Image/Audio Classification, Gis Modeling, Web Programming
summary of some projects:• Classifying tree type using 3d photogrammetry derived point clouds.(Keras/Tensorflow, pointnet, laspy, etc.) • large scale wetland mapping using segmentation neural network(Python, Keras/TensorFlow, GDAL, Google Earth Engine, etc.) Publication: https://www.mdpi.com/2072-4292/12/1/2 .• bird species identification from audio recordings (neural networks, Keras/Tensorflow, audio processing, etc.).• camera trap image analysis to determine presence/absence of animals using a deep neural network deployed in the cloud in 2015 (Caffe2, OpenCV, Python). Updated in 2016 to auto-identify domestic cows. Still running today.• large scale predictive soil/ecosite mapping (pixel based, random forest, Scikit-learn, SAGA, etc.). Report: https://www.abmi.ca/home/publications/451-500/498.html%3Btime=2016 .• web sites with web mapping (PHP, JavaScript, MySQL, OpenLayers, MapServer).• wrapping Netlogo models into a server-side application with real-time GIS post processing, web front end (PHP, JavaScript, Java, Python, GDAL).• linear feature detection from Lidar using predictive modeling and computer vision (SciPy, Pylearn2/Theano, Scikit-learn).