Basics of automated plant identification  

ABSTRACT: Historically, image-based dichotomous plant identification keys precede text-based ones by nearly one hundred years. Having lain in conceptual torpor for over 300 years, the notion of image-based identification has experienced a revival as a result of the development of modern applications which depend upon recent technological advances in electronic hardware (e.g. image sensors, network bandwidth, computer storage capacity) and software (especially image recognition systems and efficient large file browsing). There are essentially two different approaches to automated image-based recognition of plant species:  Leafsnap and Pl@ntNet. A brief discussion of the two approaches is here presented. Regardless of the approach, for successful automated plant identification, there are several dataset requirements and these are laid out in the following paper.

For the full proceedings of this workshop please click here.