The Dynamic Machine Learning projection aims to model and prognosticate complex quality trait in unfermented produce . One of the work packages is depend to create an automated method acting to make a survival of only healthy new plants to transplant into the glasshouse . WUR Vision + Robotics researchers have team up with Viscon to evolve a new applied science that can automate this horticultural unconscious process . The engineering science could enable growers to sieve out non - viable plant before moving them into the greenhouse , saving resources for healthy plants .
The project focuses on plants from tissue culture as one of the current key commercial agricultural crops , but the technology the researchers aim to develop could be applicable to various other important agricultural crop . The industrial plant start out as tissue polish , are then transplanted into trays , and lastly the young plant are transmit into a greenhouse . The sorting and option of healthy plants that are to be transplanted into the next microscope stage of culture is often done manually by growers . The visual scrutiny of the plants at each microscope stage is a laborious project . In Dutch greenhouses , the sorting process is already automated , but what makes the current research unique is that the marking of the plants and the prediction if they can survive the adaptations to the greenhouse would be done by AI .
Machine erudition and plant phenotyping could make selecting and classify healthy young plants easierSuzane Pols , team Pb of Plant Science at Viscon : " Machine learning and plant phenotyping are two blistering topics in USDA and horticulture . We need to make use of these study to bring value to the cultivation process of tissue - civilization - grown plants . The doubt is , can we transform what we visually notice in plants and the datum that we gather from those observations into utile slice of information raiser can work with ? The goal is to make the selection and categorisation of intelligent young industrial plant for transplant a fully automated and less laborious undertaking for agriculturist . "
And what is a healthy plant life ? " That is highly debated in works skill . We trust the automobile learning helps us to identify what fundamental characteristics are prognosticator of plant wellness and then helps us to separate those respectable plants so that we can help improve and make the sorting process more effective , " say Pols .
RGB camera with pecicentric lens"What we are develop is a technology that can take a photo of the roots and the shoot of the plant , and utilize paradigm processing and eventually deep - learning to key out if that plant is healthy . We desire this method acting to be automated , reliable , quotable , and scaleable , " explains Lydia Meesters , project loss leader and scientist of Computer Vision at WUR ’s Vision + Robotics curriculum .
The project has already bring about a new technical innovation : the team have developed an RGB - camera with a pericentric lens that can capture 360 grade in every direction in one exclusive image . The lens not only visualizes the bottom of the unseasoned flora plug but also the side , giving the research worker a complete image of the root of the plant .
" We are in the centre of our feasibleness survey : can we take images of the plant characteristics that are need to influence industrial plant health ? And if yes , how can we visualise these traits optimally using an wanton and scaleable technique ? That ’s something we will proceed to search in the arrive geological period , " concludes Meesters .
For more information : Viscon Plant Technologywww.viscon.eu/plant-technology
