Citdet: a benchmark dataset for citrus fruit detection

Authors: JORDAN, JAMES - University Of Texas At Arlington; MANCHING, HEATHER - North Carolina State University; Mattia, Matthew; Bowman, Kim; Hulse-Kemp, Amanda; BEKSI, WILLIAM - University Of Texas At Arlington

Submitted to: International of Electrical and Electronics Engineers (IEEE) Robotics and Automation Letters
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/11/2024
Publication Date: 10/3/2024
Citation: Jordan, J.A., Manching, H.K., Mattia, M.R., Bowman, K.D., Hulse-Kemp, A.M., Beksi, W.J. 2024. Citdet: a benchmark dataset for citrus fruit detection. International of Electrical and Electronics Engineers (IEEE) Robotics and Automation Letters. https://doi.org/10.1109/LRA.2024.3474473.
DOI: https://doi.org/10.1109/LRA.2024.3474473

Interpretive Summary: Tools for determining characteristics of plants in an automated fashion are lacking in many important agricultural systems. In this study, we have developed a tool to enable detecting citrus fruit both on the tree and on the ground. This tool allowed us to accurately capture observable yield or the amount of fruit observed by a trained human using front and back images of a tree. With this tool available, it may be possible for humans to look at yield across a much larger number of trees. Additionally, the images along with annotations or where the fruit are located on the image have been made publicly available. These images and annotations will allow other researchers in the future to use these in the process of developing new machine learning and artificial intelligence tools to benefit agriculture.

Technical Abstract: In this letter, we present a new dataset to advance the state of the art in detecting citrus fruit and accurately estimate yield on trees affected by the Huanglongbing (HLB) disease in orchard environments via imaging. Despite the fact that significant progress has been made in solving the fruit detection problem, the lack of publicly available datasets has complicated direct comparison of results. For instance, citrus detection has long been of interest in the agricultural research community, yet there is an absence of work, particularly involving public datasets of citrus affected by HLB. To address this issue, we enhance state-of-the-art fruit detection methods for use in typical orchard settings. Concretely, we provide high resolution images of citrus trees located in an area known to be highly affected by HLB, along with high-quality bounding box annotations of citrus fruit. Fruit on both the trees and the ground are labeled to allow for identification of fruit location, which contributes to advancements in yield estimation and potential measure of HLB impact via fruit drop. The dataset consists of over 30,000 bounding box annotations for fruit instances contained in 600 high-resolution images. In summary, our contributions are the following: (i) we introduce a novel dataset along with baseline performance benchmarks on multiple contemporary object detection algorithms, (ii) we show the ability to accurately capture fruit location on tree or on ground, and finally (ii) we present a correlation of our results with yield estimations.