| dc.description | The extraction of characteristics, currently, plays an important role, likewise, it is considered a complex task, allowing to obtain essential descriptors of the processed images, differentiating particular characteristics between different classes, even when they share similarity with each other, guaranteeing the delivery of information not redundant to classification algorithms. In this research, a system for the recogntion of diseases and pests in tomato plant leaves has been implemented. For this reason, a methodology represented in three modules has been developed: segmentation, feature extraction and classification; as a first instance, the images are entered into the system, which were obtained from the Plantvillage free environment dataset; subsequently, two segmentation techniques, Otsu and PCA, have been used, testing the effectiveness of each one; likewise, feature extraction has been applied to the dataset, obtaining texture descriptors with the Haralick and LBP algorithm, and chromatic descriptors through the Hu moments, Fourier descriptors, discrete cosine transform DCT and Gabor characteristics; finally, classification algorithms such as: SVM, Backpropagation, Naive Bayes, KNN and Random Forests, were tested with the characteristics obtained from the previous stages, in addition, showing the performance of each one of them. | |
