CapsNet-weed-crop-Classification
A deep net classifier for the "Nut Grass-Chinese Broccoli" dataset
We created our own dataset on two species of plants and developed a vision-based classifier, which will enable future development in weed/crop detection algorithms and precision agricultural robotics. The classifier is based on recent researches on Capsule Networks and Residual Networks. The effects of various improving methods on the neural net architectures and the image processing techniques are explored.
Please read the CS_280_report.pdf for the ideas behind this project. The .ipynb file contains all the codes need for data and model. In the future I'll make the .ipynb file into separate .py files.