Machine vision and audition integrating detection and identification method for underwater moving objects
A moving object and machine vision technology, applied in character and pattern recognition, instruments, computer parts, etc., can solve problems such as turbidity, limited light propagation distance, and lack of corresponding improvement in target detection and recognition capabilities
Active Publication Date: 2018-09-04
TIANJIN UNIV
View PDF11 Cites 15 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Sonar imaging is a wide and thick technology, its detection distance is wide and deep, but due to the reflection and refraction of sound waves caused by the undercurrent of the seabed, the unevenness of the water body, the scattering caused by the unevenness of the seabed, the detection of various underwater organisms Interference makes sonar imaging have very strong speckle noise, and sonar images have no color and cannot provide texture and shape information, causing great trouble to underwater object tracking and recognition
Recently, monocular or binocular vision technology based on optical imaging has begun to be applied to AUVs. Although optical images can provide color, texture and shape information, due to the very limited propagation distance of light in water, this technology is only suitable for Shallow sea areas, and because the water quality in shallow sea areas is relatively turbid, the ability of target detection and recognition has not been improved accordingly
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0026] The Bayesian framework is a relatively traditional effective method for target tracking and information fusion. Recently, due to the development of neural networks, especially the development of deep learning, Bayesian networks have been further developed. This research aims to use the development of a block sparse dynamic Bayesian network so that the two kinds of information can be better fused together to achieve fast and accurate tracking and identification of underwater moving targets.
[0027] The main technical route is divided into the following parts:
[0028] The first step: the construction of the acoustic vision system and optical vision system.
[0029] The optical vision system includes underwater cameras, lighting and other equipment to meet the basic requirements for obtaining optical images and video information. The task to be completed by the acoustic vision system is the automatic positioning of the target. At present, the multi-beam sonar system is ...
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
Login to View More Abstract
The invention relates to the field of the detection and identification for the underwater moving objects, and aims to make up detects of an underwater optics method and an underwater acoustic method when being separately applied to underwater moving target detection and identification, and the information can be timely and accurately intercommunicated, the tracking and identification for the underwater targets can be timely and accurately performed. The invention provides a machine vision and audition integrating detection and identification method for underwater moving objects. The method includes the following steps: the first step: the establishment of a sound vision system and light vision system; the second step: the pretreatment based on the sound vision system; the third step: the target detection and feature extraction based on the light vision system; the fourth step: the integration of the sound vision system and light vision system based on a Bayes network. The method is mainly applied to the detection and identification for the underwater moving objects.
Description
technical field [0001] The present invention relates to the field of underwater moving object detection and recognition (underwater moving object detection and recognition), in particular to the detection and recognition of underwater moving objects using machine vision and auditory fusion, specifically, to the use of sparse dynamic Bayesian networks to detect and recognize underwater moving objects. Combining the underwater vision and underwater acoustic detection methods to classify the visual images and acoustic signals of different moving objects to realize the automatic recognition of underwater moving objects. Background technique [0002] Tracking and recognition of underwater moving objects has always been a common problem in marine science and computer science and technology. Autonomous Underwater Vehicle (AUV) is an important carrier for understanding the ocean. AUV plays a very important role in the collection of marine data, exploration of seabed resources, obse...
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
Login to View More Patent Type & AuthorityApplications(China)
IPC IPC(8): G06T7/269G06T7/215G06K9/62
CPCG06T7/215G06T7/269G06F18/24323G06F18/254
Inventor张翠翠刘志磊
OwnerTIANJIN UNIV