Considering wireless sensor networks for border surveillance, one of the major concerns is sensing coverage. Breach probability can be used as a measure to analyze the tradeoffs and to determine the required number of sensors to be deployed to obtain a certain security level. In this paper, the sensing coverage of surveillance wireless sensor networks is studied by utilizing a well-known image processing technique. By applying the watershed segmentation algorithm to the sensing coverage graph, the contour points, which are followed by the breach paths, axe found. To determine the weakest breach path by utilizing the contour points, Dijkstra's shortest path algorithm is used. Under uniformly distributed random deployment, the breach probability is evaluated against the sensor model parameters, field width and the number of sensors deployed.