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

A Method for Detection of Rotating Targets in Remote Sensing Images

A remote sensing image and target detection technology, which is applied in the field of aerial image target detection, can solve the problems of high hardware requirements, slow speed, and insufficient detection robustness, and achieve strong anti-noise ability, reduced resource usage, and excellent detection accuracy.

Active Publication Date: 2022-05-06
WUHAN UNIV
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is to provide a method for detecting rotating targets in remote sensing images, so as to solve the problems of insufficient robustness, high hardware requirements and slow speed in the detection of optical remote sensing targets in the case of noise data

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
  • A Method for Detection of Rotating Targets in Remote Sensing Images
  • A Method for Detection of Rotating Targets in Remote Sensing Images
  • A Method for Detection of Rotating Targets in Remote Sensing Images

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0038] Embodiment: A kind of remote sensing image rotating object detection method, it proposes a kind of new Semantic Attention Feature Pyramid Structure (OcSaFPN) based on Octave convolution, on the basis of Faster R-CNN network, use Oct-ResNet-50 structure Replace the commonly used ResNet50 structure as the backbone feature extraction network (BackBone), design and implement the semantic attention feature pyramid structure for the Oct-ResNet series network to replace the commonly used FPN structure, and generate semantic pyramid feature information;

[0039]Then, the feature map {conv2, conv3, conv4, conv5} obtained from the Oct-resNet-50 network that divides high-frequency information and low-frequency information is sent into the OcSa structure and a top-down feature pyramid structure. In the semantic attention feature pyramid structure (OcSaFPN) based on Octave convolution;

[0040] Among them, the OcSa structure can squeeze the input 4 pairs of high and low frequency fe...

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 a remote sensing image rotating target detection method based on a global-local attention mechanism, comprising the following steps: S1: extracting depth features separated by high and low frequencies; S2: constructing a semantic attention pyramid; S3: extracting potential targets containing foreground and Features and coordinates of the top-ranked candidate regions; S4: perform RoIAlign pooling operation; S5: generate the final feature map; S6: perform category prediction and prediction of bounding box coordinates represented by five parameters. The invention can effectively reduce the resource usage of the hardware, increase the target detection speed, have excellent detection accuracy on remote sensing images, greatly improve the detection accuracy in a noisy environment, and have strong anti-noise ability.

Description

technical field [0001] The invention relates to a method for detecting a rotating target of a remote sensing image, and belongs to the technical field of aerial image target detection. Background technique [0002] Remote sensing image target detection is a basic task in remote sensing image processing and application, and its main content is to quickly determine the position and category of meaningful targets from large-scale remote sensing images. It has a wide range of applications in traffic control, airport monitoring, oil depot monitoring, offshore ship detection, military target discovery and other fields. [0003] With the rapid development of computer vision technology in recent years, the target detection technology based on neural network has surpassed traditional features such as histogram of directed gradients (HOG) and scale invariant feature transform (SIFT) with higher accuracy and faster speed. target detection method. At present, target detection methods ...

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 Patents(China)
IPC IPC(8): G06V10/25G06V20/10G06N3/04G06T3/40G06T3/60
CPCG06T3/60G06T3/4038G06V20/13G06V10/25G06V2201/07G06N3/045
Inventor 李成源王晨捷刘军罗斌
Owner WUHAN UNIV
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

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com