Comprehensive Dataset for Urban Streetlight Analysis

TL;DR

This paper presents a large, systematically collected and labeled dataset of over 800 high-resolution streetlight images from India, designed to facilitate deep learning research in urban infrastructure monitoring and smart city applications.

Contribution

It introduces a comprehensive, standardized dataset of streetlight images with binary labels, supporting the development of robust neural network models for urban infrastructure analysis.

Findings

Dataset enables training of effective deep learning models

Supports smart city surveillance and infrastructure monitoring

Facilitates future research in urban computer vision

Abstract

This article includes a comprehensive collection of over 800 high-resolution streetlight images taken systematically from India's major streets, primarily in the Chennai region. The images were methodically collected following standardized methods to assure uniformity and quality. Each image has been labelled and grouped into directories based on binary class labels, which indicate whether each streetlight is functional or not. This organized dataset is intended to make it easier to train and evaluate deep neural networks, allowing for the creation of pre-trained models that have robust feature representations. Such models have several potential uses, such as improving smart city surveillance systems, automating street infrastructure monitoring, and increasing urban management efficiency. The availability of this dataset is intended to inspire future research and development in computer vision and smart city technologies, supporting innovation and practical solutions to urban infrastructure concerns. The dataset can be accessed at https://github.com/Team16Project/Street-Light-Dataset/.