분류학습모델과 CNN 모델을 이용한 해상물체 식별 성능 평가 방법 연구

Abstract

We conducted a study to identify small objects floating at sea using machine learning technology based on hyperspectral image data. Classification learning models and CNN models were used to identify maritime objects, and the accuracy of identification of maritime objects was quantitatively analyzed. In addition, the performance between each model was compared by calculating detailed identification accuracy according to the classified objects.