In a few months , it will be clear which squad wins the Autonomous Greenhouse Challenge and thus has modernise the respectable AI algorithm for an autonomously manipulate cultivation of midget tomatoes . In analog with the participants , a team of WUR experts also dress up a nursery for autonomously controlled cultivation . investigator Pinglin Zhang explains the usefulness of this reference greenhouse and why the additional sensors and tv camera deploy by the teams can also supply worthful noesis for WUR .
In Bleiswijk , nanus tomatoes are presently growing in five glasshouse compartment controlled by AI algorithmic program germinate by the team who participate in the Autonomous Greenhouse Challenge . A sixth compartment is also growing midget tomatoes but is see with input from a squad of WUR expert . One of these is investigator Pinglin Zhang . ' Our compartment we also call the denotation greenhouse . Cultivation here is also autonomously controlled but without AI algorithms . We mark up the greenhouse as a regular agriculturalist would control the greenhouse . With this reference , we can make a dear comparison between the result of greenhouses hold in with AI algorithms and a greenhouse controlled in a commonly used autonomous way . '
Additional sensing element and camerasAnother difference is that , unlike the WUR team , the participants have been allowed to use additional sensors and cameras since this yr , says Zhang . ' In all six compartments - so admit ours - there are basic sensor , for example for measuring temperature and relative humidness inside the greenhouse and real - clip weather conditions outside . On top of this , some team have installed their own sensor and television camera . Think of a detector that measures the weight of a pot . Or a thermal photographic camera measuring the temperature around leaves and yield . So the teams have more comment they can use in control than we have in our reference greenhouse . '
data point processed by algorithmThe teams have developed their algorithmic program so that the datum collect by the sensing element and camera is straightaway march to make an optimum decision . Zhang : ' A low weight can indicate deficient water , so the plant can be given surplus water . you may also use data on weight to determine when to glean . Just like the redness of the yield that cameras record . Based on the teams ' strategy , the algorithm cipher the harvest time particular date . Whether this also ultimately lead to the most rich crop with the best caliber fruit will be hear at the end of the challenge . '
From greenhouse to implant levelAs a researcher , Zhang deals with nursery engineering science for growing crops more efficiently and sustainably . ' datum from sensing element can help growers optimise Energy Department consumption for light , passion , and breathing . Besides at the greenhouse level , I am also calculate at sensors on a smaller scale . For representative , in a project I am working on appraise the microclimate : the temperature and humidity around a plant . By placing detector penny-pinching to the plant , you collect data point on local conditions . This can give an indication of , for example , the comportment of bacterium that can cause diseases . If you have that in view , you may intervene early . '
More focal point on the plantIn the Autonomous Greenhouse Challenge , Zhang also look an increase focal point on monitor the plant with smart sensors and cameras . ' In the past , the focal point was almost exclusively on greenhouse climate . That provides valuable data for controlling the greenhouse but gives less info on how the plant itself develop . While that is ultimately the most of import thing for skillful product . Some team in the Challenge really knead with State Department - of - the - art engineering . It is also very interesting and worthful for us as WUR to see the latest developments in sensors and camera . '
future tense of autonomous growingFinally , how does Zhang conceive of the time to come of sovereign ontogeny ? ' That count on how you look at it . You have different degrees of self-directed controller of a greenhouse . Autonomous climate regulation , for example , is already well imbed in greenhouses in the Netherlands and other Western state . If you talk about autonomously defining growing strategies or apply machinelike proficiency to crop direction , that is still evenhandedly much in its infancy . That is one of the things we hope to explore further with the Challenge . How far we are exactly and whether fully autonomous grow is potential at all , we do n’t really screw yet . That in turn makes this very interesting . '
generator : Wageningen University & Research
