A NOVEL METHOD FOR ESTIMATION OF STRUCTURAL CHANGES IN POTATO CROPS FROM UAV-BASED DIGITAL SURFACE MODELS

Abstract

Climate change is arising challenges in potato crop production. More robust clones and cultivars adapted to climate change and resistant to pests and diseases, need to be developed. Several phenotyping approaches are being explored for a quick and affordable characterization of structural changes of crop varieties in real-world conditions. However, conventional methods present several problems. Field work in remote areas is time-consuming or difficult to carry out because of physical barriers. Satellite images lack the temporal resolution needed to monitor small changes in plant structure. Unmanned aerial vehicles (UAVs) provide high-resolution images which can give a better understanding of vegetation changes. Recent UAV-based studies have demonstrated the potential of the Normalized Difference Elevation Index (NDEI) to map centimeter-level changes in topography in different environments. This paper explores the potential of NDEI for identifying structural changes in experimental potato crops. In particular, it proposes an automatic method for estimation of volume changes in potato crop canopy. NDEI values were obtained by multi-temporal analysis of digital elevation models generated from UAV images using structure-from-motion (SfM) and multivision-stereo (MVS) techniques. The main components of the proposed method are: i) Image data capture from UAVs, ii) Data pre-processing, iii) NDEI calculation, and iv) Data processing and analysis. The results suggest that changes in surface elevation values, normalized using the NDEI index, are useful for rapid assessment of volume change in potato crop canopy. The NDEI reveals as an affordable technique to field measurement of phenotypic traits in potato breeding programs.