Soil crack detection: Datasets & Models

Manually inspecting soil for cracking across large agricultural fields is slow, subjective, and impossible to do consistently at scale, especially when trying to catch early signs of moisture stress or erosion. Automating soil crack detection with computer vision lets teams scan entire plots from drone or fixed-camera imagery and get objective, repeatable measurements of crack severity. This helps agronomists prioritize irrigation, flag areas of soil degradation, and correlate cracking patterns with water management decisions. The same pipeline runs the same way every day, removing inspector-to-inspector variability.

Instance segmentation delivers per-crack masks, so you can quantify cracked surface area and track how it changes across a season rather than just flagging presence. The model works with the cameras and drones you already fly, and can be deployed at the edge or in the cloud to process imagery at field scale. As your acreage or camera count grows, the workflow scales without adding manual inspection labor.

1. Upload an image: Drop any photo into the widget above to try soil crack detection for free, right on this page.
2. Preview instant results: The model detects soil crack, dry cracked soil, fissure, erosion channel, bare soil within seconds and returns an annotated image with structured predictions, so you can immediately see how well it works. Adjust the detection classes and re-run until the results fit your scenario.
3. Like what you see?: With a free Roboflow account, you can deploy this workflow with the Serverless Cloud API, customize it in your workspace, or train RF-DETR (instance segmentation) on your own images for production accuracy.
4. Get started for free: No code or training data required to begin: upload, review the results, and make it part of your stack in a single click.