A Review: Object Detection and Classification Using Side Scan Sonar Images via Deep Learning Techniques

Deep learning technique which is otherwise called either deep machine learning or deep construction based learning approaches, have newly attained remarkable accomplishment in processing images digitally for object recognition and also categorization. Accordingly, object detection and classification are quickly reached the esteem and attention from computer vision investigation society. Also, this enormous growth in imaging data has drive necessitate for detection and classification automatically through deep learning NN based classifiers. In this paper, reviewed several existing research work from the year 2015 to 2021 regarding detection of objects in underwater acoustics on side scan sonar images using various techniques such as deep based CNN, machine learning and image processing used by various researchers. Also, how the detected objects in seafloor were categorized into mine, rocks, mud and other non-mine objects were analyzed. Moreover, several CNN architectures were established for objects recognition, segmentation in underwater sea by various investigators. And finally what kind of datasets utilized by various authors found in their research works along with published year, techniques such as ML, DL, the objective of each existing work, detection accuracy, classification accuracy, and what kind of outcomes every author determined were summarized.