electric storage device, wherein the input end of the electric storage device is respectively connected with the photovoltaic power generation device and the wind driven generator; the output end of the electric power storage device is connected with the image acquisition device.