Study On Stereo Monitoring Sea Ice Image Processing And Analysis

Sea ice brings enormous challenges to offshore structures and ocean development in the seasonally glacial sea areas. Monitoring the parameter information of sea ice timely and accurately is not only an important part of ocean management, but also a significant way to ensure ocean exploitation security. In the current stage of offshore engineering, satellite remote sensing, radar detection, video monitoring and other technologies have been widely used to provide abundant original image resources for sea ice monitoring activities. This paper creates a compositive sea ice image processing and analysis software system based on stereo monitoring to extract parameter information of sea ice from satellite images, radar images and videos by integrating and improving the relative computer image processing and analysis algorithms.The system is composed by three submodules which handle different original images. It can extract concentration, size, thickness and velocity of sea ice from each kind of images by using the algorithms like binarization algorithm, image erosion and dilation, median filter, histogram equalization, linear gray level transformation, boundary tracing, maximum connected region algorithm in eight directions, features matching and others.In satellite image submodule, the complex contents of satellite image can be classified into land, sea and ice using calibration board voting algorithm according to the color. The coordinates of coastal line and boundaries can be extracted from satellite images directly with boundary tracing algorithm. In radar image submodule, multi-scale thresholds based on rectangular segmentation are used to solve the problem caused by different strength of reflected radar signal after calibrating the region of interest. The floating velocity field is obtained by using block image matching and sea ice thickness is estimated by using histogram equalization which strengthens detail information of radar images. In video image submodule, functions as auto segmentation of non-ice region, I-shaped caliper and key features matching are used to improve the accuracy and efficiency in the calculation.The system is applied for sea ice monitoring on the JZ20-2oil and gas field exploitation platform in Bohai Sea and plenty of sea ice conditions and movement information are obtained from original images during the field trials. The result of cross-comparison will be followed by a multiscale data correction model. In the end, studies have completely to establish a data fusion module for sea ice parameter information of stereo monitoring. It proofs that the image processing and analysis algorithms mentioned in this paper can extract sea ice parameter information accurately and efficiently as well. As the specialized business support software, this compositive system can make great contributions to sea ice monitoring and forecasting.