Obstacle identification is one of the critical technologies of unmanned vehicle, edge detection is the basic step of obstacle identification based on video sensor and the magnitude guarantee of identification effect. In order to meet the demand of accuracy, real-time and stability of obstacle identification, a new multiple order morphology edge detection algorithm is proposed. We adopt two-dimensional histogram oblique segmentation to locate edge, then detected edge by improving the existing mathematical morphology edge detection operators and using appropriate structuring elements and percentile. Experimental results showed the edge detected is exquisite, continuous and intact. The algorithm possesses good robustness for different noised images, cuts operation time by nearly half compared with algorithm without edge location, then makes a good foundation for subsequent processing of obstacle identification.