In order to decrease water consumption in agricultural sectors, deficit irrigation has been suggested. For preventing a heavy loss in crop yield, management of water stress is essential. Plant based methods for measuring plant water stress such as stomatal conductance are more precise than soil based methods such as soil water potential or content, but all these methods are largely time-consuming and require specialized labor and a large number of measurements. Surface temperature of a leaf or canopy can be an indicator of stomatal conductance, hence of a plant water stress. However, canopy temperature is not only affected by stomatal conductance, but also changes with environmental conditions such as air temperature and vapor pressure deficit. Crop water stress index (CWSI) and stomatal conductance index (Ig) are recently proposed to solve these dependency problems. Comparing the canopy temperature with two reference temperatures, crop water stress index can minimize the effect of environmental conditions. Thermography is a non-destructive and non-contact approach toward surface temperature measurement without interfering with plant activities. The temperature of specific options in thermal image could be obtained via supervised or unsupervised methods. The general goal of this study was to obtain CWSI using thermal images and to evaluate the potential of this method for predicting the stomatal conductance as a reference measurement of crop water stress and to automate this procedure via unsupervised computer processing. In order to achieve this goal, thermal images from olive trees under five deficit irrigation treatments with three replications were captured. Wet and dry references surfaces were developed and placed in the field of view of the camera nearby canopy. Images were capture in two directions, first perpendicular to solar travel and second back to the sun shining direction which receives more radiation than first one, in three times in a day (10, 12, 14 regional times). The automation procedure was a mixture of pixel and region based image processing applied to thermal images. The automation algorithm were coded in MATLAB and compared with supervised methods. Inputs of algorithm were thermal images and the code delivered stress indicators. Indices assisted with air temperature plus five degrees as dry reference temperature, could significantly improve the coefficient of correlation. Results revealed more powerful relationship between stress indicators calculated in first direction at morning measurements (R 2 ADJ = 0.92***) rather than other times and directions. Furthermore, Ig showed higher coefficient of correlation with stomatal conductance (R 2 ADJ = 0.94***). Automatic algorithm obtained canopy temperature, significantly similar to supervised method (R 2 ADJ = 0.99***), with even more precise relationship between stress indices and the stomatal conductance (R 2 ADJ = 0.95***). Keywords: Thermography, Canopy Temperature, Stomatal Conductance, Crop Water Stress Index, Stomatal Conductance Index, Image Processing, Fuzzy C-Means Clustering