This study proposes an approaching method of identifying sea fog by using Geostationary Ocean Color Imager (GOCI) data through applying a Convolution Neural Network Transfer Learning (CNN-TL) model. In this study, VGG19 and ResNet50, pre-trained CNN models, are used for their high identification performance. The training and testing datasets were extracted from GOCI images for the area of coastal regions of the Korean Peninsula for six days in March 2015. With varying band combinations and changing whether Transfer Learning (TL) is applied, identification experiments were executed. TL enhanced the performance of the two models. Training data of CNN-TL showed up to 96.3% accuracy in matching, both with VGG19 and ResNet50, identically. Thus, it is revealed that CNN-TL is effective for the detection of sea fog from GOCI imagery.

Ship type classification of synthetic aperture radar imagery with convolution neural network (CNN) has been faced with insufficient labeled datasets, unoptimized and noised polarization images that can deteriorate a classification performance. Meanwhile, numerous labeled text information for ships, such as length and breadth, can be easily obtained from various sources and can be utilized in a classification with k-nearest neighbor (KNN). This study proposes a method to improve the efficiency of ship type classification from Sentinel-1 dual-polarization data with 10 m pixel spacing using both CNN and KNN models. In the first stage, Sentinel-1 intensity images centered on ship positions were used in a rectangular shape to apply an image processing procedure such as head-up, padding and image augmentation. The process increased the accuracy by 33.0% and 31.7% for VH (vertical transmit and horizontal receive) and VV (vertical transmit and vertical receive) polarization compared to the CNN-based classification with original ship images, respectively. In the second step, a combined method of CNN and KNN was compared with a CNN-alone case. The f1-score of CNN alone was up to 85.0%, whereas the combination method showed up to 94.3%, which was a 9.3% increase. In the future, more details on an optimization method will be investigated through field experiments of ship classification.

Illegal fishing has been a serious threat to the conservation of seafood resources and provoked the importance of marine surveillance. There are several types of fishing vessel monitoring systems operated by Republic of Korea, for example, Vessel Monitoring System(VMS), Automatic Identification System (AIS), V-Pass and VHF-DSC. However, those methods are not adaptable directly to fishing activity monitoring. The limitation requires more human resources to determine fishing status. Thus, this study proposes a method of estimating fishing activity from V-Pass, fishing vessel position reporting system, using Hidden Markov Model (HMM). HMM is a model to determine status through probability distribution for a sequence of time-series data. First of all, fishing activity status was labeled on V-Pass data. The distribution of speed on fishing activity was computed from the labeled data and HMM was constructed from the data obtained at Socheongcho Ocean Research Station (SORS). The model was first applied to the data of SORS for a test, and then Busan for validation. The model showed 99.4% and 89.6% as test and validation accuracy, respectively. It is concluded that the HMM can be applicable to predict a fishing activity from vessel tracks.