From an Undergraduate Final-Year Project to IEEE Access
Our research on multispectral imaging for soil texture has been accepted for publication in IEEE Access (SCImago Q1; Impact Factor: 4.2).
Incredibly proud to announce that our research at the Multidisciplinary AI Research Centre, University of Peradeniya, has been accepted for publication in IEEE Access (SCImago Q1; Impact Factor: 4.2)!
This milestone is especially meaningful as it originated as an undergraduate Final Year Project, showcasing the immense value of cross-disciplinary collaboration between the Department of Electrical & Electronic Engineering and the Department of Civil Engineering at the University of Peradeniya.
Soil texture (clay, silt, and sand composition) is fundamental for agriculture and geotechnical engineering. However, conventional lab tests like sieve and hydrometer analyses are labor-intensive and often take 1 to 2 days to complete.
To create a faster, accessible alternative, we developed a custom, low-cost reflectance multispectral imaging system capturing 13 narrow spectral bands across the near-ultraviolet, visible, and near-infrared regions (365–940 nm).
By combining this imaging data with advanced machine-learning models, our framework successfully delivers a rapid, non-destructive method for:
- Direct classification of soil into the 12 USDA texture classes
- Regression to accurately estimate clay, silt, and sand percentages
- Indirect classification by mapping estimated composition onto the USDA soil texture triangle
A huge congratulations to the brilliant team and supervisors who made this multidisciplinary research a reality!
Team Members
- Sandunika Ranasinghe
- Senith Jayakody
- Mario De Silva
- Gayathri Thilakarathne
Supervisors
- Prof. Roshan Godaliyadda
- Prof. Vijitha Herath
- Prof. Parakrama Ekanayake
- Dr. Sinniah K. Navaratnarajah