Overview
What is Nachet?
The CFIA ensures that seeds or grains crossing Canadian borders meet strict phytosanitary standards to prevent the spread of invasive weed species. Nachet Interactive is an open-source web application developed to assist CFIA seed analysts in identifying regulated weed seeds quickly and accurately. It combines digital microscopy with computer vision models hosted on AzureML to streamline the seed inspection process.
Initially focused on weed seed identification, Nachet Interactive's scope is expanding to include other tasks like identifying pure seeds and insects, showcasing the application's versatility. This tool aims to support CFIA scientists by integrating AI into their workflow.
Seed Species
Currently we are training our models to identify these 15 species:
| Species |
|---|
| Ambrosia artemisiifolia |
| Ambrosia trifida |
| Ambrosia psilostachya |
| Brassica junsea |
| Brassica napus |
| Bromus hordeaceus |
| Bromus japonicus |
| Bromus secalinus |
| Carduus nutans |
| Cirsium arvense |
| Cirsium vulgare |
| Lolium temulentum |
| Solanum carolinense |
| Solanum nigrum |
| Solanum rostratum |
For more information about the CFIA's Seed Program, visit the Seeds identification page.
Web App
Visit this link to access the deployed Nachet Interactive web app: https://nachet.inspection.alpha.canada.ca/
Application Components
Nachet Interactive is structured around two key elements: a React-based frontend and a Flask-powered backend. For more information about each element, view the respective documentation below: