Indoor Plant Varieties for Computer Vision Applications: A Diverse Image Dataset

Published: 4 November 2024| Version 1 | DOI: 10.17632/ct79299k27.1
Contributors:
,
,
,
,
,

Description

The Indoor Plant Varieties Dataset is a valuable asset for researchers, plant care specialists, and AI developers, offering essential tools for the accurate identification and classification of common indoor plant species. This dataset supports the development of computer vision models tailored to indoor plant recognition, enabling users to create applications that simplify plant identification and care. Collected between October 28 and November 20, 2023, the dataset consists of 1172 high-resolution images (3000x3000 pixels, JPG format) across seven plant classes: Aglaonema, Cryptanthus, Devil’s Ivy, Heartleaf Philodendron, N-Joy Pothos, Rhaphidophora, and ZZ Plant. Each class represents unique indoor plant species frequently found in homes and offices, collected from BADC (Bangladesh Agricultural Development Corporation) in Kashimpur and nearby nurseries. Images were captured using a Xiaomi M2101K61 smartphone, ensuring consistency in quality. Domain experts (agronomists from BADC) confirmed the class labels, adding reliability to the dataset. To enhance the dataset further, Data Augmentation techniques were applied, including flipping, rotation, zoom, shear, brightness adjustments, and noise reduction. This augmentation expanded the dataset to 7000 images (1000 per class), enhancing the effectiveness of deep learning models for accurate plant classification. 1. Original Data: Number of datasets: 1172 Data format: .jpg 2. Augmented Data: Number of datasets: 7000 Data format: .jpg

Files

Institutions

Daffodil International University, City University

Categories

Computer Vision, Image Processing, Feature Extraction, Deep Learning, Agriculture

Licence