Live With Paws — AI Wildlife Intrusion Detection System

Live With Paws is an AI-powered wildlife intrusion detection system designed to detect and log animal intrusions in real-time, minimizing human-wildlife conflict using computer vision and backend alert APIs.

⚡ System Architecture Flow

sequenceDiagram
    autonumber
    participant Camera as Camera Stream
    participant CV as OpenCV Ingestion
    participant YOLO as YOLOv8 Inference Engine
    participant Pandas as Pandas Analytics Log
    participant Django as Django Backend API
    participant User as Alert Notification

    Camera->>CV: Capture Raw Video Frame
    CV->>YOLO: Preprocess & Matrix Transform (NumPy)
    YOLO-->>CV: Detection Bounding Boxes & Confidence
    alt Intrusion Detected (Confidence > Threshold)
        CV->>Pandas: Log Intrusion Event & Timestamp
        CV->>Django: POST /api/v1/alerts (Intrusion Payload)
        Django-->>User: Trigger Automated Alert Notification
    end

⚡ Core Architecture & Engineering Highlights

• Real-Time Vision Inference Pipeline: Designed and deployed a frame-by-frame image ingestion pipeline using YOLOv8 and OpenCV, optimizing array and matrix transformations via NumPy.
• Data Filtering & Behavior Analysis: Built custom data normalization and cleaning logic using Pandas to aggregate and analyze historical wildlife intrusion frequency and temporal patterns.
• Model Evaluation & Metric Tuning: Evaluated detection accuracy across Precision, Recall, and mAP metrics, tuning confidence thresholds to minimize false positive triggers in noisy outdoor environments. See related note: [[Machine Learning Model Evaluation Metrics: Precision, Recall & mAP]].
• Backend Alert Engine: Integrated the ML inference pipeline with a Django web application, building structured REST APIs to deliver automated alert notifications.

🛠️ Technology Stack

| Layer | Technologies Used | | :--- | :--- | | Computer Vision Engine | YOLOv8, OpenCV, NumPy | | Data Normalization & Logs | Pandas, SQL | | Backend & APIs | Django, REST APIs | | Language & Environment | Python 3.x, Linux |

🔗 Related Notes & Links

• Related Study Note: [[Data Visualization with Matplotlib & Pandas: Fundamentals & Practice]]
• Author: Nischal Neupane (B.Sc. CSIT, Vedas College / Tribhuvan University)
• GitHub: github.com/nischalneupanee