Measuring the Level of Mildew in Quinoa Plantations Based on Digital Image Processing

Abstract

This paper presents a computer vision system that successfully detects the level of affectation produced by Mildew present in quinoa leaves. The system consists of two subsystems in sequence: processing for the discrimination of earth with vegetation, and processing for the discrimination of vegetation with the affected leaves, which is useful for disease detection. The Lab color method and non-complex image processing techniques are used for both subsystems, such as: histogram equalization, thresholding, filtering, and tagging and connectivity algorithms. This model has been tested in different scenarios, different angles for different crops in the coastal area (where problems with the Mildew are presented with greater emphasis) during different times of the year. In addition, it has been shown that this system produces acceptable results even in difficult conditions, such as in the presence of objects outside the field, either pipes or waste.