Robotics
Current- Developed a Deep Net based tree stem/bark image segmentation approach for tree biometrics like diameter at breast height
- Developed a ML augmented Deep Net model for high dimensional Hyperspectral data to detect early signs of Spruce Bud Worm attack on forest trees
- Developed a localization platform for localizing under challenging dense forest canopy based on INS and Visual Odometry fusion
- Used Clearpath Husky UGV to setup and prepare Navigation stack for forestry autonomous traversal
- Key words: Python, Tensorflow, Keras, Machine Learning, DNN, CNN, Pattern Recognition, non-parametric statistics, hyperspectral imaging, ROS