Abstract
Even though there are numerous elements affecting crime, this research targets possible environmental factors using multi-modal techniques combining remote sensing, geospatial analysis, and photogrammetry using drone mapping. First, the correlation between vegetation and crime was examined by retrieving NDVI and Land Surface Temperature (LST) values from NASA’s Landsat-5 TM satellite imagery (30m resolution) on January 31, 2007, and comparing them with a total number of 1,012 burglary cases in the urban areas of Miami-Dade County. Poisson regression analysis revealed a positive correlation between the total number of burglaries and the mean values of NDVI (β=.03, p<.005) and LST (β=.31,p<.001) at the ZIP code level. Next, the relationship between environmental factors and crime was analyzed by using four daytime and four nighttime drone surveys (1.6 cm pixel accuracy), and 7,434 crime incidents in the city of Coral Gables between 2017 and 2022. The kernel density function was utilized to generate heat maps that display the distribution of crime incidents. The findings showed that crime incidents are frequently concentrated around the northeastern part of the city and along US-1. As evidenced from the analysis, greater levels of criminal activities were observed in neighborhoods with higher numbers of renters, newer residences, businesses, public parks, empty parcels, accessibility with public transportation and thoroughfares, and walkability with sidewalks. On the other hand, neighborhoods with greater numbers of streetlights and additional light sources, privately-owned one- or two-story building homes, and bigger parcel areas, reported fewer incidents of crime. In addition, the results revealed that one of the most desirable targets of crime is corner houses. Furthermore, one of the most effective crime prevention methods is restricting visibility and accessibility of targets. Lastly, the outcomes of this study have important implications for police departments, urban planners, and policymakers to develop evidence-based approaches for crime prevention.