ACHENY : A Standard Chenopodiaceae Image dataset for Deep Learning Models

Published: 2 August 2021| Version 1 | DOI: 10.17632/fpfty8nn7j.1


This dataset includes 27,030 images of 30 species of Chenopod from their natural habitats and is called ACHENY (Autumn Chenopod of Yazd). From each class, different numbers of images (300-1461) of the entire plants and branches were captured in the natural habitat of the plant. In summary, out of 27,030 ACHENY images, 2,703 images were dropped for testing. 24,327 remaining images, in each folding, 4,867 images were used for validation and 19,460 images were used for training. The number of images from each class used for testing, validation, and training, was 10%, 18%, and 72% respectively. Images were collected from multiple Chenopod bushes, at different camera-to-target distances and from different viewpoints and angles. The test specimens were selected from bushes distinct from the training specimens. The reason is to ensure, as far as possible, that the train and test samples are distinct and do not share common images. The Chenopodiaceae plant is important vegetation around the world. Chenopodiaceae has various species that are ecologically and financially important. These species play a significant role in biodiversity. Biodiversity protection is very critical for the survival and sustainability of each ecosystem. Plants species recognition in their natural habitats is the first process in plants diversity protection. Automatic plant recognition by using image processing and computer vision can reduce the cost and time effectively. Modern computer vision approaches are based on deep learning techniques. In order to perform a deep learning algorithm, the existence of a standard dataset is very essential. Experimental Design, Materials, and Methods 1. Camera specification and setting Imaging is performed using two different cameras: a) Nikon COOLPIX S2800 digital camera with a 1: 1 (14.9-megapixel) 3864-by-3864 resolution. b) Samsung SM-J701F mobile with a 1: 1 (3.7-megapixel) 1920-by-1920 resolution. Both cameras were utilized for image collection in natural light during days. 2. Imaging time and conditions Studied Chenopodiaceae species often have flowers and fruits in the autumn, hence imaging is performed in November and December in their habitat. Outdoors and nature have many uncontrollable factors affecting images, such as light intensity throughout the day, wind blowing, cloudy skies or sunshine, atmospheric precipitation, foggy air, and so on. Imaging was performed at different times of sunny, cloudy and windy days in natural sunlight. Some other factors also affect acquired images, such as camera-to-target distances, viewing angles, location of light sources, and so on.



Islamic Azad University


Biodiversity, Image Database, Deep Learning