Analysis of necessary sensor spatial resolution for reliable plant detection

This paper considers theoretical prerequisites for a sensor or sensor system that provides data for reliable single plant detection. The goal of the paper is to close the gap between state of the art sensor technology and the theoretical “ideal” sensor to fulfill one and only purpose: providing single plant detection data in cluttered and disordered environments. Through analysis of artificial leaf forms, the paper determines expectations of the sensor spatial resolution and its maximum noise level, and is supported by the results of a theoretical primer on predefined (artificial) shape analysis. Each shape is analyzed by elliptic Fourier descriptor invariants obtained from the shape boundary, which allow translation, rotation and scale invariant of each object. This paper analyzes the statistical significance of each descriptor and its separability from other features. Finally an approximation formula developed here can be used to compute required resolution of such sensor for the plant detection.