Optimizing the conversion of solar energy to electricity is central to the world’s future energy economy. Solar panels installations require measuring the total area of the roof and any areas on the roof that may account for gaps due to vents, pipes, roof fixtures, or tree shading. There are several options for measuring roofs (e.g., climbing the roof, using drones from the floor, manually measuring with Google Maps). However, these options are usually very expensive, risky, and time-consuming. In this work, we developed a methodology that automatically creates a three-dimensional (3D) model of the house and measures its roof using pictures of its elevation views (north, south, east, and west) and satellite images from Google Maps. After the 3D model is created, a top view projection of the house’s ceiling is used for polygonal shape matching with a previously known ceiling from Google images. The matching problem is multimodal, noisy, and highly dimensional since there are several roofs similar to the projection, they are covered by leaves and the search is done with high-dimensional images. In recent years, meta-heuristic algorithms (MHs) have arisen as effective algorithms to solve such multimodal optimization problems. Among the MHs, population based algorithms when applied with a niching strategy, called niching algorithms, have proven to be able to identify such optima. In our previous work, we have used GA to solve the above problem. In this chapter we use genetic algorithms (GA) with niching methods and design a new fitness function to better cope with the multimodality of the problem. After the best matching polygon is chosen from the set of possible solutions, the scale given by the satellite image is used to rectify the 3D house-model proportions and to estimate the space available for a solar installation.