Pharmacy People Counting & Tracking System

Computer vision-powered customer flow analysis for retail pharmacies

Leer en Espanol

---

Overview

Master's Thesis (TFM) project: a real-time system for counting and tracking people inside pharmacies using deep learning. It detects customers with YOLO, tracks them with DeepSORT, analyzes dwell time per zone, and generates heatmaps — all with built-in GDPR-compliant privacy (real-time pixelation, local processing only).

---

Key Features

Detection & Counting YOLOv5 person detection with virtual counting lines Entry/exit tracking with real-time totals Configurable confidence thresholds JSON export of all counting data	Tracking & Analytics YOLOv8 + DeepSORT persistent tracking Polygonal zone definition with dwell time Automatic heatmap generation Per-person trajectory and zone history
Privacy & Compliance Real-time face/body pixelation (toggle with P) 100% local processing — no external data transfer GDPR compliant by design	User Interface Desktop GUI built with CustomTkinter Interactive line and zone configuration tools Integrated video preview and controls One-click processing start/stop

---

Gallery

Main Interface

Control panel	Processing & real-time controls

Configuration Tools

Virtual counting lines	Polygonal zone definition

Counting & Tracking

Real-time people counting	Privacy pixelation enabled

DeepSORT trajectory tracking	Generated heatmap

---

Architecture

graph TD
    A[GUI - CustomTkinter] -->|Tools Tab| B[Configuration]
    B --> B1[Define Virtual Lines]
    B --> B2[Define Polygonal Zones]
    A -->|Processing Tab| C[Detection & Analysis]
    C --> C1[YOLOv5 Counting]
    C --> C2[YOLOv8 + DeepSORT Tracking]
    C1 --> D[Export]
    C2 --> D
    D --> D1[datos_conteo/ JSON]
    D --> D2[datos_tracking/ JSON]
    D --> D3[heatmaps/ PNG]

Loading

---

Tech Stack

---

Quick Start

Prerequisites

• Python 3.8+
• pip

Installation

git clone https://github.com/patatapython/Sistema-Tracking-Farmacia.git
cd Sistema-Tracking-Farmacia
python -m venv venv
source venv/bin/activate  # Linux/Mac
# venv\Scripts\activate   # Windows
pip install -r requirements.txt

Download YOLO Models

# YOLOv5s (~15MB)
wget https://github.com/ultralytics/yolov5/releases/download/v7.0/yolov5s.pt
# YOLOv8s (~22MB)
wget https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov8s.pt

Run

python uiFarmacia_logo.py

---

Usage

1. Configure (Tools tab): define counting lines and polygonal zones interactively
2. Process (Processing tab): select video source, start counting or tracking
3. Analyze: results are automatically exported to datos_conteo/, datos_tracking/, and heatmaps/

Output data formats

Counting (datos_conteo/*.json):

{
  "entradas": 15,
  "salidas": 12,
  "total_personas": 27,
  "personas_dentro": 3,
  "timestamp": "2025-09-05T14:42:45.598Z"
}

Tracking (datos_tracking/*.json):

{
  "timestamp": "20250905_144534",
  "total_personas": 5,
  "sistema_tracking": "DeepSORT",
  "personas": [
    {
      "id": 1,
      "zona_actual": "mostrador",
      "tiempo_en_zona_actual": 45.2,
      "historial_zonas": [
        {
          "zona": "entrada",
          "tiempo_entrada": "2025-09-05T14:40:00",
          "tiempo_salida": "2025-09-05T14:40:30",
          "duracion": 30.0
        }
      ]
    }
  ]
}

---

Project Structure

Sistema-Tracking-Farmacia/
├── uiFarmacia_logo.py      Main GUI application
├── conteo.py               YOLOv5 people counting module
├── tracking.py             YOLOv8 + DeepSORT tracking module
├── crear_linea.py          Interactive counting line creator
├── crear_zonas.py          Interactive polygonal zone creator
├── config/                 Line and zone configuration files
├── datos_conteo/           Counting results (JSON)
├── datos_tracking/         Tracking results (JSON)
├── heatmaps/               Generated heatmap images (PNG)
├── extras/Imagenes/        Screenshots for documentation
├── Logo/                   Application logo
└── requirements.txt        Python dependencies

---

Credits

Developed as a Master's Thesis (TFM) by Guillermo — patatapython

Special thanks to:

• Ultralytics for YOLO models
• OpenCV for computer vision tools

---

License

Distributed under the MIT License. See LICENSE for details.

---

_{Built with Python and open-source computer vision}