On field disease detection in olive tree with vision systems

This project contains part of the dataset used in this works.

In the present work the capability of convolutional neural networks  to extract samples of leaves in images of tree’s canopy and detect the  presence of different diseases and pests that manifest in deformation,  discoloration or direct presence in the leaves, is studied. The sample  obtained along with its location and sampling date, allows a mapping of  the diseases in the field. This mapping capability will allow better  decisions to be made when fighting these canopy diseases. An example of  those are fungus and Aceria oleae in olive leaves. The study begins with  the analysis of a data set generated in the laboratory and divided into  healthy and faulty parts. The images were captured with a RGB and a  multi-spectral with the blue, green, red, near infrared and red border  spectra. They were taken in an image laboratory with a white background  and led lighting. The objective was to carry out tests to determine the  impact of each spectral channel and the possibility of using different  types of cameras for the detection of diseases, as well as important  factors to consider for its application in the field. Then, Mask rcnn R  50 FPN 3 was used to obtain segmented leaves and Fast-r cnn inception v2  to detect leaves. Then the detected or segmented leaves were classified  with the Inception V3 network to determine which were healthy and which  were diseased. With, the combination of these tools, it is possible to  determine the disease level of an olive tree in the field.