SP / BUILD LOGCODE · DATA · SYSTEMS

A DEVELOPER IN THE MAKING / KOLKATA, INDIA

Sayan Paul.

BUILD · DEBUG · DOCUMENT

Opening the workbench...

PROJECT / 08BUILD · DEBUG · REFINE

Sapling Survival Analysis System.

A computer-vision pipeline for estimating sapling survival from drone-based orthomosaic imagery. It aligns pre/post-weeding surveys, detects plantation pits, classifies survival status, and prepares GIS-ready overlays for field verification.

PROJECT NOTE · 2026AI / COMPUTER VISION · 2026
PIT DETECTION / FIELD SURVEY

01 / THE PROBLEM

Comparing two field surveys needs consistent locations, not just two images.

The pipeline focuses on matching pre- and post-weeding imagery, detecting plantation pits, classifying their survival state, and producing overlays that a person in the field can verify.

02 / DESIGN DECISIONS

Image alignment

Treat survey alignment as a prerequisite before comparing pit positions.

Pit detection

Create consistent candidate planting-pit locations for later classification.

Survival classification

Separate model estimates from field-verified labels.

GIS-ready output

Produce map overlays and exports that fit field-review workflows.

← Back to the project explorer