Ship detection method based on deformable fast convolutional neural network

A neural network and fast convolution technology, applied in the field of image processing, to achieve the effect of improving resolution, good detection speed and accuracy, and enhancing the ability of extraction

Active Publication Date: 2021-10-01
XIAMEN UNIV
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] Although the above methods can effectively detect most of the ships in the video or image, there is still a relatively large room for improvement in the detection of speed, accuracy and small-sized targets

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
  • Ship detection method based on deformable fast convolutional neural network
  • Ship detection method based on deformable fast convolutional neural network
  • Ship detection method based on deformable fast convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0077] The implementation examples and effects of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0078] refer to figure 1 , the implementation steps of the present invention are as follows:

[0079] Step 1, get training set images.

[0080] Existing ship detection method training set images from China Computer Federation (CCF) [6] The game data provided, the data set contains various scenes such as different lighting, different shooting angles, and different weather. Most of the ships are small targets with complex backgrounds, including ports, islands, and sea interference ships. The specific process is as follows:

[0081] In the data set collection stage, the collected data sets are screened to obtain training set images, and each image in the above training set is scaled to a size of 1024×1024 to form the final training data set;

[0082] Step 2, modify the settings in the original Faster R-CNN network, and desig...

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

Ship detection method based on deformable fast convolutional neural network, involving image processing. The detection method includes a model training phase and a ship detection phase. It can be used in the civilian field, instead of manually classifying and detecting ships in specific ports, harbors, and sea areas in real time. It can be used for video surveillance or ship detection in images, and can also be used for ship detection and monitoring of military ports. It can detect military intelligence early, and provide Our military command provides combat basis and holds the initiative on the battlefield. Through the improvement of Faster R‑CNN, an end-to-end method that takes both speed and accuracy into account, and according to the unique properties of ship targets, the Faster R‑CNN basic network, RoI‑wise subnetwork and loss function are modified to obtain deformable fast The model structure of the detection network of the convolutional neural network, the experimental results show that it has better detection speed and accuracy than the original Faster R‑CNN method.

Description

technical field [0001] The invention relates to image processing, in particular to a ship detection method based on a deformable fast convolutional neural network. Background technique [0002] Sea ship target detection is a special scene of general target detection. It is the process of finding and locating targets from complex coastal and ocean backgrounds. It has a very wide application prospect and use value. [0003] In recent years, many scholars at home and abroad have done a lot of work on the research of ship target detection algorithm for optical images. It can be roughly divided into detection algorithms based on classification learning, feature discrimination, and Hough voting. However, this type of method is usually sensitive to illumination changes, shooting angles, edge noise, etc., which is easy to cause false alarms, and the algorithm is not robust. [0004] Deep learning is one of the current mainstream machine learning methods, and has achieved great suc...

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): G06K9/00G06K9/62
CPCG06V20/00G06V2201/07G06F18/214G06F18/24
Inventor 曲延云张怡晨丁瑶陈蓉李翠华
Owner XIAMEN UNIV
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