OPR OPTICAL PLANT RECOGNITION
OPR is a smart system based on image recognition, fuzzy logic and Bayesian statistics that aims to help people with no specific botanical knowledge to identify the plants species in an urban park or in a natural open space.
The image recognition system analyses a picture of a tree leaf to retrieve the main morphological characteristics that will help identify it. The user can also provide additional information about the plant through a set of simple questions: has it flowers or not, is it a tree or a bush, and the like. For more advanced users, the system includes a complete form where they can fill all the relevant aspects of the plant.
Usually, this kind of field guide is based on a dichotomous key, which is quick, but not very robust. If the user makes a mistake or lacks some key information the result will be wrong. To avoid this issue, as OPR main target are non-plant-savvy people, the classification system uses fuzzy logic and Bayesian statistics. The result is a list of possible matches with a certain degree of probability. An arbitrary mistake would lower the right match probability, but would not turn it to 0.
To make the process more accurate and fast, the OPR system also takes into account other information that may reduce the number of possible candidates, like location or date, and the distribution of the different species throughout the area.
Goals:
- Development of citizens’ environmental awareness through the knowledge about the vegetal species around them.
- Increase value of urban and natural parks integrating them in the touristic offer of the city and attracting economic activity to the area.