In this paper we present a technique for identifying the presence of stagnant water bodies in images taken in various settings. Stagnant water bodies, such as puddles, can become sites for mosquitos to grow, which increases the likelihood of the spread of diseases, such as Zika . We observe that existing techniques perform poorly on images of variable quality, variable backgrounds, and focus. We present a technique that is able to successfully identify puddles in these images with about 88% accuracy and over 95% recall. Our technique is robust to image focus, background, and quality, making it a good choice for datasets produced by various kinds of cameras, including those on mobile phones, thermal imagers and unmanned aerial vehicles (UAVs).