A computer vision solution based on landing mark for guiding an unmanned helicopter to the ship deck was studied. Factors which limited the pose estimation efficiency were analyzed and a novel information-blocking based image processing method was proposed. With the optional coordinate transformation the ideal region for image processing was obtained, and by applying the technique of line direction coarse identification, the line parameters scope was determined and the space need to be traversed in line hough transformation was narrowed. The computation of low layer image processing was reduced. Tests on real image show that the target detection and state estimations in a 768 576 pixel size image may be completed within 30�fms. The state estimations are accurate to within 2�fcm in each axis of translation and 1.5��°�� in each axis of rotation when the equivalent distance between the helicopter and the ship deck is about 10�fm,and the algorithm is effective for the control of unmanned helicopter autonomous landing on ship deck.
Wang Xiaojian, Pan Shunliang, Song Zishan, Shen Weiqun.Computer vision scheme for autonomous landing of unmanned helicopter on ship deck[J] JOURNAL OF BEIJING UNIVERSITY OF AERONAUTICS AND A, 2007,V33(06): 686-689