Research Project

Digital Transportation Infrastructure & AI

Open-source video analytics tools for Caltrans CCTV camera systems, providing benchmarks to assess computer vision solutions and comprehensive procurement specifications for digitalizing transportation infrastructure statewide.

YOLOv8 + ByteTrackCaltrans Agreement 65A1144
0
App Versions Built
0+
Software Modules
0
Assessment Areas
0
Camera Subgroups

Ground Truth Generator

A complete vehicle detection, tracking, and annotation system for generating ground-truth traffic data from CCTV footage. Uses automated AI detection with a human-in-the-loop correction interface for quality assurance.

Live demo

See the detector run in your browser

Real YOLOv8 + ByteTrack output replayed over Caltrans footage — scrub the timeline, adjust the confidence threshold, and watch the live vehicle tally.

Open demo

Native Desktop App (v2.2)

PySide6/Qt standalone app for macOS & Windows with a nine-workspace tabbed interface. v2.2.0 adds hybrid GPU packaging — a universal installer that ships CPU PyTorch and auto-fetches the matching CUDA build (cu128 for Blackwell, cu118 for older GPUs) on first run.

PySide6/QtYOLOv8ByteTrackHybrid GPU9 workspaces

Processing Pipeline

Automated detection and tracking runs in background threads. Supports multiple video formats (MP4, MOV, AVI, MKV) with JSON track data, correction layers, and COCO format export.

JSON TracksCOCO ExportCSV StatsNAS IntegrationTailscale SMB

Software Capabilities

Comprehensive tools for video processing, annotation, and ground-truth generation.

YOLOv8 Detection

Automatic vehicle detection using state-of-the-art object detection

ByteTrack Tracking

Persistent ID assignment across video frames for accurate tracking

Bounding Box Editor

Draw, resize, and adjust detection boxes frame by frame

Track Splitting

Split incorrectly merged tracks into separate vehicle paths

Track Merging

Combine fragmented detections into single continuous tracks

Undo/Redo System

Full history management with Ctrl+Z / Ctrl+Y support

ROI Drawing

Rectangle and polygon regions of interest with per-ROI counting

COCO Export

Export corrected annotations in standard COCO format for training

Performance Assessment Framework

Rigorous evaluation of computer vision technologies for processing Caltrans CCTV video recordings across five critical dimensions.

Vehicle Detection & Classification

Detecting, tracking, and classifying vehicles under varying camera resolutions, framerates, and weather conditions.

95% accuracy in normal conditions, 85% in degraded conditions

Cyclist Detection & Tracking

Assessing capability to detect and track cyclists under different lighting, traffic, and weather conditions.

95% detection along roadways and within crosswalks

Pedestrian Detection & Tracking

Evaluating accuracy of pedestrian detection and tracking across crosswalk scenarios.

95% detection within crosswalks in normal conditions

Processing Speed

Assessing real-time processing capabilities under different capture setups and computing devices.

Minimum 10 frames per second

Data Management

Analyzing input/output data volumes and determining storage, transmission, and cloud requirements.

Optimized storage per camera per day

Data Collection Strategy

Subgroup 1

Ready for Immediate Collection

Cameras that are accessible and ready for data collection without technical barriers.

Subgroup 2

1-2 Month Delay Expected

Cameras with minor delays in retrieving videos or post-processing outputs.

Subgroup 3

2+ Months Required

Cameras requiring extended time to obtain video recordings or post-processing data.

Subgroup 4

Uncertain Retrieval

Cameras of interest where retrieval of recordings or data is uncertain due to technical challenges.

Collection requirements: A minimum of 12 hours of recordings per CCTV camera category, tagged with location, resolution, framerate, camera vendor, weather conditions, lighting conditions, and traffic conditions (peak vs. off-peak).

Procurement Specifications

Measurable specifications to guide Caltrans on procurement and deployment of computer vision platforms for their CCTV network.

Infrastructure Requirements

Hardware, software, power supply, ingress protection, CPU/GPU, memory, storage, and OS specifications

Use Case Identification

Mapping road types to achievable detection accuracy levels and recommended processing approaches

Communication Standards

Broadband, 3G/4G/5G, and data transmission speed requirements for video recording and processing

Explore more research

View our other active projects and initiatives.