We present an optimum camera placement algorithm. We are motivated by the fact that the installation of security cameras is increasing rapidly. From the system cost point of view, it is desirable to observe all the area of interest by the smallest number of cameras. We propose a method for deciding optimum camera placement automatically considering camera specification such as visual distance, visual angle, and resolution. Moreover, to reduce the number of cameras, we divide the scene into regions and weight them according to their importance of observation. Then, we calculate camera locations which cover all the important regions and as many other regions as possible by considering the trade off between a number of observed regions and the number of cameras required. We formulate this problem as a set covering problem.