Julia Gersey receives ACM SenSys 2025 Best Poster Award for work on sensing city environments

Mena Davidson • June 2, 2025
Gersey is integrating computer vision-based infrastructure analysis with mobile air quality monitoring to explore urban conditions that can affect human health.
A woman gestures to an academic poster while explaining it to a man, both standing in an outdoor courtyard.
Julia Gersey presents her award-winning poster at ACM SenSys 2025 in Irvine, California. Photo courtesy of Julia Gersey

Julia Gersey, PhD student in Electrical and Computer Engineering (ECE), was honored with a Best Poster Award at the Association for Computing Machinery (ACM) Conference on Embedded Networked Sensor Systems (SenSys). Her presentation, “Sniffing Out the City – Vehicular Multimodal Sensing for Environmental and Infrastructure Analysis,” describes a new method to improve upon manual inspections of street conditions, building maintenance, pollution hazards, and more.

“We’re developing a way for a vehicle to ‘sense its surroundings’ in urban environments,” said Gersey. “In particular, we are trying to help homeless populations by exploring and improving the distribution of resources.”

The research team is using a combination of computer vision and environmental sensing to gain insights into issues like excess pollutants, trash buildup, and heat islands in cities. Computer vision can be used to categorize and analyze infrastructure like buildings, sidewalks, and green spaces. Sensing of environmental conditions like ambient temperature, humidity, volatile compounds, carbon dioxide levels, and air quality index can provide more specific information about exposures that affect human health—especially for those lacking a fixed, regular, and adequate place to spend the night.

A diverse group of seven people poses for a conference. The three students in the photo hold award certificates.
Left to right: Mani Srivastava (2025 SenSys General Co-Chair), Fatima Anwar (2025 SenSys Poster Co-Chair), Julia Gersey, Jiale Zhang, Pei Zhang, Guoliang Xing (2025 SenSys General Co-Chair), and Jesse Codling. Photo courtesy of Julia Gersey

“We collect visual data through a camera and the environmental sensors allow us to ‘see’ beyond the camera’s field of view—but, on the flip side, the visual field can actually add more context to the environmental side. For example, when the sensor is reading a high value we can visually find out why,” explained Gersey.

In the presented work, Gersey designed a custom 3D-printed case for a series of sensors mounted inside a car. An air pump draws outdoor air through a hose extended out the window and into the 3D-printed enclosure, allowing reliable sampling while shielding the sensors from weather and airflow disturbances. The team used a pre-trained DeepLabv3+ computer vision model on the publicly available RoadBotics dataset of Detroit streets, and they tested the environmental sensors by driving through downtown Ann Arbor.

Two images side by side. Left: a photo of a Detroit city street with buildings and cars. Right: an abstract color representation of the same photo, with different colors to show buildings, cars, sidewalks, and other landmarks.
An example video frame showing a Detroit street segmented into classes for roads, buildings, cars, people, sidewalks, and street signs. Image from “Poster Abstract: Sniffing Out the City – Vehicular Multimodal Sensing for Environmental and Infrastructure Analysis”

The poster and accompanying conference paper demonstrate proof of concept for use of a multimodal sensing method to replace the slow and labor-intensive manual standards for assessing urban infrastructure. Gersey plans to continue improving the modeling and implementation for the multimodal sensing system, eventually deploying the units in San Jose, in conjunction with Associate Professor Hae Young Noh at Stanford and some homeless shelters in the area.

Gersey coauthored her poster with her PhD advisor Prof. Pei Zhang, labmates Jiale Zhang and Jesse Codling, and Stanford University doctoral student Jatin Aggarwal.

Gersey has also been awarded the Department of Energy Computational Science Graduate Fellowship to support this work. Prior to joining ECE as a graduate student in August 2024, Gersey earned her bachelor’s degree in Computer Science & Applied Mathematics at Baldwin Wallace University.