In this work, an extended approach to surveying urban trees based on automatic detecting from images captured using unmanned aerial vehicles (UAVs) is presented. Such method is a cost-effective alternative to traditional measurement techniques. Through autonomous flights, UAVs capture detailed aerial imagery of urban areas, which is then processed to generate high-resolution raster images and elevation models. Machine learning algorithms are then applied to these images to identify trees, refining the detection process by eliminating false positives and estimating tree heights. Dealing with challenges such as flight time limitations and the irregularity of urban trees, the method achieves a great accuracy in tree identification, not only covering the tree detection but also the separation between sidewalk and block interior trees, the estimation of height among other important data. Even that the proposed methodology is not new, the implementation of it with particular variants results effectively to detect most of the trees in a short time and it serves as a complementary tool to ground survey systems.