Skip to content

Computer Vision · Environmental

Drone Imagery Analysis System

A scalable CV pipeline processing high-resolution aerial imagery — detecting and segmenting trees, buildings and pipelines to automate monitoring across agricultural and industrial sites.

OpenCVPyTorchYOLOv8Mask R-CNNCOCO

70%

Labelling efficiency gain

Edge

Optimised for Jetson Nano

What we built

01

Modular CV pipeline

An architecture handling data ingestion through to model training.

02

Automated annotation

A YOLO pre-labelling tool with human review, cutting annotation time 70%.

03

Edge optimisation

Inference tuned for deployment on Jetson Nano hardware in the field.

04

Detection & segmentation

YOLOv8 for fast localisation, Mask R-CNN for pixel-level area measurement.