VIIRS Day/Night Band Image Processing System for Boat Detection in Java Sea, Indonesia

VIIRS is a sensor onboard NOAA JPSS satellite series. Day/Night Band (DNB) is one of VIIRS outputs that offers overnight monitoring of quantitative low visible and near-infrared light emitted by the earth’s surface. VIIRS DNB composites can generally be used for visual interpretation. In order to get better interpretation and precise result, this paper provides a development of VIIRS DNB image processing system for boat detection in Java Sea, Indonesia. Boat detection is chosen as the preliminary target of this image processing since the light presence in the ocean is brighter than the background. An image processing system was developed to process the raw data of VIIRS DNB in SDR (Sensor Data Record) and EDR (Environmental Data Record) level. Pre-processing and mask-processing are two main processes implemented in this system. Radiance value of VIIRS DNB is applied to mathematical operations as well as wiener and median filters during the preprocessing step. In mask-processing, the system focuses on categorizing and eliminating other lights such as lightning, gas flare, and ionospheric energetic particle using some thresholds based on spike height index (SHI). During the implementation, by taking a sample area in Western Java Sea, an average of 454 boats is detected at night. The image processing time takes less than 4 minutes for one input data. The output data come in CSV format including longitude and latitude of boat, which can be used to monitor marine activities at nighttime and help stakeholders to counter illegal fishing in Indonesia.