🌌 ZED2i Perception & Data Intelligence

Advanced Stereo Vision Pipeline & Synchronized Multi-Camera Acquisition

---

🔭 Project Overview

This repository serves as a professional-grade ecosystem for StereoLabs ZED2i and Intel RealSense D435i integration. It bridges the gap between raw spatial data acquisition and high-level AI perception.

🧩 Core Ecosystem

graph TD
    Root[ZED2i Project Root] --> Stereo[Stereo Depth Intelligence]
    Root --> Data[Data Collection Suite]
    
    Stereo --> MaskRCNN[Mask R-CNN Detection]
    Stereo --> Hungarian[Hungarian Matching]
    Stereo --> Distance[CM-Level Depth]
    
    Data --> Dual[Synchronized ZED/RealSense]
    Data --> IMU[High-Freq IMU Logging]
    Data --> Specialized[Custom Collection Options]

Loading

---

📂 Navigation & Modules

Module	Description	Key Documentation
Perception	Deep learning based depth estimation using stereo pairs.	🧠 Perception README
Data Engine	Synchronized multi-camera recording (ZED + RealSense).	📹 Data Engine README
Specialized	Custom acquisition profiles for specific research.	🛠️ Specialized README

---

🚀 Quick Start

Tip

Ensure you have the ZED SDK and RealSense SDK installed on your system before running the collection scripts.

1. Synchronized Data Collection

# Capture from ZED and RealSense simultaneously
python default/code/Working/script.py --fps 30 --output_dir ./recordings

2. Stereo Depth Finding

# Analyze stereo images for object distances
python stereoImageProcessing/code/StereoCameraFindingDepth/Stereo_Image_All2.py

---

🛠️ Technology Stack

• Computer Vision: OpenCV, StereoLabs SDK (pyzed), Intel RealSense SDK (pyrealsense2)
• Deep Learning: PyTorch, Torchvision (Mask R-CNN ResNet50)
• Data Science: NumPy, Pandas, SciPy (Linear Sum Assignment)
• Visualization: Matplotlib

---

Designed for precision spatial intelligence.