Contracts awarded under Innovative Solutions Canada

Wildfire predictions using satellite data and machine learning

Current wildfire prediction tools primarily use climate and meteorological data from Earth-based weather stations. This data can be sparse and costly to collect.

Coanda is using artificial intelligence and machine learning to access and draw useful information from a large ensemble of diverse satellite data in order to increase our capability to predict wildfire events autonomously. Apart from the direct benefits on public health and safety, this will translate into better use of resources and money, and reduction in losses due to wildfires.

Satellite mapping and monitoring of shoals through artificial intelligence

Having reliable data on the depth of the oceans along Canada's coasts is a priority for the Government of Canada. Because global warming has led to heavier maritime traffic in the Canadian Arctic, there is a growing need for accurate maps of this region. Unfortunately, for a number of Canada's coastlines, the data is out of date, or even non-existent.

Until now, ocean depth has been measured primarily through field surveys, using methods that are very costly and time-consuming. Effigis Geo-Solutions has come up with a solution that aims to map and monitor the depth of coastal waters using satellite images processed with the help of artificial intelligence. The system will be able to detect changes that seabeds may undergo, before those seabeds pose a threat.

Artificial intelligence tool to improve maritime risks and manage vessels

Maritime traffic is in constant progression with increases in ship numbers, size and routes, including increased traffic in the Arctic region. We need to develop solutions that will protect the environment, reduce ship emissions, enhance marine safety and security, and support trade and commerce.

While tools to support safe maritime traffic are already in use, Global Spatial Technology Solutions (GSTS) is proposing a next-generation global vessel management concept that enables dynamic vessel management to enhance maritime safety, security and operations. Their project consists in applying artificial intelligence and big-data methodologies to analyze satellite imagery, vessels' Automatic Identification System (AIS) combined with oceanographic, weather and environmental data sets. This next-generation capability will enhance OCIANA, the GSTS Maritime Risk and Vessel Management tool. Through automated analysis of vessel trajectories, the advanced OCIANA platform will detect maritime risk and threat situations early on to provide decision support information to the appropriate authorities and avoid disruptive events. By enhancing the risk management of global vessel traffic, this innovative tool will reduce the risk of collisions and groundings, minimize damage to fisheries and marine life, prevent environmental contamination, and support the operations of Canadian safety and security programs globally.

Development of a crop disease risk assessment Web tool for the Canadian Prairies

There are currently no tools available in Canada that integrate geospatial data in near real time to assess crop disease risk. Such risk factors, being highly variable in space and time, are difficult to monitor and measure effectively.

Scientists are interested in developing a crop disease risk assessment Web tool that would be applicable to all of the Canadian Prairies. While a first model has been developed for a small test site, with this project, H2O Geomatics will produce geospatial data for all the risk factors for the entire agricultural region of the Prairies. H2O Geomatics will use artificial intelligence to unlock the value of this big dataset to assess the potential of disease development.

This novel Web tool will go far beyond what is currently available, to allow farmers to efficiently access many different information layers as part of one single platform in near real time. This knowledge will help producers make the best decisions on how to mitigate environmental risks and adopt the most successful agricultural management practices.

Autonomous Tasking Optimization via Machine-learning (ATOM)

The number of satellite data vendors is continuously growing, and so are the demands for satellite imagery. The processes to purchase such data can be complex, tedious and time-consuming. As part of its mission to make satellite data accessible to the world, SkyWatch has developed the EarthCache platform, which provides a simplified digital interface for the distribution of Earth observation data.

The next step for SkyWatch is the automation and optimization of satellite tasking through machine learning and artificial intelligence. The enhanced platform will predict the optimal satellite to capture data therefore improving coordination and efficiency in task planning. By having access to such a simple and incredibly powerful platform to capture and share satellite data, the market will be in a better position to leverage satellites and use them in the most efficient manner possible.

Application of Novel AI Techniques to Satellite Big Data Analysis in Support of Maritime Risk Management

The project aims to develop a novel approach to classify vessels solely from time series of Satellite Automatic Identification System (SAIS) positional data (e.g., longitude, latitude, speed over ground, course over ground) by employing advanced data analysis and artificial intelligence (AI) models to reveal vessel trajectory idiosyncrasies. In particular, the approach exploits positional information to derive identification information entirely independent of the vessel's stated identifiers such as the Maritime Mobile Service Identity (MMSI). This approach aims to directly enhance the utility of the SAIS data stream by enabling sophisticated applications that depend on datasets with reliable vessel IDs.

Autonomous Tasking Optimization via Machine-learning (ATOM)

SkyWatch believes that Earth observation should be simple and accessible for everyone. They have developed the EarthCache platform for aggregating Earth observation data supply to customers and the TerraStream platform for simplifying the planning of satellite tasking and data management for satellite operators. This project involves incorporating the concepts developed in the first phase to operationalize improvements in the capabilities of both of these platforms. The improvements include operationalizing multi-mission space asset optimization; increasing filtering capabilities, SAR tasking and archive integration; improving machine learning models used to predict radio frequency interference, cost optimize tasking, and enhance the probability of collection; and building additional distribution controls for governments and large enterprises.