Target detection method based on sonar image

Abstract

The invention discloses a target detection method based on a sonar image. The method comprises the following steps: obtaining sonar images in real time, taking at least four continuous sonar images, and carrying out image addition operation on the at least four continuous sonar images; de-noising the sonar images by adopting a median filtering algorithm for the images after the addition operation, and carrying out binarization processing by taking a pixel mean value of the sonar images as a threshold value; performing density clustering on binarized images to obtain a plurality of connected regions; calculating the area of the connected regions, and screening the connected regions by setting a threshold value to select suspected target regions; and solving directions and length-width ratios of bounding rectangles of the suspected target regions, and obtaining a final target object region by setting a threshold value. The method can guide a vehicle to sense a surrounding environment through sonar images in an underwater navigation process, is suitable for sonar image processing under the condition of more underwater noise, and solves the defect of likely missed detection when the performance of a processor is insufficient. Therefore, the vehicle can carry out effective underwater obstacle avoidance navigation, and a target detection function can be realized.

Description

technical field [0001] The invention relates to a target detection method based on sonar images, which is mainly applied to obstacle avoidance navigation and target detection of underwater unmanned vehicles. Background technique [0002] Underwater operating vehicles need to sense the surrounding environment when performing underwater operating tasks. Multi-beam forward-looking sonar can detect the front fan-shaped area, but due to the influence of water flow disturbance, foreign object interference and other factors, especially in the working conditions near the bottom, near the water surface and near the shore, the reverberation and side lobe effects are serious, resulting in sonar The image noise is serious, and more accurate image perception data cannot be obtained directly, so it is easy to misjudgment only by processing a single image, and processing each image will easily lead to increased pressure on the processor. Contents of the invention [0003] In order to im...

AI Technical Summary

Problems solved by technology

Multi-beam forward-looking sonar can detect the front fan-shaped area, but due to the influence of water flow disturbance, foreign object interference and other factors, especially in the working conditions near the bottom, near the water surface and near the shore, the reverberation and side lobe effects are serious, resulting in sonar The image noise is serious, and it is impossible to directly obtain more accurate image perception data, so it is easy to misjudge only a single image, and each image is processed, which will easily lead to increased pressure on the processor

Images