해상에서 소형 객체 인식을 위한 데이터 전처리 방안 연구

Abstract

Implementing good performance deep learning requires a large amount of high-quality data. However, in areas where the amount of data is limited, data collection takes a lot of time and cost. This study attempts to detect small object-sized submarine masts based on the environmental characteristics of the sea, which limit data collection and reduce the amount and quality of data due to the characteristics of submarine data used in the study. This study aims to improve recognition performance through preprocessing techniques with a small amount of data, and it can be seen that recognition performance has improved based on mAP.