Tea canopy detection algorithm based on multi-sensor data fusion
Year: 2025
Type: article
Abstract: The intelligent tea harvesting machine needs to follow the undulating changes of the tea canopy surface to plucking tea shoots automatically. To address the issue of precise localization of tea canopies, we propose a vertical detection model and algorithm based on multi-sensor information integration. Firstly, the feature analysis on the detection data of tea tree canopies is conducted to identify recognition pattern characteristics, leading to the construction of a canopy detection model. Subsequently, within this framework, an “Onion Peeling” algorithm is designed for detecting tea tree canopies by fusing multi-sensor data. And its convergence is demonstrated along with an error analysis. Finally, we apply this algorithm in developing an intelligent shape-adaptive tea-picking machine. The experiment results indicate that the proposed algorithm effectively identifies and eliminates outliers while accurately localizing the position of tea tree canopies within specified precision limits. The main contributions are as follows: 1) Construct a multi-sensor-based tea canopy position detection system, finding out the recognition pattern: concentrated data distribution area in data range; 2) Proposing the “Onion Peeling” algorithm for feature region detection, which can exclude outliers such as empty sensor readings and anomalous tea shoots, and accurately approximate the target area. Our method and algorithms outperform the-state-of-the-art technologies, providing effective technical support for adaptive harvesting of high-quality tea.
Source: Smart Agricultural Technology
Authors: Yu Han, Zhiyu Song
Language: English
Topic: Smart Agriculture and AI
Subfield: Plant Science
Domain: Life Sciences
SDG: Zero hunger
Open Access status: gold
Estimated APC paid: $1,770
Awards: G3893326267, G8554151646
OpenAlex ID: W4415218191