Optical remote sensing image airplane target detection method based on rotary positioning network

A technology for optical remote sensing images and aircraft targets, applied in the field of image detection, can solve the problems of insufficient data information, poor detection effect of dense scenes, poor training effect, etc., achieve high learning efficiency, improve detection accuracy and recall rate, The effect of improving the detection accuracy

Inactive Publication Date: 2020-05-08
NANJING UNIV OF POSTS & TELECOMM
View PDF0 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Aircraft target detection often has problems such as insufficient data inf

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
  • Optical remote sensing image airplane target detection method based on rotary positioning network
  • Optical remote sensing image airplane target detection method based on rotary positioning network
  • Optical remote sensing image airplane target detection method based on rotary positioning network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0033] The data set used in this embodiment is divided, wherein 20% is divided into a test set and 80% is a training set. In order to make the hyperparameter setting more accurate during training, 20% of the training set is divided into a verification set to ensure The robustness of the network model is better.

[0034] After the data set is divided, such as image 3 As shown, start sending pictures to the network.

[0035] This method comprises the following steps,

[0036] Step 1: First initialize the parameters except the feature extraction network, and then send the dataset images into the feature extraction network ResNet in batches, and the feature extraction network ResNet uses the official weight value. The three-dimensional array of the picture is extracted by the residual module of the ResNet network and becomes a deeper dimensional array. This array represents the feature information in the picture, including context information, receptive field information and ai...

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

An optical remote sensing image airplane target detection method based on a rotary positioning network comprises the following steps that S1, a picture is input into a ResNet feature network, multi-dimensional feature information is extracted, and a feature map represented by a multi-dimensional array is obtained; S2, the feature map is inputted into a cumulative feature pyramid network, fusing high-level semantics and bottom-level semantics, and extracting a receiving domain and context information of high-level semantics and target position information of the bottom-level semantics; S3, a RPN sub-network generates an anchor box in the features output by the cumulative feature pyramid network, boundary regression and foreground classification are carried out on the anchor box, and a horizontal proposal box is obtained; and S4, the horizontal proposal frame is sent into a rotation area positioning network for rotation and zooming to generate a rotation target frame, and finally an aircraft target rotation detection frame is outputted. According to the invention, the cumulative feature pyramid network is used for feature fusion, and the rotating rectangular bounding box is adopted,so that the background redundancy after framing of the target is reduced, and the detection result is more accurate in a dense scene.

Description

technical field [0001] The invention belongs to the technical field of image detection, and in particular relates to an optical remote sensing image aircraft target detection method based on a rotation positioning network. Background technique [0002] Aircraft target detection is an important way to evaluate the function and importance of the airport and grasp the enemy's dynamics. With the maturity of remote sensing image technology, the resolution of remote sensing images becomes higher and the amount of information contained increases. Therefore, more and more aircraft target detection methods use remote sensing images as the basic structure of detection. Under this trend, machine learning methods that have become popular in recent years have also been widely used in this field. [0003] Although the technology is becoming more and more mature, there are still many difficulties and challenges in the aircraft target detection method: 1. The aircraft on the airport apron ...

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): G06K9/00G06K9/46G06K9/62
CPCG06V20/13G06V10/40G06F18/241G06F18/214
Inventor 周亮李陈刘希鹏康彬陈建新
Owner NANJING UNIV OF POSTS & TELECOMM
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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap