Sonar target detection method and system based on time history accumulated images

Abstract

The invention discloses a sonar target detection method and system based on time history accumulated images, and the method comprises the following steps: 1, carrying out the data interpolation, coordinate transformation and back projection of original sonar data, and forming a polar coordinate PPI sonar image; step 2, accumulating each wave beam in the original sonar data in the time direction to form a time history image of a fixed period; step 3, carrying out frame separation processing; step 4, designing an image enhancement algorithm to perform line feature enhancement on the global and local time history images after the frame separation processing; 5, performing linear target detection on the enhanced process image by using an improved multi-scale LSD algorithm; step 6, performing post-processing on a linear target detection result; and 7, performing moving target detection on the PPI sonar image sequence by using a DBT technology, and performing data fusion with a linear detection result to obtain a final detection result. The method has the advantages of enhancing the signal-to-noise ratio of the small target, being high in background interference resistance and the like, and can realize accurate detection of the moving target in the complex underwater environment.

Description

technical field [0001] The invention relates to a sonar image target detection technology, in particular to a sonar target detection method and system based on time-history cumulative images, which can be applied to the detection, observation and tracking of underwater security fields and other underwater targets, belonging to image processing and water under the field of target detection technology. Background technique [0002] In recent years, with the rapid development of technologies such as artificial intelligence and smart oceans, unattended defense systems such as underwater security have gradually become research hotspots and difficulties. Among them, the detection, tracking and observation of underwater targets such as underwater frogmen and UUVs The most critical technology, sonar is the technical equipment that "does my part" to achieve this purpose, and sonar target detection is the technical basis. However, due to the complex and changeable underwater environm...

AI Technical Summary

Problems solved by technology

Although these two types of methods have achieved automatic detection, it is still difficult to achieve a win-win situation of accuracy and real-time performance. Their limitations are mainly as follows:

[0004] (1) Low target signal-to-noise ratio: because the sonar working environment is subject to passive and active interference such as ocean noise, reverberation, and ship noise, and the interference is often irregular can be followed, causing the target to be easily confused with the noise

(2) High false alarm rate: The threshold detection of the DBT method mainly relies on methods such as inter-frame difference of image sequences and background modeling, but the threshold needs to be modified according to different environments. If the threshold is too high, the target with low target intensity (such as frog People, etc.) are prone to missed detection, and if the threshold is too low, there will be too many false detections

Constant false alarm (CFAR) technology, under certain requirements of the system for the false alarm rate caused by noise and other interference, performs adaptive threshold detection on the target. Although it solves the above limitations, it only uses single-frame image information. poor performance in the environment

However, the TBD method also has large errors in data association due to noise and other problems, resulting in high follow-up false alarms.

(3) Lack of information in sonar images: sonar images only contain the intensity and motion information of the target, and cannot obtain prior knowledge such as the color and shape of the target like optical images, which greatly increases the difficulty in the design of detection algorithms

(4) Data scarcity: Due to the limitations of information confidentiality and high data acquisition costs, it is difficult to obtain sonar image data including frogmen, UUV and other targets, making it difficult for the popular deep learning target detection technology to exert its advantages in sonar image processing

Images