Unmanned ship intelligent target detection method and system adopting binocular camera

A target detection and binocular camera technology, applied in file system, file system management, computer components, etc., can solve problems such as large weight file of SSD algorithm, complex lightweight network structure, and slow network operation speed, so as to facilitate training and upgrades, easy deployment and implementation, and fast detection of effects

Pending Publication Date: 2022-08-05
SHANGHAI JIAO TONG UNIV
View PDF1 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The single-stage YOLOv3 deep network and the two-stage Faster RCNN deep network have a complex network structure, a large number of deep network layers, long training time, and slow network operation speed, making it difficult to deploy on edge devices such as Raspberry Pi and JetsonNANO; and Xception is lightweight The network structure is relatively complex and is generally used in the field of object classification; while the weight file of the SSD algorithm is larger, the network is more complex, and it is difficult to deploy on edge devices such as Raspberry Pi and Jetson NANO to achieve quasi-real-time detection functions

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
  • Unmanned ship intelligent target detection method and system adopting binocular camera
  • Unmanned ship intelligent target detection method and system adopting binocular camera
  • Unmanned ship intelligent target detection method and system adopting binocular camera

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0063] According to an unmanned boat intelligent target detection method using a binocular camera provided by the present invention, the method includes the following steps:

[0064] Step S1: Create a system directory according to the Pascal VOC format.

[0065] Step S2: Download or take a selfie on the network to create the graphics of the obstacles, and process them into color pictures and place them in the JPEGImages directory; the obstacles are ships, buoys, reefs and other common obstacles; the color pictures are 416×416.

[0066] Step S3: Use BBOX or similar software to annotate the picture, and place the generated txt file and xml file in the labels directory and the Annotations directory of the system directory, respectively.

[0067] Step S4: randomly generate the training set train.txt file, the test set test.txt file and the validation set val.txt file according to the ratio of 7:2:1, and place them in the Main directory under ImageSets in the system directory.

[...

Embodiment 2

[0084] Embodiment 2 is a preferred example of Embodiment 1, in order to describe the present invention in more detail.

[0085] The present invention also provides an unmanned boat intelligent target detection system using a binocular camera, and the system includes the following modules:

[0086] Module M1: Create a system catalog in the Pascal VOC format.

[0087] Module M2: Download or take a selfie on the Internet to build the graphics of obstacles, and process them into color pictures, which are placed in the JPEGImages directory; obstacles are ships, buoys, reefs and other common obstacles; color pictures are 416×416.

[0088] Module M3: Use BBOX or similar software to annotate pictures, and place the generated txt file and xml file in the labels directory and the Annotations directory of the system directory, respectively.

[0089] Module M4: randomly generate training set train.txt file, test set test.txt file and validation set val.txt file according to the ratio of ...

Embodiment 3

[0106] Embodiment 3 is a preferred example of Embodiment 1, in order to describe the present invention in more detail.

[0107] Aiming at the shortcomings of the prior art, the present invention provides an intelligent target detection system for an unmanned boat using a binocular camera and a Parnet lightweight network, and improves the Parnet network originally used for the object classification function into a target detection network. It is implemented as an object detection network on the Darknet deep learning framework. After training, images captured by the binocular camera can be used to detect boats, buoys, reefs, and other common obstacles in real time, along with their relative positions and sizes.

[0108] The present invention is realized by the following technical solutions, and the present invention includes: the described unmanned boat intelligent target detection system using binocular camera and Parnet lightweight network, the core of which is binocular camer...

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 provides an unmanned ship intelligent target detection method and system adopting a binocular camera. The method comprises the following steps: establishing a system directory; downloading or performing selfie on a network to establish a graph of the obstacle, processing the graph into a color picture, and placing the color picture in a directory; marking a picture by using software, and respectively placing the generated txt file and xml file in a system directory; randomly generating a training set, a test set and a verification set according to a proportion of 7: 2: 1, and placing the sets in a system directory; a voc.names file is newly built in a system directory, the content is the name of an obstacle, and one file is built in each row; a voc.data file is newly built in a system directory, and a parnet.cfg file is newly built in the system directory and used for defining a lightweight network structure. The target detection network realized by the lightweight network is relatively shallow, simple in structure and easy to realize; the target detection network realized by the lightweight network is symmetrical in structure, few in parameters and convenient to train and upgrade.

Description

technical field [0001] The invention relates to the technical field of intelligent control of unmanned boats, in particular to a method and system for detecting an unmanned boat intelligent target using a binocular camera, and in particular to an unmanned boat intelligent target using a binocular camera and a Parnet lightweight network Detection Systems. Background technique [0002] Unmanned boats are mobile platforms that can perform water quality testing, topographic mapping, fish censuses, security patrols and other related operations in the fields of oceans, lakes and rivers, and have broad application prospects. Especially with the establishment of my country's maritime strategy and the emphasis on maritime rights, small unmanned boats with path planning and intelligent obstacle detection capabilities have become an important research object and platform in the field of intelligent control at home and abroad. In order to improve the reliability of its movement, the in...

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 Applications(China)
IPC IPC(8): G06V10/764G06N3/04G06V10/774G06V10/82G06V10/22G06T7/70G06F16/11G06F8/61G06F8/41
CPCG06V10/764G06V10/774G06V10/82G06F8/61G06F8/41G06F16/11G06T7/70G06V10/22G06V2201/07G06T2207/20081G06T2207/20084G06T2207/10024G06N3/045
Inventor 宋立博费燕琼
Owner SHANGHAI JIAO TONG 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