Litcius/Paper detail

Alternative Procedure to Improve the Positioning Accuracy of Orthomosaic Images Acquired with Agisoft Metashape and DJI P4 Multispectral for Crop Growth Observation

Toshihiro Sakamoto, Daisuke Ogawa, Satoko Hiura, Nobusuke Iwasaki

2022Photogrammetric Engineering & Remote Sensing14 citationsDOIOpen Access PDF

Abstract

Vegetation indices (VIs), such as the green chlorophyll index and normalized difference vegetation index, are calculated from visible and near-infrared band images for plant diagnosis in crop breeding and field management. The DJI P4 Multispectral drone combined with the Agisoft Metashape Structure from Motion/Multi View Stereo software is some of the most cost-effective equipment for creating high-resolution orthomosaic VI images. However, the manufacturer's procedure results in remarkable location estimation inaccuracy (average error: 3.27–3.45 cm) and alignment errors between spectral bands (average error: 2.80–2.84 cm). We developed alternative processing procedures to overcome these issues, and we achieved a higher positioning accuracy (average error: 1.32–1.38 cm) and better alignment accuracy between spectral bands (average error: 0.26–0.32 cm). The proposed procedure enables precise VI analysis, especially when using the green chlorophyll index for corn, and may help accelerate the application of remote sensing techniques to agriculture.

Topics & Concepts

Multispectral imageRemote sensingVegetation (pathology)Multispectral pattern recognitionPrecision agricultureSpectral bandsEnvironmental scienceLeaf area indexComputer scienceGeographyAgronomyAgricultureBiologyMedicineArchaeologyPathologyRemote Sensing in AgricultureLeaf Properties and Growth MeasurementRemote Sensing and LiDAR Applications
Alternative Procedure to Improve the Positioning Accuracy of Orthomosaic Images Acquired with Agisoft Metashape and DJI P4 Multispectral for Crop Growth Observation | Litcius