saisanjay1716/SmartFence-Wildlife-Dataset

SmartFence AI Dataset

Overview

The SmartFence AI Dataset is a custom computer vision dataset developed for an intelligent wildlife protection and monitoring system.

It is designed to support object detection models that identify humans, domestic animals, and wild animals. It can be used to develop AI-powered wildlife monitoring systems, animal detection applications, and human-wildlife conflict prevention solutions.

Dataset Information

Attribute	Details
Dataset Name	SmartFence AI Dataset
Task	Object Detection
Framework	YOLOv8
Number of Classes	14
Total Images	21,814
Annotation Format	YOLO
Dataset Splits	Training, Validation, Testing

Dataset Splits

Split	Images	Labels
Training	16,694	16,694
Validation	3,786	3,786
Testing	1,334	1,334
Total	21,814	21,814

Classes

The dataset contains 14 classes:

ID	Class
0	Humans
1	Cows
2	Goats
3	Dogs
4	Cats
5	Elephant
6	Deer
7	Bear
8	Gaur
9	Wild Boar
10	Tiger
11	Leopard
12	No Animal or Human
13	Chicken

Dataset Structure

SmartFence-Wildlife-Dataset/
│
├── README.md
├── dataset.yaml
│
├── TRAIN/
│   ├── images/
│   │   ├── part_001/
│   │   └── part_002/
│   └── labels/
│       ├── part_001/
│       └── part_002/
│
├── VAL/
│   ├── images/
│   │   └── part_001/
│   └── labels/
│       └── part_001/
│
├── TEST/
│   ├── images/
│   │   └── part_001/
│   └── labels/
│       └── part_001/
│
├── evaluation/
└── tested images/

The dataset is divided into three subsets:

• TRAIN: Used for model training.
• VAL: Used to evaluate model performance during training.
• TEST: Used for final model evaluation.

The training images are divided into two parts to accommodate repository file-count restrictions. This does not change the original images or annotations.

Annotation Format

The dataset uses YOLO object detection annotation format.

Each image has a corresponding .txt annotation file.

Each annotation follows this structure:

class_id x_center y_center width height

The bounding box coordinates are normalized between 0 and 1.

Dataset Configuration

The dataset.yaml file defines the dataset paths and class names.

The configuration uses the reorganized directory structure:

path: .
train:
  - TRAIN/images/part_001
  - TRAIN/images/part_002
val: VAL/images/part_001
test: TEST/images/part_001

nc: 14
names:
  0: humans
  1: cows
  2: goats
  3: dogs
  4: cats
  5: elephant
  6: deer
  7: bear
  8: gaur
  9: wild boar
  10: tiger
  11: leopard
  12: no animal or human
  13: chicken

Intended Applications

• Wildlife detection and monitoring.
• Human intrusion detection.
• Forest and agricultural area surveillance.
• Animal movement monitoring.
• AI-based wildlife protection systems.

Usage

The dataset can be used with Ultralytics YOLO models.

Install the required package:

pip install ultralytics

Download the dataset and use its dataset.yaml configuration file.

Example training command:

from ultralytics import YOLO

model = YOLO("yolov8n.pt")

model.train(
    data="dataset.yaml",
    epochs=100,
    imgsz=640,
    batch=16
)

License and Attribution

Dataset source: To be specified.

License: To be specified.

Please check the original dataset sources and their licensing and attribution requirements before redistributing or using this dataset.

Project

SmartFence AI: Intelligent Wildlife Protection System

This dataset supports the development of computer vision solutions for wildlife protection and intelligent surveillance.