Loop Closure Detection Using Local 3D Deep Descriptors

We present a simple yet effective method to address loop closure detection in simultaneous localisation and mapping using local 3D deep descriptors (L3Ds). L3Ds are emerging compact representations of patches extracted from point clouds that learnt data learning algorithm. propose novel overlap measure for by computing the metric error between points correspond mutually-nearest-neighbour after registering candidate cloud its estimated relative pose. This approach enables us accurately detect loops estimate six degrees-of-freedom poses case small overlaps. compare our L3D-based with recent approaches on LiDAR achieve state-of-the-art accuracy. Additionally, we embed RESLAM, edge-based SLAM system, perform evaluation real-world RGBD-TUM synthetic ICL datasets. Our RESLAM better accuracy compared original strategy.