GRID: A Python Package for Aerial High-Throughput Phenotyping
Aerial images have the potential for high throughput phenotyping agricultural experimental fields. Currently, the image analysis relies on either intensive framing manually, or tools requiring stringent structures. We developed a python package, GRID (GReenfield Image Decoder) to overcome these challenges. GRID has a graphical user interface to crop area of interest. Users can see the instant graphic results of tuning parameters such as number clustered objects, degree of noise reduction, and cutoff to balance overall and local structure. With map file, GRID automatically connects plot identifications to the recognized imagery plots. For each plot. the output includes the vertical and horizontal coordinates, the total number of pixels, average NDVI (normalized difference vegetation index based on either near Infrared or regular RGB (red-green-blue) sensors. With GRID, users are completely liberated from manually drawing plot edges, or lines to define the image structure. The GRID executable file, user manual, tutorials, and example datasets are freely available at GitHub