Analysis of Models and Algorithms for Measuring the Area of Vine Leaves

Abstract

The area of the vine leaves is an important indicator for determining the quantity of leaf mass, making connections with the influence of the environment, improving the methods of growing the vineyards. Satellite and aviation measurements for now have the drawback that the images obtained are of low resolution and do not allow the measurement of the area of individual leaf. A solution to this problem is the use of unmanned aerial vehicles, which provide digital images of a small height (1-3m) and autonomous robots to navigate in the vineyards. These systems use video cameras operating in the visible, near infrared range and thermo cameras. The measurement of individual leaves and the search for links to the foliage and plant indicators is due to the fact that in order to do this when crawling the array, it is necessary to use low energy consumption devices. These devices also have poor computing resources. In this report a comparative analysis of 16 models describing the relationship between the area and the main dimensions of the leaf - long and short axis is made. Three of these models have been selected to describe this relationship with sufficient precision. They are compared with the 4 algorithm for measuring the area of the vine leaves. The results obtained show that the measurement error, the data processing time between the algorithms used and the models are comparable. The analyzes made suggest that the choice of a method for measuring the area of vine leaves depends on the desired accuracy, the time of receipt, the processing and the analysis of the results of what equipment the user has access to.