Effective Techniques for the Analysis of Hyperspectral Images to Detect Black Sigatoka Disease Based on Unique Learning Principles

The present study assesses cutting distributed edge intelligence methods with unique learning theories for hyperspectral imaging-based early detection of the black sigatoka disease. We go over the learning characteristics of the techniques used, which will aid researchers in better comprehending the conditions for the necessary data and choosing a strategy that will work for their research needs. With the continuous progress of development, artificial intelligence has made good progress in the analysis and recognition of images, which has also triggered some researchers to explore the area of combining machine learning and deep learning with hyperspectral images and achieve some progress. The current chemical methods for controlling plant diseases have an adverse effect on the environment and raise the cost of production. Creating efficient crop protection strategies requires accurate and early disease detection of the disease. A set of hyperspectral images of banana leaves inoculated with a conidial suspension of black sigatoka fungus (Pseudocercospora fijiensis) was used to train and validate machine learning models. Support vector machine (SVM), multilayer perceptron (MLP), neural networks, N-way partial least square–discriminant analysis (NPLS-DA), and partial least square–penalized logistic regression (PLS-PLR) were selected due to their high predictive power. When the spectral signatures of the misclassified leaves were compared with the average spectra of the healthy and diseased leaves, the similarities and differences that explained their erroneous classification could be observed. The models were assessed using the metrics of AUC, precision, sensitivity, prediction, and F1 score. The experimental outcomes demonstrate that the PLS-PLR, SVM, and MLP models enable the successful and highly reliable early detection of black sigatoka disease, positioning them as robust and highly reliable HSI classification methods for the early detection of plant disease and allowing for the evaluation of chemical and biological control of phytopathogens.