Version 1 (modified by 6 days ago) ( diff ) | ,
---|
SASS Project Weekly Progress
The Streetscape Application Services Stack (SASS) project focuses on real-time data collection and synchronization from multiple urban sensors, including Reolink cameras, IMUs, and GoPros. Below is a week-by-week breakdown of progress.
Week 1: Project Setup and Goals
- Defined core objectives of the SASS system: synchronization, modularity, and smart city relevance.
- Introducing ourselves, setting up the initial team structure, and responsibilities.
Week 2: Reading Papers about Multi-modal Fusion
- Discussed with Professor Ortiz about making SASS identify a person in need more discreetly.
- Read additional papers to check the feasibility of the proposed framework.
- Read about multi-modal fusion.
Week 3: System Design and Website Development
- Began developing a website to present a portfolio about the project.
- Set up a System Design Diagram for the data collecting phase.
- Set up all the cameras we are going to use and look into the Sensorlogger App.
Week 4: Devices + IMU Merge Efforts
- Installed Reolink, GoPro cameras, and the SensorLogger App, and basic data collection scripts for each of them.
- Investigated frame-by-frame video analysis from Reolink and GoPro.
- Began efforts to merge IMU and Reolink outputs into a unified logging format.
- Same with the GoPro.
- Integrated IMU sensors with custom Python-based
sensorlogger
script.
Week 5: Synchronization Debugging
- Analyzed misalignments in timestamps between sensor types.
- Attempted to merge all Python files into one. As a result, the Reolink/SensorLogger file and the GoPro/SensorLogger file were used separately.
- System Design Diagram was updated. (As shown in the Google slide).
Week 7: Infrastructure and Database Discussions
- Explored alternatives to flat CSV files; began shifting toward a structured database backend.
- Discussed metadata tagging and retrieval use cases.
- Updated diagrams to reflect router roles, LAN/IP separation, and new sensor locations.
- Started experimenting with YOLOv5 object detection to help define the best camera angles for installations.
Week 8: Unified Logger Development
- Finalize camera alignment with YOLO results.
- Started collecting data.
Next Steps
- Still collecting data.
- Test the unified data logger under full deployment.
Questions or contributions? Please contact us.
Note:
See TracWiki
for help on using the wiki.