Turbid water quality imaging target detection system and method based on logical stochastic resonance

Abstract

The invention provides a logical random resonance-based turbid water imaging target detection system and method. According to the detection system and method, a low-quality image captured in turbid water is expanded in a one-dimensional manner according to a row direction; the signal-to-noise ratio of the low-quality image is estimated; one-dimensional signals obtained after the expansion are inputted into a logical random resonance unit; Gaussian white noises corresponding to the estimated signal-to-noise ratio are added into the logical random resonance unit; the system parameters of the logical random resonance unit are adaptively optimized by a genetic network unit; the output signals of the logical random resonance unit are inputted into a thresholder; and background pixels and target pixels in an original image are distinguished, imaging target detection can be realized. With the detection system and method of the invention adopted, target information in imaging can be detected out quickly and accurately, a vision-based underwater target detection distance can be effectively increased; and the system parameters of the logical random resonance unit can be adjusted adaptively. The method can be applied to the detection of different targets in different water.

Description

technical field [0001] The invention relates to the field of underwater computer vision, in particular to a system and method for detecting turbid water quality imaging targets based on logic stochastic resonance. Background technique [0002] Oceans, lakes and other waters contain abundant natural resources, which have very high research value. People have been trying to find various ways to understand and develop the underwater environment. However, due to the large absorption and scattering of light by water bodies, especially in turbid water bodies, the background noise is more abundant, making underwater imaging extremely blurred, which restricts the development of technologies such as target detection based on underwater vision. There are two main methods of traditional underwater imaging processing: image restoration and image enhancement. Image restoration requires accurate physical models, which are difficult to obtain and have large temporal and spatial variations...

AI Technical Summary

Problems solved by technology

[0003] This application provides a target detection system and method for turbid water quality imaging based on logical stochastic resonance, which overcomes the problem of strong noise interference in turbid water quality imaging, and can quickly and accurately detect target information in imaging without the need for previous image preprocessing , which effectively improves the detection distance of underwater targets based on vision, and because the system parameters of the logic stochastic resonance unit can be adaptively adjusted, this method can be applied to the detection of different targets with different water qualities

Images