Fast and Robust Underwater Obstacle Detection in Acoustic Vision

Sonar is a vital instrument for underwater detection. BlueView(BV) as an acoustic vision sensor is subject to structural noise during detection. This phenomenon has a negative impact on the detection and image restoration of sonar. In this paper, we propose an improved Bayes estimation method for ac...

Full description

Saved in:
Bibliographic Details
Main Authors: Liu, Hengrui, Chen, Deshan, Zhang, Di, Zhou, Peng
Format: Conference Proceeding
Language:English
Subjects:
Acoustics
Estimation
Image edge detection
information fusion
Instruments
Object detection
robust detection
Sonar
structural noise
underwater obstacles
Vision sensors
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Sonar is a vital instrument for underwater detection. BlueView(BV) as an acoustic vision sensor is subject to structural noise during detection. This phenomenon has a negative impact on the detection and image restoration of sonar. In this paper, we propose an improved Bayes estimation method for acoustic vision object detection in the BV forward looking sonar images. The method is based on Gaussian mixture model and to represent background prior probability model for each pixel. The conditional probability model construction via an improved sigmoid model. As verified experimentally, the proposed method shows a fast yet robust result.
ISSN:2325-0925
DOI:10.1109/ICSSE50014.2020.9219258