An image dataset of fruitfly species (Bactrocera Zonata and Bactrocera Dorsalis) and automated species classification through object detection

Sana Tariq; Ayesha Hakim; Awais Ahmad Siddiqi; Muhammad Owais

doi:10.1016/j.dib.2022.108366

An image dataset of fruitfly species (Bactrocera Zonata and Bactrocera Dorsalis) and automated species classification through object detection

Data Brief. 2022 Jun 9:43:108366. doi: 10.1016/j.dib.2022.108366. eCollection 2022 Aug.

Authors

Sana Tariq¹, Ayesha Hakim¹, Awais Ahmad Siddiqi², Muhammad Owais¹

Affiliations

¹ Department of Computer Science, Muhammad Nawaz Shareef University of Agriculture (MNSUA), Multan, Pakistan.
² School of Software, Tsinghua University, P.R. China.

Abstract

This data article describes the image dataset collection and annotation of the two most common fruitfly species Bactrocera Zonata and Bactrocera Dorsalis. The dataset is released as a collection of more than 2000 images captured through two sources: images of specially reared fruitfly species in laboratory captured by (48-megapixels) smartphone camera, and images of fruitflies captured by (8-megapixels) Raspberry Pi camera through insect traps installed in fruit orchards. Each image sample is associated with a ground truth label that mentions the fruit fly species. The dataset has been classified and annotated using the object detection method into two fruitfly species with an average 85% accuracy. The results of classification and annotation have been validated by expert entomologists by manually examining test samples in a laboratory setting. This dataset is best suited for developing smart monitoring systems to provide advisory services to farmers through mobile applications that provides real-time information about fruitfly species for effective control and management.

Keywords: Artificial intelligence; Bactrocera Dorsalis; Bactrocera Zonata; Fruitfly; IoT; Microcontroller; Object detection; Smart trap.