BeeWay: Elephant Detection and Forest Surveillance 🐘🌲

Overview

BeeWay is a cost-effective, AI-powered solution to mitigate human-elephant conflicts. By combining ESP32-CAM modules, motion sensors, and lightweight AI models trained on Edge Impulse, the system detects elephants in real time, triggers bee sound deterrents, and sends Telegram alerts using ESP-NOW communication.

---

Features 🎯

1. Real-time Detection: Elephant detection using a lightweight AI model on ESP32-CAM.
2. Non-lethal Deterrent: Plays bee sounds via the DF Mini Player to scare elephants.
3. Instant Alerts: Real-time notifications via Telegram with images and timestamps.
4. ESP-NOW Communication: Low-latency data transfer between ESP32-CAM and ESP32.
5. Scalable and Cost-Effective: Easy deployment in remote areas using low-cost hardware.
6. Energy Efficient: Optimized power consumption for long-term deployment.
7. Weather Resistant: Designed for outdoor operation in forest environments.

---

Workflow 🚀

1. Motion Detection

   • PIR motion sensor detects movement and triggers the ESP32-CAM.

2. Elephant Detection

   • The ESP32-CAM captures an image and runs a custom Edge Impulse AI model to confirm if an elephant is detected.

3. Communication via ESP-NOW

   • Detection data is sent to the ESP32 main board for further actions.

4. Sound Deterrent

   • The DF Mini Player plays pre-recorded bee sounds to deter elephants.

5. Real-time Alerts

   • The ESP32 main board triggers the Telegram Bot, sending an image and timestamp to authorized users.

---

Hardware Requirements 🛠️

Component	Description
ESP32-CAM	Captures images and processes AI inference.
ESP32	Acts as the master board for notifications.
PIR Motion Sensor	Detects motion to trigger image capture.
DF Mini Player	Plays bee sounds to deter elephants.
SD Card	Local storage for detected images.
Power Supply	Battery pack or solar power for remote operation.

---

Pin Configuration 🔌

ESP32-CAM Pin Mapping 🛜

Component	Pin Mapping
PIR Sensor	GPIO 14
DF Mini Player	RX: GPIO 16, TX: GPIO 17
SD Card	GPIOs 2, 4, 12, 13, 14, 15

ESP32 Main Board 🛜

Component	Pin Mapping
ESP-NOW	WiFi Channels (Auto Set / Individual WIFI network wth WPA 3 Security)
Telegram Trigger	Configurable and Will be Communicated with the ESP-NOW protocol

---

Software and Tools 🧑‍💻

1. Edge Impulse

   • Train and optimize AI models for elephant detection.
   • Export the model as TFLite for Microcontrollers for deployment.

2. Arduino IDE

   • Develop and upload code for ESP32 and ESP32-CAM boards.

3. Telegram Bot API

   • Sends real-time notifications with images to authorized users.

4. ESP-NOW Protocol

   • Low-latency wireless communication between ESP32 devices.

---

Necessary Arduino Packages 📦

To ensure the BeeWay project works seamlessly on the ESP32-CAM and ESP32, install the following necessary Arduino packages:

1. ESP32 Board Support

   • Package: ESP32 by Espressif Systems
   • Installation:
     • Go to Arduino IDE > File > Preferences.
     • In the Additional Boards Manager URLs field, add:

       https://dl.espressif.com/dl/package_esp32_index.json
       

     • Go to Tools > Boards > Boards Manager, search for ESP32, and install it.

2. Edge Impulse SDK for Arduino

   • Package: Edge Impulse SDK
   • Installation:
     • In the Arduino IDE, go to Sketch > Include Library > Manage Libraries.
     • Search for Edge Impulse and install the Edge Impulse Inference Library.

3. DFMiniMp3

   • Package: DFMiniMp3
   • Installation:
     • In the Arduino IDE, go to Sketch > Include Library > Manage Libraries.
     • Search for DFMiniMp3 and install it to control the DF Mini Player.

4. UniversalTelegramBot

   • Package: UniversalTelegramBot
   • Installation:
     • In the Arduino IDE, go to Sketch > Include Library > Manage Libraries.
     • Search for UniversalTelegramBot and install it to send Telegram alerts.

5. WiFi Library

   • Package: WiFi by Arduino (for ESP32)
   • Installation:
     • This library is typically installed automatically with the ESP32 board package.
     • If not, go to Sketch > Include Library > Manage Libraries, search for WiFi, and install it.

