In the recovery of oil spill, in order to improve the oil recovery ships automation degree and work efficiency, aimed at the problem of mechanical arm fast capturing oil fence front buoy on the oil recovery ship, an accurate positioning method is put forward based on monocular vision. The method was used of computer to control the camera to acquire image information, by the Canny operator edge detection method to extract the edge of the buoy, using of the Hough transformed the image space corresponding to the parameter space, and utilizing the three dimensional (3D) space parametric clustering to extract the buoys center coordinates. Finally, the coordinate data is transmitted to the lower computer, driving the robot arm to complete the task of fast capturing buoy and expanding the oil boom.