As part of my University RA, I developed a Deep Learning model capable of detecting diseased wheat leaves amongst the healthy ones with favourable scores and help the user(mostly farmers) classify the healthy vs diseased wheat plants.
The dataset contains 4800 images in total with
- 1279 Healthy Wheat
- 939 Wheat Loose Smut
- 1622 Leaf Rust
- 960 (after refining) Crown and Root Rot
A new type of technique has been utilised (by browsing multiple sources) of classifying each type of plant image into a binary array (4 arrays used) for a favourable and easy classification portrayal alongside using Neural Networks.
For a small cycle of 30 epochs only (faster model result check), a favourable score of 87.31% accuracy has been achieved.
Feel free to contact me on my LinkedIn.