technical advancements in gadfly and disease reconnoitring are transforming a labor - intensive sphere into a more effective , information - driven version of itself . Now , as artificial intelligence operation ( AI ) is start to be developed to aid harvest output , growers must be more critical than ever in assessing the benefits of these other - stage solutions . Dr Mikkel Grum , Research and Development Director at global craw plague and disease chromosome mapping experts , Scarab Solutions , talk about the latest AI developments and explain why farm and harvest protection managers need to continue to concentre on the engineering augment human labor instead of holding out for the as yet unrealized hope of AI .
TheFood and Agriculture Organization of the United Nations estimatesthat between 20 to 40 percent of global harvest production is lost each year to pests and diseases , costing the global thriftiness around $ 220 billion . Pests such as thrips , aphids , leafage miners , mites , whitefly and caterpillars , and disease such as blights , mildews , botrytis and stalk and root rots are plebeian throughout all mood zones .
It is true that to become even more effective , craw management will require meliorate technique as well as technology . Many believe AI holds the solvent .
AI takes its first steps in horticulturePest and disease monitoring is a childbed - intensive process , requiring scouts to accurately evaluate works and crop health as they move across the greenhouse , field or farm . AI - driven image depth psychology aims to help automatize the craw surveillance process .
In gardening , late developments let in a‘robot scout’equipped with near - infrared image camera to detect powdery mold and image psychoanalysis to predict bud and heyday yields , and theIRIS Scout Robot . There is aremote pest monitoring systemusing political machine learnedness ( ML ) to conduct image analytic thinking on pheromone traps , and a orotund number proposing drone and satellite imagery as the fundament of future crop management operations .
More widely - spread head use eccentric promote the use of smartphone applications toscan photos for sign of pest and diseases , often present as quick , or nigh ready for prime time use .
At first glance this sounds plausible . Many have heard that Google ’s image analysis is now salutary than man at recognizing cats and dogs in figure of speech , or that inbreast malignant neoplastic disease enquiry , AI - image analysis now detects cancer on mammograms with more efficiency and accuracy than expert radiologist . So surely , using image analysis to identify harvest pests and diseases on photos direct with a smartphone ca n’t be that far off . Not so fast .
realness paint a less fortunate pictureCurrent efforts to habituate picture realization applied science in smartphones fall short of their promise to provide both a granular insight and actionable overview of farms and greenhouses .
As highlighted in a recentScientific American article , the statistics used to present how well image psychoanalysis works are often misinform at best . The most common ‘ pairing mental testing ’ that tests the ability to liken two image and state which of the two has the pest or disease , give much gamy percentages for truth than an analytic thinking of multiple images with no cognition of whether any of them have the pest or disease .
… and can create problems such as false positives .
Using inaccurate or skew resultant role gleaned from AI as a footing for pesticide purpose can cause more hurt than good , as illustrated by the issue of sour positives .
Let ’s ideate an imaging organization that gives a delusive positive of blight just five pct of the time , a very conservative figure of speech even by the claims of accuracy of any current app . In a theatre of operations full of blight , this would n’t pose a problem , but now allow ’s take a field that does n’t have any occurrence of the disease . If you took 2,000 images in that force field – which is the phone number of observance decimal point a skilled scout manages in a 24-hour interval – you would get 100 positive termination !
Does the grower turn on this result , or do they scrutinize the 100 “ positive ” locations to insure whether they really do have this problem ? Multiply this by the other cuss and disease that the image analytic thinking system is also checking for and perhaps has even higher untrue - positive rate for , and you have the workings of a hard-nosed nightmare . The higher the phone number of false positive , the more resourcefulness are require to conduct independent verification of event — think all the gains of automation are lost .
car versus humanThe approach also needs to be put in context . work comparing site where either there is AI or no harvest reconnoiter technology at all do not paint a naturalistic painting , because in some case there is already a organization in home that helps tape and psychoanalyse data compile by human lookout man .
In thebreast Crab researchcase , as tumors are not visible to the human middle , the Dr. and AI are looking at the same image . In a greenhouse circumstance , however , icon analytic thinking is much less effective than human aid to detail . A watch can move their foreland and turn over leaves to see the problem from multiple angles and with a magnifying chalk , therefore has a significantly better view of the issue than a smartphone epitome would have .
augment human skills with mobile applied science – smartphones make people smarterSimply put , farm and greenhouses still want the great unwashed to walk around , open up the crop canopy and twist over leaf and use a magnifying glass where needed . To build on this requires technologies that enable scouts to do their jobs more accurately , quicker and to a greater result — not technologies that ignore their expertise .
Smartphones are and will continue to be fundamental in this process — but not primarily as an AI tool . A more naturalistic and proven function of roving applications is for information collection and mapping purposes . Instead of using smartphones to take pic for AI to analyse , harvest auspices managers should indue scouts to use their inspection acquisition and memorialise the results as they go .
Training plays an crucial role in this process , not the least in tightening scouting timelines — another pain point AI expects to solve . Correct identification and marking of pests and diseases , exhaustive knowledge of sampling protocol and the technique to speed up the process are all involve to harmonise the execution and accuracy of scouts across the integral farm , which is key for winner .
AI may help guide scouts towards aright identify pests or diseases they amount across that they do n’t know , but most harvest scouting is about maintain track of the very dynamic distribution of a well - known exercise set of pests and diseases .
Digital mapping and scouting applied science enable man to glean novel insightsIf we couple up the data recorded by scouts with geographical info , the termination make datasets , bring home the bacon a clear audited account track for traceability and data point visualization alternative such as digital maps , charts and graphical record — and worthful assist to easily describe unequaled and recurring trouble and patterns with few , if any , false positive .
Digital mathematical function is where pestis and disease exploratory survey technology meet human expertise to optimise outcomes . At Scarab Solutions , we see this every day as clients useScarab Precisioncrop pest and disease reconnoitring and function solutions to provide a solid basis to nail infestation hotspots , determine the right pesticide use or biological control agent and reduce harvest departure by enhanced farm management .
As their datasets grow , crop protection manager can not only measure their own progress , but in some case benchmark against pest and disease physical body in their region , using anonymized datum from other farms .
Not yet AI ’s meter to glow , but we will always need the human touchWhile AI - drive image psychoanalysis remains a talking point in the diligence , the technology has a long path to go before it can farm authentic , precise and actionable use cases . Today , GPS - tracking , fluid data collection and rendition tools are the most effective and lucrative technology solutions for pestis and disease management for crops .
As horticulture undergo a technological transformation , artificial intelligence should not be examine as a reserve for existing processes , but as an extension of human intelligence . AI - driven effigy psychoanalysis will come with drones and robots in some setting , but that ’s a story for another fourth dimension .
For more information : Scarab Solutionswww.scarab-solutions.com
