DisPlacing Objects: Improving Dynamic Vehicle Detection via Visual Place Recognition under Adverse Conditions

TL;DR

This paper presents a method that uses visual place recognition and prior maps to enhance dynamic vehicle detection accuracy in adverse conditions without requiring detailed 3D maps or pixel-level correspondences.

Contribution

It introduces a novel algorithm combining visual place recognition and a neural network classifier to refine vehicle detections using prior maps, improving detection performance under challenging conditions.

Findings

Improved vehicle detection accuracy across multiple datasets.

VPR-based approach eliminates need for precise localization.

Neural network classifier effectively refines initial detections.

Abstract

Can knowing where you are assist in perceiving objects in your surroundings, especially under adverse weather and lighting conditions? In this work we investigate whether a prior map can be leveraged to aid in the detection of dynamic objects in a scene without the need for a 3D map or pixel-level map-query correspondences. We contribute an algorithm which refines an initial set of candidate object detections and produces a refined subset of highly accurate detections using a prior map. We begin by using visual place recognition (VPR) to retrieve a reference map image for a given query image, then use a binary classification neural network that compares the query and mapping image regions to validate the query detection. Once our classification network is trained, on approximately 1000 query-map image pairs, it is able to improve the performance of vehicle detection when combined with an existing off-the-shelf vehicle detector. We demonstrate our approach using standard datasets across two cities (Oxford and Zurich) under different settings of train-test separation of map-query traverse pairs. We further emphasize the performance gains of our approach against alternative design choices and show that VPR suffices for the task, eliminating the need for precise ground truth localization.