Oil spill on the water bodies has adverse effects on coastal and marine ecology. Oil spill contingency planning is of utmost importance in order to plan for mitigation and remediation of the oceanic oil spill. Remote sensing technologies are used for monitoring the oil spills on the ocean and coastal region. Airborne and satellite sensors such as optical, infrared, ultraviolet, radar and microwave sensors are available for remote surveillance of the ocean. Synthetic Aperture Radar (SAR) is used most extensively for oil-spill monitoring because of its capability to operate during day/night and cloud-cover condition. This study detects the possible oil spill regions on fully polarimetric Uninhabited Aerial Vehicle - Synthetic Aperture Radar (UAVSAR) images. The UAVSAR image is decomposed using Cloude-Pottier polarimetric decomposition technique to obtain entropy and alpha parameters. In addition, other polarimetric features such as co-polar correlation and degree of polarization are obtained for the UAVSAR images. These features are used to with fuzzy logic based classification to detect oil spill on the SAR images. The experimental results show the effectiveness of the proposed method.
Measuring water quality of bays, estuaries, and gulfs is a complicated and time-consuming process. YSI Sonde is an instrument used to measure water quality parameters such as pH, temperature, salinity, and dissolved oxygen. This instrument is taken to water bodies in a boat trip and researchers note down different parameters displayed by the instrument’s display monitor. In this project, a mobile application is developed for Android platform that allows a user to take a picture of the YSI Sonde monitor, extract text from the image and store it in a file on the phone. The image captured by the application is first processed to remove perspective distortion. Probabilistic Hough line transform is used to identify lines in the image and the corner of the image is then obtained by determining the intersection of the detected horizontal and vertical lines. The image is warped using the perspective transformation matrix, obtained from the corner points of the source image and the destination image, hence, removing the perspective distortion. Mathematical morphology operation, black-hat is used to correct the shading of the image. The image is binarized using Otsu’s binarization technique and is then passed to the Optical Character Recognition (OCR) software for character recognition. The extracted information is stored in a file on the phone and can be retrieved later for analysis. The algorithm was tested on 60 different images of YSI Sonde with different perspective features and shading. Experimental results, in comparison to ground-truth results, demonstrate the effectiveness of the proposed method.