Intelligent Mapping of the Farm Using Low-Cost, GPS-Enabled Cameras During Existing Farm Activities

With the advent of precision agriculture the way we farm is changing. Technologies exist to addprecision to all aspects of the farm; however these practices all require information that is notreadily available to the small farm. Current field mapping technology has a financial and timecommitment that puts it out of reach of small farms. The aim of the proposed research is to enableprecision agriculture on the farm by augmenting existing farm activities where the farmer drives thewhole field with a machine vision system inside the cab of the farming implement.The action camera market has brought rugged wide-field-of-view GPS-enabled cameras into thereach of the common consumer and thus the small farm. The rapidly growing field of artificialintellegence (AI) has lead to the development of embedded AI systems such as the NVIDIA Jetsonwith higher capabilities and lower costs than most desktop computers. The combinaiton of devicesenable automated mapping of the farm by identifying anomalies in the field adjusting for variationsin lighting and finally robust object characterization. This system must be ever evolving where itself determines when it fails to characterize an object and enters a learning mode where with onlyminimal human interaction it can learn how to label the anomalies present in the field. Combinedthis hardware and software package named Groundskeeperâ„¢ can deliver actionable informationneeded to enable precision agriculture on the small farm at a reasonable price.