Satellite/Aerial Image Segmentation

Satellite/Aerial Image Segmentation

Project Highlights

To find the different segments in satellite image (i.e.buildings, roads, trees, crops and water) for smart city planning

Our Solution

We have developed a Deep Neural Network by using the following data

• The dataset consists of 8-band commercial grade satellite imagery taken from Space Net dataset.
• Train collection contains few tiff files for each of the 24 locations.
• Every location has an 8-channel image containing spectral information of several wavelength channels (red, red edge, coastal, blue, green, yellow, near-IR1 and near-IR2). These files are located in data/mband/ directory.
• Also available are correctly segmented images of each training location, called mask. These files contain information about 5 different classes: buildings, roads, trees, crops and water (note that the original Kaggle contest had 10 classes).
• Resolution for satellite images is 16-bit. However, mask-files are 8-bit.