Use of Approximate Set Differences to Infer the Movement of Objects in Point Clouds
AuthorClifford, James A.
AdvisorOlson, Eric J.
StatisticsView Usage Statistics
To support the mathematical eld of change detection, we present a novel detec- tion algorithm based on approximate set di erences. Our goal is to determine how e ective the algorithm is for detecting the movement of objects and what factors a ect its performance. A theoretical foundation for this algorithm is built by prov- ing mathematical theorems relating to its e ectiveness. To optimize calculations, we developed a method of recursively boxing a point cloud into a tree structure. We then tested computational e ciency by applying the algorithm to synthetic data. Further optimization was performed by parallelizing the calculation of the approxi- mate set di erence. Finally, the e ectiveness of the algorithm for detecting changes was determined by testing data from a 3D scanner. The results indicate that the approximate set di erence is an e cient algorithm that can e ectively detect changes in three-dimensional data in O(n log2 n) time.