CURRENT RESEARCH • REMOTE SENSING • GIS • ARTIFICIAL INTELLIGENCE
Drone & Satellite Imagery and Crime
Using remote sensing, geospatial technology and artificial intelligence to measure the built environment and better understand crime and place.
This interdisciplinary project combines high-resolution drone imagery, satellite data, GIS and machine learning to identify environmental characteristics that may influence crime, victimization and community safety.
PROJECT OVERVIEW
Seeing Crime Environments From Above
The physical environment has long been central to theories of crime and place. Buildings, parking areas, vegetation, street layout, property conditions and other elements of urban design can influence visibility, accessibility, guardianship and opportunities for offending.
Drone and satellite imagery make it possible to observe these characteristics systematically and at geographic scales that would be difficult to achieve through traditional field observation alone.
By combining remote sensing, GIS and artificial intelligence, the project transforms imagery into measurable environmental variables that can be examined alongside spatial patterns of crime and community safety.
PROJECT AT A GLANCE
Remote Sensing for Crime Research
RESEARCH FOCUS
Built environment, crime and community safety
PRIMARY DATA
Drone & satellite imagery
CORE METHODS
GIS • Remote sensing • Machine learning
PROJECT STATUS
Active research & technology development
RESEARCH FOCUS
Turning Images Into Environmental Data
The project explores how new geospatial technologies can improve the measurement of physical environments relevant to crime and public safety.
01
Object Detection
Can artificial intelligence reliably identify environmental objects and features from drone and satellite imagery?
02
Built Environment
How do buildings, vegetation, parking areas, infrastructure and other physical characteristics relate to crime?
03
Physical Disorder
Can remote sensing provide scalable and systematic measures of environmental deterioration and disorder?
04
Spatial Crime Analysis
Can imagery-derived environmental measures improve spatial explanation and prediction of crime patterns?
ENVIRONMENTAL MEASUREMENT
What Can We See From Above?
High-resolution imagery can be transformed into measurable environmental variables. Machine learning makes it possible to identify and classify large numbers of physical features more consistently and efficiently than traditional manual observation.
Buildings & Structures
Parking Areas
Trees & Vegetation
Pavement & Surface Conditions
Urban Infrastructure
Signs of Physical Disorder
RESEARCH WORKFLOW
From Imagery to Crime Analysis
The workflow converts raw aerial imagery into geospatial information that can be analyzed alongside crime data.
01
Collect Imagery
Acquire high-resolution drone imagery and appropriate satellite imagery for selected study areas.
02
Process Imagery
Generate orthomosaics, spatial imagery layers and three-dimensional representations of study environments.
03
Apply AI
Use machine learning and object detection to identify and classify physical features within the imagery.
04
Analyze Crime
Integrate imagery-derived variables with crime data to examine environmental relationships and spatial patterns.
TECHNOLOGY & INFRASTRUCTURE
Processing Imagery at Research Scale
A single drone flight can produce hundreds of high-resolution images. Converting those images into orthomosaics, three-dimensional point clouds and machine-learning inputs requires substantial storage, specialized software and accelerated computing.
The project is developing a cloud-enabled geospatial workflow capable of supporting imagery processing, deep-learning analysis and large satellite datasets. The computing environment has been developed with consultation and technical support from Amazon Web Services specialists.
COMPUTING ENVIRONMENT
Built for Large Geospatial Datasets
GPU-enabled computing, GIS software and scalable storage support processing from raw imagery through spatial analysis.
GEOSPATIAL WORKFLOW
From Raw Images to Spatial Data
01
Acquire
Collect high-resolution drone flights and satellite imagery.
02
Process
Generate orthomosaics, imagery products and three-dimensional point clouds.
03
Detect
Apply machine-learning models to identify built-environment features.
04
Analyze
Integrate environmental measures with spatial crime data and models.
GEOSPATIAL COMPUTING ENVIRONMENT
GPU Computing
Accelerated processing for large imagery datasets and deep-learning models.
ArcGIS Pro
Advanced GIS, imagery interpretation and spatial-analysis workflows.
Drone2Map
Drone photogrammetry, orthomosaic generation and three-dimensional products.
Cloud Storage
Scalable storage for high-resolution drone and satellite imagery.
RESEARCH TEAM
Research Associates
The project brings together expertise in criminology, spatial analysis, GIS, remote sensing and geospatial technology.
RESEARCH ASSOCIATE
Dr. Olga B. Semukhina
Professor & Department Head
Department of Criminal Justice
RESEARCH ASSOCIATE
Dr. Yuen “Yolanda” Tsang
Assistant Professor
Geography & GIS
RESEARCH ASSOCIATE
Dr. Opeyemi Zubair
Adjunct Faculty
GIS & Remote Sensing
DRONES • REMOTE SENSING • GIS • CRIMINOLOGY
Interested in Remote-Sensing Crime Research?
We welcome interdisciplinary collaboration involving drone technology, satellite imagery, GIS, remote sensing, machine learning, environmental criminology and spatial crime analysis.