Can Urban Blight Be Accessed with Vision-language Models: A Case Study in Detroit

2026-08-03

AI Summary

This study addresses the high cost and prolonged duration of traditional residential decay assessments, which hinder timely urban renewal and public health interventions. It proposes a novel approach that integrates multi-view street-level imagery with open-source large vision-language models, leveraging structured prompt engineering to evaluate critical housing attributes—such as roof integrity, wall damage, and window/door deterioration—and outputs both binary judgments and probabilistic decay estimates. The work introduces an innovative XGBoost-based ensemble stacking framework coupled with a weighted scoring mechanism to effectively fuse multi-perspective information, substantially enhancing model accuracy, robustness, and generalizability. Experimental results demonstrate consistent superiority over individual base models across diverse residential settings, enabling low-cost, scalable, and dynamic monitoring of housing conditions.

Abstract

Addressing urban blight has seen increased focus in the past 15 years. Assessing urban blight is essential for guiding urban planning, targeting rehabilitation, and safeguarding public health, yet traditional residential blight surveys are difficult to maintain at scale due to the labor-intensive cost and long-term cycle. This study introduced a scalable framework for estimating residential blight using open-source large vision-language models on multiple views. Structured prompts guided models to evaluate housing attributes, including roof integrity, wall damage, and broken or boarded openings, producing both binary assessments and probabilistic estimates of disrepair. To evaluate the performance of these visual assessments, we compared professional human annotations of these features across several models, including an ensemble stacking approach based on XGBoost and a weighted scoring system. Results showed that (i) multiple street views can contribute to the improvement of accuracy, (ii) large vision-language models have different strengths of inference, (iii) the ensemble learner outperforms individual base models, enhancing robustness across all residential conditions and blight assessment. The practical application of the method allows low-cost tracking and management of housing stock conditions, providing a regularly updatable complement to traditional blight surveys.

Problem

• urban blight
• residential decay
• housing condition assessment
• scalable survey
• vision-language models

Innovation

• vision-language models
• urban blight assessment
• structured prompting
• ensemble learning
• multi-view street imagery