Grant: Discovery Grants Program - Individual

Remote sensing, the science of aerial or satellite data capture of the earth, provides crucial information across land, oceans, and atmosphere. As a growing part of the world economy, the remote sensing market is expected to reach nearly $30B by 2030. Given the vast and increasing amount of remote sensing data being produced, the global challenge, and the basis of my research program, is creating groundbreaking intelligent algorithms which are urgently needed to increase throughput, reduce turnaround time, save money, and provide information for decision making while scientific experts must be trained to manage projects and lead future missions. My research focuses on innovative design and operational implementation of computer vision algorithms to monitor earth's oceans. Key images are produced by a set of three Canadian satellites called the RADARSAT Constellation Mission (RCM) used for monitoring of ice-infested waters in and around Canada. The Canadian Ice Service (CIS) is the government agency tasked with producing required ice maps using RCM imagery. CIS personnel manually interpret 10,000 images annually, providing ice maps over huge (350km x 350km) regions to infer ice thickness, extent, and strength to facilitate decision support systems for shipping navigation, ice breaking, climate modelling, and on-ice northern transportation. My team conducts novel research in machine learning to automatically interpret this imagery accurately, solutions that provide far more detail than a human operator. We aim to develop a viable new paradigm for operational ice classification, devising algorithms that consistently determine the ice type of a particular pixel. Knowing ice types and extents is crucial for navigation and ship captains do not enter ice-infested waters without this information - planned algorithms can deliver this to them. We are developing algorithms that can detect oil spills and polynyas (open water surrounded by ice) in SAR imagery. Some ships illegally dump bilge oil, and this leaves an oil trail in oceans that must be identified by satellite imagery. New algorithms will be created to detect such oil spills so that perpetrators can be identified. Polynyas, important for on-ice navigation around northern communities, are now more unpredictable due to global warming, so we will create new methods to identify these small regions. We plan to build dependable algorithms that count whales in aerial imagery, important to Canada for the environment, tourism, and sustenance. Other northern nations would be interested in utilizing maps derived from my team's efforts. My NSERC-enabled research in computer vision required to solve these problems will continue to be successfully applied to other domains such as video sports analytics and image understanding while simultaneously training HQP to tackle intelligent systems problems and gain leadership roles in industry.