Weston Seeding Stronger Communities

Seeding Food Innovation

Awarded Project 2018

Enabling the next revolution in global food production through automatically labelled data sets and machine learning

Project Description

We envision a future where it will be possible to lavish the same attention on individual plants in a large prairie crop farm as one might on those in a backyard garden. As camera sensors shrink in size, and self-driving vehicles continue to improve, such an idea is no longer the realm of science fiction. The remaining piece of the puzzle, however, is the need for a very large number of pre-identified images of crop plants and weeds with which to train a computer to recognize one from the other. Our research project is to develop the means to automatically generate and label such images and to make the resulting data sets openly available for use by Canadian researchers and companies. This approach puts the missing puzzle piece into the hands of more innovators more quickly, leading to new solutions for global food production sooner.

Relevance to the field of food innovation

By tackling the fundamental crux of this problem, our work will help enable the many new innovations in food production needed for a growing population in a world with diminishing resources. For example, autonomous vehicles fitted with imaging sensors and plant recognition software could one day operate around the clock, inspecting hundreds of hectares of farmland plant-by-plant, monitoring crop health, identifying weeds, pests and disease, and taking action to improve growing conditions and eliminate threats. This would greatly increase crop yields and help reduce herbicide and pesticide use.

Anticipated outcome

Our anticipated outcome is five-fold: (1) the development of methods to automatically generate and label images of crop plants and weeds; (2) the development of plant recognition software using these image datasets; (3) the open access sharing of our methods, datasets, and software; (4) the exploration of other sensor technologies; and (5) the testing of our software in the field using autonomous vehicles.

Grantees:

Dr. Christopher Bidinosti

Dr. Christopher Bidinosti

Dr. Bidinosti is an Associate Professor in the Department of Physics at the University of Winnipeg. He received his PhD in physics from the University of British Columbia, studying high-temperature superconductivity. As a postdoctoral fellow at L’École Normale Supéreiure (Paris) and Simon Fraser University, he began working on nuclear magnetic resonance (NMR) and magnetic resonance imaging (MRI). » More Info

Dr. Christopher Henry

Dr. Christopher Henry

Dr. Henry, P. Eng. (Applied Computer Science, University of Winnipeg) is an Associate Professor and obtained his Ph.D. from the Department of Electrical and Computer Engineering at the University of Manitoba. He has many years of experience working in reinforcement learning and with theoretical frameworks for modelling human perception. » More Info

Dr. Carl Gutwin

Dr. Carl Gutwin

Dr. Gutwin (Computer Science and Plant Phenotyping and Imaging Research Centre, University of Saskatchewan) is an expert in Human-Computer Interaction and Information Visualization. He is a Pillar Co-Lead (Computational Informatics of Crop Phenotype Data) in the $48M Plant Phenotyping and Imaging Research Centre (P2IRC), and the co-directory of the University of Saskatchewan » More Info

Dr. Ian Stavness

Dr. Ian Stavness

Dr. Stavness (Computer Science and Plant Phenotyping and Imaging Research Centre, University of Saskatchewan) leads the Data Analysis for Rapid Plant Phenotyping project of the Plant Phenotyping and Imaging Research Centre at the USask. His research group is at the forefront of applying deep learning methods to image-based plant phenotyping. » More Info

Dr. Rafael Otfinowski

Dr. Rafael Otfinowski

Dr. Otfinowski (Biology, University of Winnipeg) has extensive experience designing greenhouse, field, and natural experiments. His primary research focuses on understanding links between plants and soils to conserve, manage, and restore prairie ecosystems. He uses statistical models, multivariate and spatial analyses, including GIS, to predict how the structure, composition, and diversity of restored prairie communities » More Info

Dr. Ed Cloutis

Dr. Ed Cloutis

Dr. Cloutis is a Professor in the Department of Geography at the University of Winnipeg. He has over 25 years of expertise in the applications of optical spectroscopy to a wide range of applications, including agricultural and environmental monitoring, analysis of cultural artefacts, and planetary exploration. He has served as Director of the University of Winnipeg’s Institute of Urban Studies » More Info

GrantDr. Jonathan ZiprickeeName

Dr. Jonathan Ziprick

Dr. Ziprick (Applied Computer Education, Red River College) obtained his PhD in Physics from the University of Waterloo while residing at the prestigious Perimeter Institute for Theoretical Physics. He has more than 10 years experience doing research across both academia and industry in the varied fields of computer science, mechanical engineering, physics and mathematics. » More Info