A Distributed Approach of Accompany Vehicle Discovery

Accompany vehicle discovery is a hot topic in criminal investigation department with regard to massive vehicle data retrieval. In this paper, we study accompany vehicle discovery from traffic monitor plate recognition data. Existing methods are inefficient to query accompany vehicle. To address the problem, we propose a distributed algorithm to find accompany vehicle pair using sliding window algorithm with new prune strategies. We first load traffic monitor plate recognition data into memory by Spark, get candidate vehicle’s upper bound of statistical data to prune invalidate candidates of vehicle and perform verification on these candidates to get final results. We devise some new prune strategies to find high quality candidates. Experiments with real Traffic Monitor Plate Recognition (TMPR) data in a distributed environment verify that our distributed approach can discover accompany vehicles efficiency. We also analyze the performance affecting factors from the experiments.