The aim of this presentation is to present a fusion of methods for better object detection in satellite imagery, and apply it to a quantitative analysis of human rights crises. Refugee camps in and around Syria will serve as a case study for this presentation. This research illustrates how traditional remote sensing methods such as, Spectral Angle Analysis (SAM), and machine learning methods, specifically Convolutional Neural Networks, can be used for refugee tent detection. SAM, as a method of tent detection, was used over a time series of images, and achieved an accuracy of 88 percent in the Rubkan refugee camp on the border between Syria and Jordan. To scale this analysis over larger areas of satellite imagery, we will discuss using SAM to generate labeled training data for CNNs and the result of applying them over a large area.