Image-free localization and tracking of multi-targets based on single pixel detection

Abstract

Using single pixel detection method to locate and track the target without image has the advantages of low sampling rate and high tracking speed. However, previous image-free positioning scheme can only locate and track one moving object at a time. When there are multiple moving objects, entanglement solutions will be generated. This paper proposes an image-free localization and tracking scheme of multi-targets, which can simultaneously track multiple targets under low sampling rate. The scheme consists of three steps: single pixel measuring, matching and positioning. Firstly, we design two groups of patterns of digital mirror device and project them to the targets in different angles successively, then two sets of bucket detector values are obtained. Secondly, we treat the vector of bucket detector values as virtual images and use the image matching method to get the possible locations of multiple targets. Finally, by comparing the possible locations obtained from the two sets of bucket detector values, we can get the correct positions of the targets. We verify the effectiveness of the proposed method under different noise conditions with simulations. This method improves the previous image-free localization and tracking scheme, it can simultaneously locate and track multiple targets in the case of low sampling rate, which provides an idea for future multi-target localization and tracking schemes.