System and Method for Selective Treatment of Crops Using Machine Vision

Abstract

There is provided a system for customized application of herbicides, comprising: a processor(s) executing a code for: feeding test images corresponding to a target agricultural field into a machine learning model trained on a training dataset of sample images of sample agricultural field(s) labelled with ground truth of weed parameters, selecting specific weed parameter(s) of according to performance metric(s) of the model, setting up instructions for triggering application of a first herbicide to a portion of the target agricultural field in response to an outcome of the model indicating likelihood of the specific weed parameter(s) being depicted in an input image of the portion of the target agricultural field, and setting up instructions for triggering application of a second herbicide to the portion of the target agricultural field in response to the outcome of the model indicating non-likelihood of the specific weed parameter(s) being depicted in the input image.

What is claimed is: 
     
         1 . A system for dynamic application of herbicides to a target agricultural field, comprising:
 an optical imaging apparatus that produces acquired images of vegetation in said target agricultural field;   at least one hardware processor executing at least one machine learning model to detect weeds in said agricultural field based on said acquired images;   wherein said at least one machine learning model is trained on at least one corresponding training data set comprising a plurality of records which in turn correspond to a plurality of sample images of one or more weeds and one or more ground truth labels representing one or more weed parameters.   
     
     
         2 . A system for controlled application of sprayable chemicals to an agricultural field, comprising:
 an image acquisition sensor configured and arranged to acquire image data of at least a portion of said agricultural field;   a first chemical storage container, containing a first herbicide that is specific to treat for a target growth;   a spot sprayer, that controllably sprays said first herbicide from said first chemical storage container onto a target area of said target growth;   a second chemical storage container, containing a second herbicide that is not specific to said target growth;   a broad sprayer that sprays said second herbicide from said second chemical storage container onto a broad area of said agricultural field;   a controller that executes computer readable instructions that determine operation of said spot sprayer responsive to at least said acquired image data as well as input from any of a machine database or a manual input.   
     
     
         3 . The system of  claim 1 , said controller comprising a processor configured and arranged to determine a target growth parameter in said portion of the agricultural field, and said spray controller triggering said spot sprayer to apply said first herbicide based on said target growth parameter. 
     
     
         4 . The system of  claim 1 , further comprising a mobile farm vehicle on which said first and second storage containers, said spot sprayer and said broad sprayer are disposed. 
     
     
         5 . The system of  claim 1 , the controller employing said machine database if the machine database generates a target growth parameter with a likelihood above a set threshold, and employing said manual input if the machine database input does not generate a target growth parameter with a likelihood above a set threshold. 
     
     
         6 . The system of  claim 1 , said second herbicide comprising a residual herbicide chemical. 
     
     
         7 . The system of  claim 1 , wherein said manual input corresponds to any of: a visual examination of said field and a domain expert input comprising knowledge of the field or target growth. 
     
     
         8 . A method for applying chemicals to an agricultural field, comprising:
 acquiring a plurality of images representing portions of said agricultural field using one or more image capture sensors;   in a processor that executes programmed instructions, receiving said plurality of images and processing said images using a machine learning classifier;   determining a performance metric corresponding to an accuracy or classification of a target growth;   if said performance metric is below a predetermined threshold within an area of said field represented by said images, spraying a first herbicide onto said area of said field;   if said, performance, metric is at or above a predetermined threshold within, said, area of said, field, represented by said images, spraying a second herbicide onto said area of sad field.   
     
     
         9 . The method of  claim 8 , wherein said first herbicide is broadcast sprayed onto a broad area within said area of said field. 
     
     
         10 . The method of  claim 8 , wherein said second herbicide is spot sprayed onto an area narrower than said broad area.