DETERMINING SUITABLE SPATIAL RESOLUTION TO ESTIMATE NITROGEN STATUS IN MD2 PINEAPPLE CROP CULTIVATED ON MINERAL SOIL

• Published in: Journal of Theoretical and Applied Information Technology
• Main Author: Hasni R.; Mohidin H.; Khan M.Y.M.A.; Narawi A.; Banchit A.; Tamrin K.F.; Jack R.; Jos S.; Salam N.D.
• Format: Article
• Language: English
• Published: Little Lion Scientific 2022
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85125367900&partnerID=40&md5=7f40820e2bd160dd0a533be659d8e866

Strengthening of Malaysian food security requires optimum utilisation of agricultural technology to sustainably increase productivity and yield. Digital nutrient monitoring enables more efficient and timely field estimation to complement existing conventional method. However, high UAV acquisition and computational costs can be overwhelming especially when periodical monitoring is involved. This study attempted to improve UAV feasibility by identifying the suitable spatial resolution (SR) to estimate Nitrogen (N) status in MD2 pineapple (Ananas comosus var. MD2) crop on mineral soil. Two field plots, respectively representing Alluvial and Red-Yellow Podzolic (RYP) soils, were built in Samarahan Campus Farm of Universiti Teknologi MARA Sarawak, Malaysia. This Randomised Complete Block Design (RCBD) based experiment was comprised of five treatments, ten replicates, and five different combinations of NPK fertiliser and MD2 pineapple leaf biochar application. N status of crop canopy was sampled using non-destructive and destructive methods; respectively involving DJI Phantom 4 Multispectral UAV, SPAD-502 Chlorophyll Meter, and D-leaf extraction. Scores of four vegetation indices (NRI, VARI, GCI and RECI) representing Predicted N, were regressed against Observed N of D-leaf Total N Content. SPAD Chlorophyll Meter provided Predicted Relative N status. This study compared the capability of SR between 0.47 and 4.01 cm to detect crop canopy and support Predicted-Observed N Status regression. Detection capability in this study corresponded with SR, yet not solely with canopy width. The highest resolutions of SR0.75 (Alluvial) and SR0.47 (RYP) were able to detect all sample crop canopies, and yield the highest Predicted-Observed N correlation based on NRI and VARI estimations. Detectability was largely influenced by canopy width, number of leaves, and crop symmetries. Lower SR estimations were affected by deteriorating pixel purity and biased sample representation. Therefore, SR of below 1.0 cm is recommended for MD2 Pineapple crop N status estimation on mineral soil. © 2022 Little Lion Scientific