Volume 18, No. 6, 2021
A Study Of Image Identification For Underwater Images Using ANN Technique
Dr. A. Divya , Dr. M. Sathya
Abstract
Research on a Indian oceanographic research vessel, Polaris includes developing holographic image software for the ship. With a weight of 260 tonnes, a length of 36.98 metres, and a breadth of 6.80 metres, this ship can reach 11 knots of speed. For this reason, it has performed a number of underwater rescue and exploratory operations. A wide range of aspects, includes water temperature, water deflection, and phantom circumstances, as well as the weather, oceanic current, algae, and light replication from schools of fish, could obstruct underwater research efforts. For marine exploration, immediate picture analysis is essential. Scientists built Artificial Neural Network(ANN)-based credit software to quickly identify an underwater object's category to meet Polaris's need for immediate identification. A neural network was used to improve shape recognition in low-resolution underwater photos, resulting in a 95% recognition rate on average. Using a neural network instead of manual recognition proved to be more accurate.
Pages: 2599-2614
Keywords: Image processing, Neural network, moment invariants, underwater images, image recognition.