Urban Object Detection Kit: A System for Collection and Analysis of Street-Level Imagery

Abstract

In this paper, we propose Urban Object Detection Kit, a system for the real-time collection and analysis of street-level imagery. The system is affordable and portable and allows local government agencies to receive actionable intelligence about the objects on the streets. This system can be attached to service vehicles, such as garbage trucks, parking scanners and maintenance cars, thus allowing for large-scale deployment. This will, in turn, result in street-level imagery captured at a high collection frequency, while covering a large geographical region. Unlike more traditional panoramic street-level imagery, the data collected by this system has a higher frequency, making it suitable for the highly dynamic nature of city streets. For example, the proposed system allows for real-time detection of urban objects and potential issues that require the attention of city services. It paves the way for easy deployment and testing of multimedia information retrieval algorithms in a dynamic realworld setting. We showcase the usefulness of object detection for identifying issues in public spaces that occur within a limited time span. Finally, we make the kit, as well as the data collected using it, openly available for the research community. CCS CONCEPTS• Computing methodologies → Visual inspection; Mobile agents; Distributed computing methodologies; • Information systems → Multimedia and multimodal retrieval.

StreetAware: A High-Resolution Synchronized Multimodal Urban Scene Dataset

“…Charitidis et al released a paper in 2023 [ 27 ] in which they utilized several state-of-the-art computer vision approaches, including Cascade R-CNN [ 28 ] and RetinaFace [ 29 ] architectures for object detection, the ByteTrack method [ 30 ] for object tracking, DNET architecture [ 31 ] for depth estimation, and DeepLabv3+ architecture [ 32 ] for semantic segmentation to detect and geotag urban features from visual data. Object detection systems have also been specifically developed for the collection and analysis of street-level imagery in real-time [ 33 ]. In “Smart City Intersections: Intelligence Nodes for Future Metropolises” [ 1 ], Kostec et al detail intersections as intelligence nodes using high-bandwidth, low-latency services for monitoring pedestrians and cloud-connected vehicles in real-time.…”

Zurich Like New

“…Santani et al [8] studied Ma3Route, a mobile social media channel to crowdsource traffic reports in the city of Nairobi. Sukel et al [19] used over 500K citizen reports from the city of Amsterdam and implemented a classification task of urban micro-events, e.g., loud boat, [20]. In this paper, we add to this literature by following a similar methodology and contribute a full multimodal analysis of ZüriWieNeu with data spanning over seven years, using both descriptive analyses and machine learning applied on contextual cues, text, and images.…”

PlastOPol: A Collaborative Data-driven Solution for Marine Litter Detection and Monitoring

“…The goal is to reduce the volunteers' personnel demands [6] and personnel fatigue in clean-up and monitoring activities. More recently, several solutions for recognizing litter were developed using mobile devices [17], [18]. These contributions used cameras to collect images, while the classification was performed on a server and the results were sent back to the users.…”