On the use of IMUs in the PnP problem

In this paper the problem of estimating the relative orientation and position between a camera and an object is investigated. It is assumed that both the camera and the object are provided with an Inertial Measurement Unit (IMU) capable of measuring their attitude with respect to the gravity and the earth magnetic vectors. Furthermore, the object is assumed to contain a feature of n points, the position of which is known in the object coordinate frame. An algorithm is proposed, which uses the image provided by the camera and the information provided by the IMUs, to solve the PnP problem, i.e., to estimate the relative pose of the object in the camera reference frame. Two special cases will be studied. The first is the case where all the attitude information given by the IMU is used. In the second case only the measurements provided by inclinometers are used, neglecting those coming from the magnetometers, because they are usually quite noisy. The effectiveness of the proposed algorithms is tested either by numerical simulations and by experimental tests with cameras.