Optimized vessel detection in marine environment using hybrid adaptive cuckoo search algorithm

摘要

Detection of vessels in a marine environment is a challenging task due to the complexity of identifying small objects, necessitating a detection algorithm to discriminate between variants in vessel-based geometry. False detection can be an issue due to sea clutter. The huge neural network-based computation models used for data analysis produce results that reflect the changing environment, based on iterations and common computation assumptions. To obtain the best results, the proposed research uses an optimization model with a classifier. This experimental model provides better accuracy than other detection systems, even with thousands of vessel images. This hybrid, adaptive cuckoo search-based optimization model produces the best results in dynamic sea-clutter regions, and outputs show lower false alarms in ports and other coastal surveillance regions. (C) 2019 Elsevier Ltd. All rights reserved.