An intelligent image recognition method and system for a monitoring system of a nature reserve

A technology of nature reserves and monitoring systems, applied in the field of computer vision, can solve the problems of not establishing an accuracy baseline and achieve the effect of improving accuracy

Active Publication Date: 2022-02-15
常灵逸
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] At present, computer vision rarely involves attempts in natu

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • An intelligent image recognition method and system for a monitoring system of a nature reserve
  • An intelligent image recognition method and system for a monitoring system of a nature reserve
  • An intelligent image recognition method and system for a monitoring system of a nature reserve

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0037] Embodiment 1: The embodiment of the present invention provides a solution for video surveillance in national nature reserves, see figure 1 , the method includes:

[0038] With high-definition image data as input, through the ImageCrop module of image cutting, output (2*3-1)*(2*3-1)=25 sub-image images;

[0039] Each sub-image is used as an input, and the corresponding tensor is output through the deep learning model Model1 (the contents of the tensor are the corresponding probabilities of scenery, people, fishing, and sailing);

[0040] Then these 5*5 tensors are used as input, and after being sent to the neural network Model2, the final image classification result is output.

[0041] It should be noted that the pixels of high-definition images are fixed, all of which are 1280*1920, so they can be cropped according to the specified cropping strategy.

[0042] Furthermore, after cropping, the format should be 5*5 tensors, and the shape of each tensor is 360*640*3, and ...

Embodiment 2

[0057] The embodiment of the present invention provides a system for video surveillance in national nature reserves, which can realize the software environment on which the solution for video surveillance in national nature reserves relies and provide stable, mature, and standardized Web services. See figure 2 , the video monitoring system of the National Nature Reserve includes:

[0058] The process Process1 obtains the frame data of the specified camera from the video server VideoServer every 1 second in the form of RTSP stream, that is, the original image; then Process1 saves the frame data to a specific path in the disk Disk;

[0059] The process Process2 independently reads the frame data from this path and uploads it to the specified interface of the National Nature Reserve Video Surveillance Service MonitorServer in the form of http. After the interface returns that the upload is successful, it deletes the successfully uploaded pictures on the disk.

[0060] The client...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses an intelligent image recognition method and system for a monitoring system of a nature reserve. The original image is obtained based on the monitoring system of a nature reserve, and the original image is cut by a window to obtain a data set after cutting the original image. Among them, The data set consists of several sub-images, and the format of each sub-image is 360*640; based on the Xception model and the training weight of the Xception model, by inputting several sub-image data sets after cutting the original image, the corresponding sub-image is obtained. Tensor; then the tensor of this group of sub-images is used as input, and is sent into the fully connected neural network model to obtain the tensor of the original image, thereby obtaining the true category of the original image. The present invention makes the behaviors of tourists, fishing and sailing. The early warning achieved an accuracy rate of 91.4%.

Description

technical field [0001] The invention belongs to the field of computer vision and relates to an intelligent image recognition method and system for a monitoring system of a nature reserve. Background technique [0002] With the rapid development of AI technology and subversive applications in certain fields, AI's position in industry is becoming more and more obvious, and its popularity and penetration in business are increasing. As one of the most important components of AI technology, computer vision has been applied in many aspects of life. The accuracy of AI determines the degree to which an AI technology is recognized and promoted by the market, so AI technology with high accuracy is very important. [0003] At present, computer vision rarely involves attempts in nature reserves, let alone establish an accuracy baseline in this field. One reason is that most of the monitoring devices used to capture images in nature reserves are installed at a distance of more than ten...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06V20/10G06V10/764G06V10/82G06K9/62G06N3/04G06N3/08G06T7/11
CPCG06T7/11G06N3/04G06N3/08G06T2207/20021G06T2207/20081G06V20/38G06F18/241
Inventor 常灵逸
Owner 常灵逸
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products