Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Method for detecting and identifying dense weak and small targets in wide remote sensing image

A technology for small targets and remote sensing images, which is applied in the field of remote sensing image processing to achieve the effects of improving detection and recognition capabilities, improving robustness, and avoiding falling into local optimum

Pending Publication Date: 2022-07-29
NO 54 INST OF CHINA ELECTRONICS SCI & TECH GRP
View PDF0 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] In view of this, the present invention provides a detection and recognition method of dense weak and small targets in wide-format remote sensing images, which can effectively solve the problem of detection and recognition of dense and weak targets in optical remote sensing images

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
  • Method for detecting and identifying dense weak and small targets in wide remote sensing image
  • Method for detecting and identifying dense weak and small targets in wide remote sensing image
  • Method for detecting and identifying dense weak and small targets in wide remote sensing image

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0043] The present invention will be further described below with reference to the accompanying drawings and embodiments, and the present invention includes but is not limited to the following embodiments.

[0044] A method for detecting and identifying dense, weak and small targets in a wide-format remote sensing image, comprising the following steps:

[0045] (1) Data preparation: data augmentation is performed on the image data in the remote sensing image dataset, and more sample data with more scales are obtained by means of scaling and other means;

[0046] (2) Feature extraction: Extract the training samples from the training sample set, add the Transformer encoder, and improve the YOLOv5 network to extract the multi-scale features of the training samples;

[0047] (3) Multi-scale prediction: Add weak and small target detection heads for detection and identification of weak and small targets, and combine the original detection heads to achieve multi-scale target predicti...

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 provides a method for detecting and identifying dense weak and small targets in a wide remote sensing image. According to the method, sample amplification is carried out by means of Mosaic, rotation, MixUp, zooming and the like, and abundant and multi-scale training samples are obtained; secondly, improving a YOLOv5 network by utilizing Transform, adding a weak and small target detection head, enhancing the characterization capability of the model on global information and context information in an image, and improving the weak and small target detection and recognition performance in a dense scene; meanwhile, a multi-model fusion prediction method is used, the prediction precision is improved, and the model is prevented from falling into local optimum; and finally, detecting and identifying dense weak and small targets in the wide remote sensing image by using a block detection method. According to the method, multi-model fusion prediction is used, the model target detection and identification capability can be improved, and the model is prevented from falling into local optimum.

Description

technical field [0001] The invention relates to the field of remote sensing image processing, in particular to a method for detecting and identifying dense and weak targets in wide-format remote sensing images, which can be used for detecting and identifying weak and small targets in remote sensing images. Background technique [0002] Remote sensing image target detection and recognition are widely used in various fields such as national defense and military. However, due to the inherent characteristics of some targets, such as small physical size, fuzzy edge information, dense distribution, etc., they appear weak and small in remote sensing images, which greatly increases the difficulty of target detection and identification. In practical scenes, there are common applications of detection and recognition of dense and small targets, such as vehicle recognition in remote sensing images, tent detection, etc. Therefore, the detection and recognition of weak and small targets i...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06V20/10G06V10/25G06V10/774G06V10/80G06N3/04
CPCG06V20/10G06V10/25G06V10/774G06V10/806G06V2201/08G06N3/045
Inventor 张萌月王港陈金勇王敏武晓博
Owner NO 54 INST OF CHINA ELECTRONICS SCI & TECH GRP
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products