We recently highlighted a new application called PlantNet that aims to identify plants through a picture and we were genuinely shocked by just now popular the post went. It seems millions of people out there have a need for an app that can identify pics of flowers, plants, herbs etc.
PlantNet works by lots of people contributing to an ongoing flower database to build a better system of identification but because of this style and because of other teething problems, many people are claiming that it’s not identifying some fairly simple plants correctly.
The app we highlighted is still in development and should improve as more and more people contribute. However, if you’re serious of identifying plants and flowers, you know how important it is that these apps get the job done right.
There are actually a couple of alternative apps that work by employing the work of trained botanists rather than automatic recognition. The makers of these apps claim their systems are much more accurate and, although it may take longer to get an answer, the results are provided by experts.
PlantSnapp - an app started by an amateur botanist from Leeds University. It works to have plants and flowers identified by a team of experts, they have a high rating in the app store and lots of positive reviews saying the service works.
FlowerChecker - which was started by three doctoral students in the Czech Republic. They originally built the app with the aim of collecting enough data on plants so they could build an automatic system of identification. The idea for them was that they would identify the plants people sent them for free because their payment was the data in the picture, however, so many people started to ask for their services that they started to charge a fee.
Both the apps above charge for identifying plants, users buy credits for the app which are then used to pay for the advice of experts. The models for both apps are the same but for us Plant Snapp seems to be a little ahead in userbility.
All these companies seem to be working towards the end goal of automatic recognition but this requires a huge database of pictures before it can work correctly. If you want plant identification that works well it seems human experts are still needed.