6. ArduinoJson

   • Package: ArduinoJson
   • Installation:
     • In the Arduino IDE, go to Sketch > Include Library > Manage Libraries.
     • Search for ArduinoJson and install it for JSON parsing in Telegram messages.

7. HardwareSerial

   • Package: HardwareSerial
   • Installation:
     • The HardwareSerial library is built into the ESP32 board package and doesn't require separate installation.
     • You can access it in your code by including the following:

       HardwareSerial mySerial(1); // Use UART1 for communication

8. SoftwareSerial

   • Package: SoftwareSerial
   • Installation:
     • This library allows serial communication on other pins if needed, but it's typically not required for ESP32 (which has multiple hardware UARTs). However, if used, you can install it via Sketch > Include Library > Manage Libraries.

Make sure to install these packages to ensure that the BeeWay system works seamlessly with your ESP32-CAM and ESP32 devices. Also, ensure that you correctly configure the hardware serial and software serial as needed for the communication between the ESP32-CAM and other components.

---

Model Integration with Edge Impulse 🎯

Steps to Import and Deploy Edge Impulse Model on ESP32-CAM

1. Train the Model on Edge Impulse

   • Upload elephant image datasets to Edge Impulse.
   • Train and optimize the model using the EON Tuner for microcontrollers.

2. Download the Model

   • Once the training is complete, go to the Deployment section in Edge Impulse.
   • Select "ESP32 - CAM" and download the complete ZIP file (includes TFLite model, inference library, and necessary headers).

3. Import the Model to Arduino IDE

   • Extract the downloaded ZIP file.

   • Copy the folder to the Arduino libraries directory:

     <Arduino>/libraries/edge_impulse_inferencing

4. Include the Model in Your Code

   • Or Just import from the Examples and then select the board ESP32 - CAM
   • In your Arduino code, include the necessary Edge Impulse header files:

   #include "edge-impulse-sdk/classifier/ei_run_classifier.h"

5. Upload the Code to ESP32-CAM

   • Open your Arduino IDE and load the code to the ESP32-CAM board.
   • Verify that the TFLite model runs inference locally and detects elephants as expected.

---

Setup Instructions 🛠️

1. Edge Impulse Model

   • Train and deploy the model using Edge Impulse.

   • Download the model and import the extracted ZIP file into the Arduino libraries folder:

     <Arduino>/libraries/edge_impulse_inferencing
     

2. Hardware Setup

   • Connect the ESP32-CAM, PIR sensor, DF Mini Player, and ESP32 boards as per the pin configuration above.

3. Telegram Bot Setup

   • Create a bot on Telegram using BotFather and get the Bot Token.
   • Replace the Bot Token and Chat ID in the ESP32 code.

4. ESP-NOW Configuration

   • Ensure that the ESP32-CAM and ESP32 main board are on the same WiFi channel for ESP-NOW communication to work.

5. Upload Code

   • Upload the respective .ino files to the ESP32-CAM and ESP32 using the Arduino IDE.
   • Once uploaded, the system will begin detecting elephants, playing deterrent sounds, and sending alerts via Telegram.

---

Future Improvements 🌟

• Extend the detection range using LoRa communication for larger areas.
• Enhance detection accuracy with multi-sensor fusion (e.g., combining thermal imaging or sound-based detection).

---

License 📜

This project is open-source and licensed under the MIT License.

---

BeeWay ensures safer human-wildlife coexistence using cutting-edge technology while promoting conservation. 🌍🐘⚡add

Power Management 🔋

Power Consumption

• ESP32-CAM (Active): ~180mA
• ESP32-CAM (Deep Sleep): ~6mA
• PIR Sensor: ~65μA
• DF Mini Player (Playing): ~20mA
• DF Mini Player (Standby): ~10μA

Power Optimization

1. Deep Sleep Mode

   • ESP32-CAM enters deep sleep when no motion is detected
   • Wake-up triggered by PIR sensor interrupt

2. Solar Power Integration

   • Recommended solar panel: 5V/2A
   • Battery capacity: 3.7V/5000mAh Li-ion
   • Charge controller for battery protection

---

Troubleshooting Guide 🔧

Common Issues and Solutions

1. ESP32-CAM Won't Connect

   • Check WiFi credentials
   • Verify GPIO0 is properly connected for programming
   • Reset board and try again

2. No Telegram Alerts

   • Verify bot token and chat ID
   • Check internet connectivity
   • Ensure proper ESP-NOW communication

3. Poor Detection Accuracy

   • Clean camera lens
   • Adjust camera position
   • Retrain model with more diverse dataset

---