This paper proposes a small target detection and geolocation software. In order to address the difficulty in detecting pixel-level small targets based on their shape, we utilize K-means clustering based on their characteristic colors for detection. To determine the latitude and longitude of a specific location within a satellite captured remote sensing image, we perform image registration with a map to calculate the pixel coordinates of the small target. Finally, the folium library is used to parse the latitude and longitude of that location. The experimental results are consistent with manual annotation and show high accuracy.