Detection of Conductive Objects with Electrical Capacitance Tomography

We present an Electrical Capacitance Tomography (ECT) based object detection system, which can work with conductive objects. For grounded objects, the coupling to the distant ground is considered in the forward problem by means of coupling conductivities. For the reconstruction of the material distribution within the Region of Interest (ROI) we use a linear Bayesian inversion method in combination with a non-linear artifact reduction. The system achieves a high object detection rate in the ROI and allows to select a subregion of the ROI, where objects should be detected. This is beneficial, e.g., for robot grasper applications. The proposed light-weight signal processing chain, which has real time capability, allows to integrate this sensing technology on resource limited mobile platforms.