Artificial intelligence monitors and predicts quality of roses13 October 2016
IQ-Flora is an innovative system that monitors the product quality of roses in the chain in 'real time'. The first module of the prototype has just been completed, the other components will be tested in November 2016.
In the 'Quality-controlled rose chain' project, Royal FloraHolland is working together with Wageningen University & Research to develop IQ-Flora, a system that uses artificial intelligence and sensors to monitor and predict the product quality of roses. IQ-Flora aims to allow us to control the entire post-harvest phase - from producer to retailer - continuously on the basis of expected vase life, flower opening and possible ethylene damage. This will ensure that consumers receive a rose of higher quality while the chain partners will experience less waste and lower logistical costs.
Prototype predicts with software, sensors and app
The project team is working to develop a prototype. The core of this prototype is a software module for quality prediction, a method borrowed from the field of artificial intelligence. In the module, experts from Royal FloraHolland and the University of Wageningen have charted the various influential factors on the quality of flowers and their expected impact.
These predictions are supplemented with data from sensors and additional quality information via a smartphone app. Based on these operational data, the calculation rules from the model are increasingly refined, and thus the reliability of the prediction improves. The current quality estimate is also calculated and then reported (together with other quality information) via a web application.
Start of test phase
The module for quality predictions has been completed and was recently presented at the Greentech trade fair. The other modules of the prototype are still being developed and will be tested in November by Van Dijk Flora and Roseconnect.
Subscribe to our newsletter and receive each week:
- current and relevant news
- based on your own profile
Do not display againSubscribe