Mobile application for the detection of black Sigatoka

Black Sigatoka is one of the main problems that affect the quality and production of the banana crop, it´s because of this, the development of systems to detect diseases, generate an important tool for the monitoring and control carried out by the farmer. The proposed system leverages hardware on mobile devices to implement computer vision techniques to determine the percentage of affected area of the plant. The smartphone is used to acquire data and capture the disease through images. The detection of diseased pixels is then performed through a segmentation algorithm with histogram analysis. A model for the calculation of the affected area is then computed. Finally, the information is presented through the user interface. To validate the proposed method, a database is created with images taken by the application to compare it´s efficiency through the RMS error between manual segmentation and the result of the algorithm. Finally, usability and response time tests are performed.