The USDA Agricultural Research Service described its developing DASH infrastructure in a September 28 report. The platform brings together data, models and sensing equipment for crop-stress detection and weed or disease mapping.
Its National Agricultural Image Repository, AgIR, contains more than 50 terabytes of crop and weed images with metadata. BenchBot, an automated modular imaging system, helps create and label the training material. ModCam takes related computer-vision capabilities into the field using cameras attached to equipment such as spray booms.
ARS identifies applications in breeding trials, cover-crop nitrogen management and weed mapping. Work with sugar beet grower associations is investigating improved real-time mapping of cercospora leaf spot. USDA’s SCINet computing infrastructure supports processing and model development. The announcement describes a growing research-to-deployment platform, rather than evidence that every proposed application is already commercially validated.
