Location: Application Technology ResearchTitle: Development of a LiDAR-guided section-based tree canopy density measurement system for precision spray applications
|MAHMUD, MD SULTAN - Pennsylvania State University|
|ZAHID, AZLAN - Pennsylvania State University|
|HE, LONG - Pennsylvania State University|
|CHOI, DAEUN - Pennsylvania State University|
|KRAWCZYK, GRZEGORZ - Pennsylvania State University|
|HEINEMANN, PAUL - Pennsylvania State University|
Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/11/2021
Publication Date: 2/20/2021
Citation: Mahmud, M., Zahid, A., He, L., Choi, D., Krawczyk, G., Zhu, H., Heinemann, P. 2021. Development of a LiDAR-guided section-based tree canopy density measurement system for precision spray applications. Computers and Electronics in Agriculture. 182. Article 106053. https://doi.org/10.1016/j.compag.2021.106053.
Interpretive Summary: Accurate measurement of apple tree canopy foliage density can help growers better manage production practices with advanced technologies in orchards. Despite great efforts have been made to characterize fruit tree canopies, challenges still exist to accurately determine leaf densities at different positions inside canopy due to irregular and nonuniform canopy architecture. In this research, apple tree canopy densities were assessed using a ground-based LiDAR guided sensing system. Point cloud data were acquired from two apple varieties in two orchards and were processed with a specially designed algorithm. Tree trunk, trellis wires, and support poles were extracted to separate the canopy points from the acquired data. Prediction models of apple tree leaves were developed for the two varieties with different sizes and shapes. The canopy volume of individual apple trees was measured using an alpha shape algorithm applying different values. Apple leaves automatically estimated by the developed algorithm were compared with manual counts. Results reported a strong correlation between manually counted leaves and acquired point cloud data from the LiDAR sensor. Additionally, the canopy density maps from the point cloud data could identify the high, moderate, and low density of leaves as well as no leaf regions within the apple trees. The canopy density information will be used to guide precision orchard management strategy development which include precision operations of spraying pesticides pruning trees and estimating yields. For spray operations, it will be integrated into intelligent sprayers to adjust both liquid and air volumes based on the appearance of the canopy characteristics in each tree section in real time.
Technical Abstract: An unmanned ground-based canopy density measurement system to support precision spraying in apple orchards was developed to precisely apply pesticides to orchard canopies. The automated measurement system was comprised of a light detection and ranging (LiDAR) sensor, an interface box for data transmission, and a laptop computer. A data processing and analysis algorithm was also developed to measure point cloud indices from the LiDAR sensor to describe the distribution of tree canopy density within four sections according to the position of the trellis wires. Experiments were conducted in two orchard sites, one with GoldRush (larger trees) and the other one with Fuji (smaller trees) apple trees. Tree leaves were counted manually from each section separated by trellis wires. Field evaluation results showed a strong correlation of 0.95 (R2 = 89.30%) between point cloud data and number of leaves for the Fuji block and a correlation of 0.82 (R2 = 67.16%) was obtained for the GoldRush block. A strong correlation of 0.98 (R2 = 95.90%) was achieved in the relationship between canopy volume and number of leaves. Finally, a canopy density map was generated to provide a graphical view of the tree canopy density in different sections. Since accurate canopy density information was computed, it is anticipated that the developed prototype system can guide the sprayer unit for reducing excessive pesticide use in orchards.