Case Studies

Soya Bean Row Detection & Gap Analysis

Improving crop spacing accuracy using aerial mapping

Mapped 150 acres across Galnewa, Talawa, and Maliyadevapura, covering 8,024 soybean plots to analyse row structures and spacing consistency. An image-processing algorithm was developed using orthomosaic data to detect crop rows and identify spacing gaps. By generating binary masks and analysing pixel distributions, the system accurately counted rows and highlighted plots with deviations from standard spacing—enabling targeted intervention for improved yield efficiency.

Corn Tassel Detection & Crop Counting

AI-powered plant counting for yield estimation

Developed a tassel detection and counting system using aerial imagery and object detection models. After preprocessing and noise removal, tassels were identified using Faster R-CNN (Detectron) trained on annotated data. The system successfully counted 4,060 tassels in a sample image, enabling estimation of plant population and identification of low-yield zones. This provides farmers with actionable insights to optimise crop management and improve productivity.

Yield Prediction Optimisation for Soya & Corn

Machine learning models to improve forecast accuracy

Developed a regression-based machine learning model to estimate crop yield using mapped field data from 150 acres and 8,024 plots. Despite limited plot-level inputs, the model achieved ~93% accuracy for soybean yield prediction. For corn, combining tassel detection with yield data resulted in ~95.6% estimation accuracy. This approach enables scalable, data-driven forecasting to support better agricultural planning and decision-making.