Put Your Title Here Using Capital and Lowercase Letters

Abstract

A Thermal-Reflectance StressImaging (TRIS) system was used to diagnose water stress in a Sunagoke moss sample. The samples thermal images were analyzed to give its surface temperatures. Its visible reflectance images were analyzed to give standard and probability-distance weighted textural features and color information features. A Crop Water Stress Index (CWSI) was used to indicate the samples water status. The texture and color features were used to diagnose the samples water stress using neural networks (NN). The sample exhibited water related stress at water contents below 1.5gg-1 (grams of water per unit mass of sample) and above 3.0gg-1. A clear correlation was established between the samples water status and its canopy temperature, GLCM texture and color information features. Probability-distance re-weighting of textural features and use of color information features increased water stress diagnostic power of the NN models. Since the setup used in this study was able to detect both flooding and drought water stress in a plant which exhibits a high degree of desiccation tolerance, the TRIS system is robust enough for real-time water stress diagnosis, analysis and monitoring in plants. In addition, use of probability-distance re-weighted texture and color information features, improves stress diagnosis and detection in plants.