Grape Detection in Vineyards Final project by Yael Peretz & - PowerPoint PPT Presentation

Grape Detection in Vineyards Final project by Yael Peretz & Amir Yeger

Introduction • Detecting grapes via color recognition and clustering.

Goals • Circle recognition – unsuccessful • Color filtering • Clustering • Noise removal Noise removal • Filling in non-desired filtered areas

Implementations Implementation #1: • Color filtering • Clustering • Result – better grape detection, poorer Result – better grape detection, poorer background filtering!

Implementations • Implementation #1 – Color filtering

Implementations • Implementation #1 – Clustering

Implementations • Implementation #1 – Result

Implementations Implementation #2: • Color filtering • Result – better background filtering, poorer grape detection! grape detection!

Implementations • Implementation #2 – Color Filtering

Implementations • Implementation #2 – Color Filtering

Implementations • Implementation #2 – Result

Results - 1

Results - 1 Implementation -1 Implementation -2

Results - 2

Results - 2 Implementation -1 Implementation -2

Result - 3

Result - 3 Implementation -1 Implementation -2

Results – 4

Result – 4 Implementation -1 Implementation -2

Conclusions • Color filtering – not the perfect solution • Clustering – efficient, but with the use of good color filtering • Both algorithms achieved better results on Both algorithms achieved better results on images whose grapes had a lighter color than images with darker grapes

Questions?

No…? ok, thank you