Survey of the application of machine learning to automated greenhouses

Abstract: This survey examines previous reports and experiments to look at the possibilities of creating an automated greenhouse, as well as the possibilities an automated greenhouse could provide: such as a much more efficient water consumption, utilizing space more efficiently and providing food to a bigger population without the downsides of more transportation.This report also researches what methods have been tested and used previously: A lot of the previous methods have been to solve problems within the topic of plant phenotyping rather than pursuing an actual automated greenhouse, by object detection and by that counting leaves. These methods have been a big inspiration for the report, but had little relevance to the conclusion. It is a report where I have tried to create a plan on how to start, how to think during the start and how to proceed a project in this context. The conclusion I have reached is that the tech previous to this report was not ready to complete this task, but it has advanced so that it is now possible and some have been successful in this context during the time this report was written. I also came to the conclusion that cameras and object detection might not be essential to complete an automated greenhouse, but rather an optional accessory.