USA
September 29, 2026
A DASH researcher services the BenchBot control box. (Photo by Kayla Driver, ARS)
ARS’s Digital Agricultural Systems Hub (DASH) aims to harness computer vision and artificial intelligence (AI) to help researchers and producers better understand what is happening in their fields and respond more quickly to emerging challenges.
American farmers are being confronted with a growing set of pressures including weather variability, herbicide-resistant weeds, soil degradation, labor challenges, and increasing production costs. At the same time, fewer farmers are managing larger operations and must make increasingly complex decisions across thousands of acres.
To address these challenges, farmers need affordable precision agriculture tools that provide real-time, site-specific insights to improve productivity, reduce inputs, and build resilient farming systems.
DASH is designed to help meet that need by unlocking precision agriculture for farmers through computer vision solutions like weed and disease mapping.
Developed by ARS scientists and collaborators, DASH is a growing national digital platform that develops and integrates large agricultural computer vision training datasets, machine learning models, and deployable sensing technologies. The platform provides open-source tools that enable researchers, breeders, and producers to detect crop stress, identify weeds and pests, and make data-driven management decisions across fields and farms.
A DASH researcher walks among hundreds of plants as BenchBot glides above. (Photo by Kayla Driver, ARS)
“DASH is helping us turn years of agricultural research into deployable digital tools,” said Steven Mirsky, ARS’s Director of Digital Agriculture and DASH co-director in Beltsville, MD. “We are building systems that allow us to see what is happening in agricultural fields in unprecedented detail and translate that information into actionable knowledge for farmers.”
To build these tools, researchers are integrating imagery and sensor data from satellites, drones, field robots, and farm equipment to capture real-world conditions across agricultural landscapes.
“We are combining these data streams with AI to unlock and scale precision agriculture for farmers — from individual plants to whole farms,” Mirsky said.
Several key technologies power the DASH platform.
One of these resources is the National Agricultural Image Repository (AgIR) — an open-source, cloud-based library containing over 50 terabytes of high-resolution, metadata-rich images of crops and weeds. These images form the foundation for training computer vision systems that allow machines to recognize plants and agricultural stresses in the field.
The imagery is generated in part by BenchBot, a modular automated camera system developed by ARS researchers to rapidly collect and annotate or label agricultural images to be used for training computer vision models.
“BenchBot allows us to generate large, high-quality datasets that are essential for training artificial intelligence models,” Mirsky said. “These models can then be used to improve plant breeding, accelerate crop research, and support everyday decision-making on farms.”
This digital system hits the field via ModCam, a modular camera system mounted on farm equipment such as spray booms, which uses computer vision AI models trained from the AgIR image repository to map and identify plants in breeding trials, research plots, and farm fields.
These tools allow scientists to rapidly phenotype crops for breeding programs while also creating practical field-scale tools that support precision agriculture.
“We are taking the knowledge generated through agricultural research and turning it into deployable technologies,” said Amanda Hulse-Kemp, DASH Co-Director in Raleigh, NC.
“For example, DASH is using computer vision to enable precision nitrogen management with cover crops, as well as mapping weeds in plant science research plots and on farmers’ fields with the goal of developing integrated weed management systems. We are currently working with sugar beet grower associations to look into cercospora leaf spot (CLS) in Fargo, ND, where we are hoping to develop new ways to map CLS in real time, which will hopefully empower new ways for the growers to manage the disease.”
All DASH data and models are processed through USDA’s SCINet, the agency’s high-performance computing system that enables large-scale data analysis and AI development across ARS research programs.
By combining advanced sensing technologies, large-scale data resources, and artificial intelligence, DASH is helping to bridge the gap between agricultural research and on-farm innovation.
As the platform grows, it is expected to support a new generation of digital agriculture tools that strengthen the resilience, productivity, and sustainability of American agriculture.
For more information, visit the DASH website
Topics
Species
