This paper presents a new automated method for the detection and determination of building heights using their cast shadows. The approach consists in applying image processing using PCA and segmentation for the detection and recognition of buildings and their shadows. The height of the buildings is deduced by knowing the length of their shadow projected on the ground, the position (azimuth and zenith) of the sun and the sensor at the time of acquisition. These shadow analyses were carried out on a free satellite image from Google Earth. The results of the height calculations are used for the three-dimensional modelling of the buildings.The 3D models produced can be used for strategic decisions in the professional field and for urban monitoring and surveillance, as well as for various research studies on the relationship between building heights and natural and man-made phenomena: energy consumption and land subsidence. Our method, which requires a good precision of the geometric characteristics of the proposed remotely sensed data, has outperformed the majority of existing research as an automated approach to exploiting the shadows of several buildings in a single satellite image and their 3D reconstruction.