Tea canopy detection algorithm based on multi-sensor data fusion
•Conduct feature analysis on the detection data of tea tree canopies to identify recognition pattern characteristics, leading to the construction of a canopy detection model.•Design an Onion Peeling algorithm for detecting tea tree canopies by fusing multi-sensor data; its convergence is demonstrate...
Saved in:
| Published in: | Smart agricultural technology Vol. 12; p. 101497 |
|---|---|
| Main Authors: | , |
| Format: | Journal Article |
| Language: | English |
| Published: |
Elsevier B.V
01.12.2025
Elsevier |
| Subjects: | |
| ISSN: | 2772-3755, 2772-3755 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | •Conduct feature analysis on the detection data of tea tree canopies to identify recognition pattern characteristics, leading to the construction of a canopy detection model.•Design an Onion Peeling algorithm for detecting tea tree canopies by fusing multi-sensor data; its convergence is demonstrated along with an error analysis.•Apply the algorithm in developing an intelligent shape-adaptive tea-plucking machine equipped with multi high accuracy sensors. The proposed algorithm effectively identifies and eliminates outliers while accurately localizing the position of tea tree canopies within specified precision limits.
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. |
|---|---|
| ISSN: | 2772-3755 2772-3755 |
| DOI: | 10.1016/j.atech.2025.101497 |