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

Robust sonar target detection method based on dual-path feature fusion network

A feature fusion and target detection technology, applied in neural learning methods, biological neural network models, instruments, etc., can solve the problem of susceptible to formation noise interference, false detection or missed detection of small underwater targets, inability to achieve high precision, strong robustness Problems such as the real-time performance of the rod system, to achieve the effect of improving the detection effect and improving the detection speed

Active Publication Date: 2020-12-08
HANGZHOU DIANZI UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0009] The contrast and signal-to-noise ratio are low, and they are susceptible to formation noise interference. In this case, the current underwater target detection and recognition methods still have many bottlenecks, such as incomplete or slow extraction of target features in sonar images, and small underwater targets. Due to low contrast and signal-to-noise ratio, it is falsely detected or missed, and it is impossible to achieve high precision, strong robustness, and system real-time performance at the same time. Therefore, sonar target detection and recognition methods also need continuous development and innovation

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
  • Robust sonar target detection method based on dual-path feature fusion network
  • Robust sonar target detection method based on dual-path feature fusion network
  • Robust sonar target detection method based on dual-path feature fusion network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0047] The present invention will be further described below in conjunction with accompanying drawing.

[0048] Such as figure 1 As shown, the specific steps of the robust sonar target detection method based on the dual-path feature fusion network are as follows:

[0049] Step 1, such as figure 2 As shown, a dual-path feature fusion network is built.

[0050] The dual-path feature fusion network consists of an initial convolutional layer, a dual-path module (Dpn), a fusion transition module, a densely connected module, and a final convolutional layer. Such as image 3 Shown; where the initial convolutional layer is a 3×3 convolutional layer, which is used to reduce the feature map size from 416×416 to 208×208. There are five dual path modules. The five dual-path modules are arranged in sequence. Each dual-path module consists of a 3×3 convolutional layer and two 1×1 convolutional layers. The 3×3 convolutional layers within the dual-path module are placed between two 1×...

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 robust sonar target detection method based on a dual-path feature fusion network. The traditional image processing method uses the method of image segmentation to distinguish the background from the target; the present invention is as follows: 1. Build a dual-path feature fusion network. 2. Train the dual-path feature fusion network obtained in step 1. 3. Sonar image generation and feature extraction. Fourth, classify and detect the target frame of the sonar image in combination with the default frame. The present invention integrates deep learning technology into target detection, inputs sonar images generated by sonar data into a network model, and completes feature extraction, target detection, and target classification in the model at one time, thereby greatly improving the detection speed. The present invention can extract more deep-level features, and fundamentally optimizes target classification and regression. The present invention adopts multi-scale dense connection to fuse multi-level features and improve the detection effect of small and medium targets.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence and underwater acoustic electronic information, and in particular relates to a robust sonar target detection method based on a dual-path feature fusion network. Background technique [0002] With the continuous iteration and development of computer science and technology, underwater detection technology has been greatly promoted in recent years, and has a wide range of applications in military and civil fields, such as military confrontation, dangerous target screening, target tracking, etc.; in other The field also plays an important role in underwater rescue, seabed resource exploration, tracking and protection of endangered species, and seabed modeling. [0003] Underwater target detection and recognition is an important part of modern sonar systems and underwater acoustic countermeasures. It is the research focus of each country's maritime security. It has been widely concerned...

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): G06K9/00G06K9/32G06K9/62G06N3/04G06N3/08
CPCG06N3/084G06V10/25G06V2201/07G06N3/045G06F2218/08G06F2218/12G06F18/23213G06F18/253
Inventor 孔万增贾明洋洪吉晨张建海周文晖
Owner HANGZHOU DIANZI 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