Approach for Indirect Measurement of Chlorophyll and Phenophases of Maize Plant Using Image Processing

An approach for Indirect Measurement of Chlorophyll and Phenophases of Maize Plant Using Image Processing is proposed in the paper. The proposed approach for indirect measurement of chlorophyll of maize plants and two main phases of plants: sweep and the presence of 11 leaves could be applied to a whole field or a part of it. Four color spaces are used in the research - RGB, HSV, Lab and XYZ. By applying the plant separation indices, it is possible to remove from the image pixels, that are not the object of the study. This allows only maize plants to be analyzed. Three informative color characteristics were found to distinguish the two classes of maize plants - a(Lab), b(Lab) and H(HSV). The color features are the first three informative features using the ReliefF algorithm. The accuracy for class 1 (sweeping phase) is $86,7 \%$ and for class 2 (phase 11 leaves) – 93,3%. Using ground measurements of the cholorphyll in plants for class 1 (sweeping phase) the values are 1,975 – 2,761 and for class 2 (phase 11 leaves): 1,449 – 1,842. Experimental verification of the accuracy of the proposed approach will be evaluated by increasing the number of field measurements and photographs.