Monitoring of potato crops based on multispectral image feature extraction with vegetation indices

• Published in: Multidimensional systems and signal processing Vol. 33; no. 2; pp. 683 - 709
• Main Authors: Meivel, S., Maheswari, S.
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.06.2022; Springer Nature B.V
• Subjects: Artificial Intelligence; Circuits and Systems; Crop growth; Damage detection; Electrical Engineering; Engineering; Estimation; Evaluation; Feature extraction; Image resolution; Mathematical analysis; Monitoring; Parameters; Plants (botany); Potatoes; Remote sensing; Signal,Image and Speech Processing; Soil testing; Soils; Vegetation; Vegetation growth; Weather; 3D model; NDVI image; Drone survey; Remote sensing Tif; QGIS; Greenness timing algorithms
• ISSN: 0923-6082; 1573-0824
• DOI: 10.1007/s11045-021-00809-5

The number of individuals working in agriculture is decreasing due to a labor shortage. When switching from manual to automation mode, image mapping must survey soil testing and plant vegetation development. The existing studies met several shortcomings in terms of higher cost remote sensing tools, higher execution time, higher computational complexity, and so on. To tackle these issues, we proposed a multispectral images feature extraction with vegetation indices for potato crops monitoring. This article discusses the many characteristics that are used to identify plants, such as plant count, plant height estimation, plant area evaluation, plant distance, crop vegetation growth detection, damaged area identification, and higher-lower vegetation. Calculating Crop growth days and minimizing leaf or root damage are provided by QGIS and Pix4Dmapper software. It measured all soil testing parameters as well as plant vegetative development. We provided a raster function for aggregation based on a low-resolution image for estimating plant area evaluation and parameters in a growing zone for higher cultivation plants. This research revealed 98% vegetation indices as well as plant characteristics. Various key parameters such as leaf area density, plants coordinates, plant height, the proportion of diffuse light for incident sunlight, and solar zenith angle are examined. The Quality Report of Vegetation Indices Value resulted in an accuracy of 96% when good or terrible weather conditions exist. All Vegetation indices values are compared with other existing technique results in terms of maximum and minimum vegetation to verify its superiority.