Service robots are characterized by autonomously performing indoor tasks in unstructured environments, this condition of the environment prevents the prior programming of the map, which requires reactive behavior. These robots require real-time and cost-effective identification of obstacles in the environment, which includes not only distance information, but also depth information. This paper shows a strategy to estimate the position of obstacles in unknown environments. This strategy is characterized by low computational cost and real-time operation. The environments are selected because they are those usual to human beings, and this also influences our design, since we look for functional and morphological equivalence with human beings. This equivalence corresponds to the installation of two cameras in our robotic platform to form a stereoscopic system equivalent to the human. The images captured simultaneously are analyzed by a bacterial interaction scheme to define points on the obstacles. Our strategy showed a high performance in controlled environments. The scheme was able to establish distances to different points of the obstacle with 95% accuracy for distances between 0.8 and 2 m.