Feature comparison from laser speckle imaging as a novel tool for identifying infections in tomato leaves

Abstract

In the present work we propose the use of laser speckle imaging as non-invasive and marker-free qualitative technique that highlights significant differences between healthy and diseased leaves through image analysis. We observed distinct visual variations, indicating the detectable impact of viral infection on leaf structure. Additionally, leaves tend to homogenize visually over time due to degradation, as evidenced by decreased measured feature vector distances after a week. Statistical analyses, including MANOVA and ANOVA, underscored the significance of parameters like contrast, energy, and variance in distinguishing leaf health states. These findings emphasize the potential of laser speckle imaging for plant health monitoring. Furthermore, our results validate the utility of Apple's Vision library, specifically the VNGenerateImageFeaturePrintRequest() function, as a powerful tool for plant disease diagnosis and health tracking.