John Deere See & Spray: AI Slashes Herbicide Use by 50%

The Challenge

In conventional agriculture, farmers rely on blanket spraying of herbicides across entire fields, leading to significant waste. This approach applies chemicals indiscriminately to crops and weeds alike, resulting in high costs for inputs—herbicides can account for 10-20% of variable farming expenses—and environmental harm through soil contamination, water runoff, and accelerated weed resistance [1][3]. Globally, weeds cause up to 34% yield losses, but overuse of herbicides exacerbates resistance in over 500 species, threatening food security [6].

For row crops like cotton, corn, and soybeans, distinguishing weeds from crops is particularly challenging due to visual similarities, varying field conditions (light, dust, speed), and the need for real-time decisions at 15 mph spraying speeds. Labor shortages and rising chemical prices in 2025 further pressured farmers, with U.S. herbicide costs exceeding $6B annually [2][4]. Traditional methods failed to balance efficacy, cost, and sustainability.

Quantitative Results

• 5 million acres treated in 2025
• 31 million gallons of herbicide mix saved
• Nearly 50% reduction in non-residual herbicide use
• 77+ weed species detected accurately
• Up to 90% less chemical in clean crop areas
• ROI within 1-2 seasons for adopters

Implementation Details

Development Timeline and Approach

John Deere's journey with See & Spray began with the 2017 acquisition of Blue River Technology, pioneers in AI weed control, accelerating from prototype to commercial launch.[3] The Premium version debuted in 2020 for cotton, expanding to corn and soybeans by 2022. In 2024, the more affordable See & Spray Select launched for retrofitting existing sprayers, broadening adoption. By 2025, it covered 5 million acres, with ongoing updates via over-the-air software for new weed/crop models.[1][7]

Technical Stack and AI Architecture

The core is computer vision object detection using convolutional neural networks (CNNs) like custom YOLO variants, trained on millions of field images labeled for 77+ weeds, crops, and backgrounds. Boom-mounted cameras (one every 10 inches, up to 140 per 120-ft boom) stream at 20+ frames/sec, processed by edge GPUs for <50ms inference—essential at 15 mph. Nozzles (up to 4 per nozzle body) pulse individually, with droplet sizes optimized to minimize drift.[2][5] Integration with John Deere's Operations Center provides data analytics, mapping treated areas for variable rate future applications.

Deployment and Farmer Integration

Implementation is seamless: New buyers get factory-installed on models like R4025 Sprayer; retrofits for Select take hours. Farmers calibrate via app for crop type, with AI auto-adapting. Field trials showed 99% crop protection (no spray) and 80-90% weed kill with less volume. Global rollout targets Brazil/Europe by 2026, addressing labor shortages.[4]

Challenges Overcome

Key hurdles included lighting variability (dawn/dusk, shadows), solved by multi-spectral imaging and data augmentation; real-time speed via optimized models (TensorRT); and weed diversity, tackled with continuous farmer-submitted data for retraining. Initial high costs dropped 40% with Select, achieving payback in 1 season at $20-30/acre savings. Regulatory approvals for reduced chem use boosted adoption.[2][6]

Overall, the phased approach—pilot, premium, mass-market—drove scalability, with 2025 marking mainstream impact amid climate pressures.