Smart Field Robot for Lime Farming in Bến Lức (Long An): Weed, Pest and Spray Decision Support in Canal-Based Orchards IoT, data collection and AI analysis

Capsaicin cooperates with The Fruit Republic and Hyves.Tech to bring technology into agriculture. The goal is to increase produce quality and quantity while reducing chemicals and waste. To actualize this, a smart farm management ecosystem was developed.

 

The first deployment is an autonomous robot, equipped with YOLO computer vision algorithm and built on the Robot Operating System 2. It uses an inertial sensor, encoders, GPS RTK, depth camera, and LiDAR to map, localize, and navigate its surroundings. The frame is modular, strong, and affordable. It can carry payloads such as sprayers, rotary cutters, or seed planters.

Next, sensor nodes and multispectral cameras are placed around the farm. Commercial cameras are expensive, but based on tree traits, ours was made from cheap and off-the-shelf components. After heavy calibration, the multispectral images identify tree health close to the ground-truth. Main soil and air data include humidity, temperature, PH, CO2, dust, and pressure. 

Data, detections, video streams, and images are sent to the Web Dashboard for analysis and actionable recommendations. LoRa was used for transmitting long-range, low-bandwidth data, and Zenoh for high-bandwidth, visual data through 4G.

The project addresses basic human needs, long-term food security demands, and efficient farming management.