Open Access

ARTICLE

Design a Computer Vision Approach to Localize, Detect and Count Rice Seedlings Captured by a UAV-Mounted Camera

Trong Hieu Luu¹, Phan Nguyen Ky Phuc², Quang Hieu Ngo^1,*, Thanh Tam Nguyen³, Huu Cuong Nguyen¹

1 College of Engineering, Can Tho University, Can Tho City, 910900, Vietnam
2 School of Industrial Engineering and Management, International University, Vietnam National University, Ho Chi Minh City, 700000, Vietnam
3 Mekong Delta Development Research Institute, Can Tho University, Can Tho City, 910900, Vietnam

* Corresponding Author: Quang Hieu Ngo. Email:

(This article belongs to the Special Issue: Artificial Intelligence Algorithms and Applications)

Computers, Materials & Continua 2025, 83(3), 5643-5656. https://doi.org/10.32604/cmc.2025.064007

Received 01 February 2025; Accepted 07 April 2025; Issue published 19 May 2025

Abstract

This study presents a drone-based aerial imaging method for automated rice seedling detection and counting in paddy fields. Utilizing a drone equipped with a high-resolution camera, images are captured 14 days post-sowing at a consistent altitude of six meters, employing autonomous flight for uniform data acquisition. The approach effectively addresses the distinct growth patterns of both single and clustered rice seedlings at this early stage. The methodology follows a two-step process: first, the GoogleNet deep learning network identifies the location and center points of rice plants. Then, the U-Net deep learning network performs classification and counting of individual plants and clusters. This combination of deep learning models achieved a 90% accuracy rate in classifying and counting both single and clustered seedlings. To validate the method’s effectiveness, results were compared against traditional manual counting conducted by agricultural experts. The comparison revealed minimal discrepancies, with a variance of only 2–4 clumps per square meter, confirming the reliability of the proposed method. This automated approach offers significant benefits by providing an efficient, accurate, and scalable solution for monitoring seedling growth. It enables farmers to optimize fertilizer and pesticide application, improve resource allocation, and enhance overall crop management, ultimately contributing to increased agricultural productivity.

---

Keywords

---