The invention relates to the technical field of low-frequency
sonar systems, in particular to an active
azimuth history graph suspected target prompting method based on
image segmentation, which comprises the following steps: firstly, carrying out binarization
processing on an active
azimuth history graph through a maximum between-class
variance method, then, carrying out
morphological processing and connected
region analysis on a
binary image to realize 8 connected region marking, and finally, carrying out
image segmentation on the image; secondly, calculating the area and the
mass center of each connected region, filtering the region with the large area to obtain a discrete target region
mask image, multiplying the discrete target region
mask image with an original
binary image to obtain a discrete target region image, and finally, carrying out BFS search on the
mass center of each discrete target region to obtain a discrete target region image; and calculating the sum of the Euclidean distances between the active
azimuth history graph and five nearest connected regions, marking the
mass centers of five discrete target regions with the maximum
Euclidean distance sum on the active azimuth history graph, and realizing isolated
bright spot positioning, thereby realizing automatic prompting of suspected targets in the active azimuth history graph of the
sonar system so as to reduce the difficulty of target identification of a worker.