New algorithm of spatial positioning based on mono-vision

To solve the problem of camera spatial positioning based on mono-vision, we propose what we believe to be a new algorithm, which is essentially based on the following idea; we find the relative attitude parameters between the camera and the special marked point with the spatial geometrical information of the marked point; then we infer the camera's spatial positioning coordinates with such geometrical information. Instead of complicated nonlinear calculations for each frame, the new algorithm obtains the initial values for consecutive attitude calculations from a particular attitude between the camera and the marked object, and then it calculates the attitude tracking of the next frame with the constraints between two consecutive frames. Section 3 of the full paper gives and analyzes experimental results, thus showing preliminarily that the new algorithm not only guarantees the tracking precision but also reduces significantly the amount of calculation in experimental and engineering applications.