Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

A ship detection method and system based on scene multi-dimensional features

A ship detection and multi-dimensional feature technology, applied in the field of computer vision, can solve problems such as insufficient accuracy, and achieve the effects of improving accuracy and speed, saving supervision costs, and high robustness

Active Publication Date: 2018-07-20
ZHUHAI DAHENGQIN TECH DEV CO LTD
View PDF3 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

But the accuracy is still not good enough

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 ship detection method and system based on scene multi-dimensional features
  • A ship detection method and system based on scene multi-dimensional features
  • A ship detection method and system based on scene multi-dimensional features

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0031]The invention proposes a ship detection method based on a deep learning network combined with scene features. First build an image sample library, and mark the ship images to obtain enough samples. Then, the coastline is obtained by edge detection and Hough transform, and the edge detection result is used as the fourth dimension of the image to construct a deep learning network to convolve the image. Then, a sliding window is used to generate a region proposal box in the area between the coastlines, because in the roundabout image, the boat will only appear on the water surface, and the region proposal methods of other deep learning methods are all region proposals for the entire image. Then use the true value of the ship position to get the loss function of the proposed box, train the entire network, and output the trained model. Finally, use the trained model to perform ship detection on the test data. It mainly includes four processes: sample library construction, c...

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 present invention provides a ship detection method and system based on multi-dimensional features of the scene, including building a ship image sample library, extracting all edges of each frame of image as the fourth dimension of the image; extracting the coastline, making the sea surface area the area where the ship appears; constructing The Faster RCNN-like convolutional network is used as a deep learning network, and the sample data is input into the deep learning network; the RPN network is constructed, and the sliding window is used to generate area suggestion boxes of different sizes in the area where ships appear. Position training model; based on the trained model on the detection image, ship detection is performed on the part between the coastlines. The present invention avoids the interference of land houses by extracting the coastline, and only proposes areas for the ship area, which improves the accuracy and speed of the area suggestion frame; and adds edge features as the fourth dimension of the image in the target detection, which improves the detection precision and speed.

Description

technical field [0001] The invention belongs to the field of computer vision, and relates to a ship detection method and system for constructing a deep learning network model based on scene multidimensional features. Background technique [0002] In today's society, video surveillance cameras are ubiquitous, and multiple surveillance images are displayed simultaneously on the video wall of the surveillance center. If only relying on human eyes to observe and detect, it is easy to miss abnormal events. Studies have shown that professional monitoring personnel will miss 95% of the behavior after 22 minutes if they only monitor 2 monitors, and cannot effectively prevent the occurrence of criminal behavior in advance. The intelligent monitoring probe improves the active early warning capability of the real-time monitoring system. When a relevant dangerous situation is detected, an early warning is issued, which is conducive to the relevant departments to take timely measures. O...

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): G06K9/46G06K9/62G06N3/04G06V10/25G06V10/44G06V10/764G06V10/774
CPCG06V10/44G06N3/045G06F18/23213G06F18/24G06F18/214G06T7/12G06T7/11G06T7/73G06T2207/20061G06T2207/20081G06T2207/20084G06T2207/30232G06N3/08G06N20/10Y02A10/40G06V20/52G06V10/25G06V10/82G06V10/763G06V10/764G06V10/774G06T7/215G06T7/251G06T7/262G06T7/269G06T7/60G06T7/75G06V20/47G06V20/49G06T2207/10016
Inventor 邓练兵
Owner ZHUHAI DAHENGQIN TECH DEV CO LTD
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
Eureka Blog
Learn More
PatSnap group products