Agriculture annotation: plant & leaf-disease datasets
Agricultural computer vision starts with a deceptively hard problem: everything is green, organic, overlapping, and variable. A leaf-disease model must distinguish early blight from nutrient deficiency from insect damage from ordinary senescence-distinctions that live in subtle texture, color gradients, and lesion morphology.
What we annotate in agricultural imagery
Disease datasets usually combine levels: a classification label for the disease, a segmentation mask for the diseased region, and attributes for severity. Every label carries attributes, so one annotation pass produces data for a classifier, a segmenter, and a severity grader simultaneously.
| Label type | Agriculture use | Examples |
|---|---|---|
| Instance segmentation | Per-leaf, per-fruit, per-plant masks | individual leaves, fruits for yield, weeds vs. crop |
| Semantic segmentation | Field & canopy analysis | healthy tissue vs. lesion area, soil, canopy cover |
| Bounding boxes | Detection at field scale | plants, pests, flowers, fruit clusters |
| Classification | Disease identification | early blight, late blight, rust, mosaic virus, healthy |
| Severity attributes | Grading & progression models | lesion coverage %, growth stage, severity grade |
AI-assisted labeling on organic shapes
Organic boundaries are where click-to-segment assistance shines. Leaves have serrated edges, lesions have irregular margins, and neither is fun to trace by hand. A click on a leaf returns its full boundary-serrations included-as an editable polygon; a rough box around a lesion cluster returns the lesion region. The annotator's time goes into the judgment calls (is this blight or burn?) rather than the geometry.
Class balance is the silent killer of agricultural models: diseases are rare relative to healthy tissue. Every delivery includes class-distribution reporting, and annotators flag underrepresented classes during labeling so collection can be redirected while the field season is still open.
Greenhouse video & plant-level tracking
Controlled-environment agriculture adds a temporal dimension field photos lack. A plant labeled once in a camera pass is tracked through the traverse with a persistent identity, and identities are linked across daily passes during review-producing per-plant timelines that disease-progression models, growth-rate estimators, and treatment-response studies all train from.
Formats & delivery
Agriculture datasets export as classification folder structures or CSV (for disease classifiers), COCO JSON and YOLO (for detection), and PNG masks (for segmentation)-with versioned splits that keep images from the same field or survey flight together, preventing the leakage that inflates validation accuracy. Multi-source datasets (drone + phone + greenhouse camera) are tagged by source.