<![CDATA[Blast disease (P. oryzae) is one of the main diseases in rice that can cause losses of up to 61% . This study aims to determine the effectiveness of infrared images in early detection of Pyricularia oryzae symptoms in rice seeds with a comparison of conventional methods and to determine the level of accuracy of infrared image analysis on the pathogenic fungus Pyricularia oryzae. This study was conducted from July to December 2025 at the Plant Health Laboratory I, Faculty of Agriculture, UPN "Veteran" East Java. This study used infrared images in detect the pathogen Pyricularia oryzae in rice plants; observations began with Image Acquisition, Image Preprocessing, Segmentation, Visualization, and Validation. The results showed that infection symptoms began to be detected on the 3rd day through pseudo-coloring images before the seeds showed visual symptoms conventionally, because the RGB image only showed color changes on the 4th to 5th day. Validation using regression analysis showed a strong relationship between the results of image estimation and direct observation in the laboratory. The regression model obtained is y = 0.7468x + 0.0845 with an R² value of 0.7033, which means that 70.33% of the variation in observation results can be explained by the image estimation results. The analysis in this study has the potential to be an early detection method because it can detect Pyricularia oryzae fungus on rice seeds with the percentage of symptoms obtained through image processing in the range of 30–52% and is more accurate than manual observation.]]>
