Current irrigation practices depend largely on imprecise applications of water over fields with varying degrees of heterogeneity. In most cases, the amount of water applied over a given field is determined by the amount the most water-stressed part of the field needs. This equates to over-watering most of the field in order to satisfy the needs of one part of the field. This approach not only wastes resources, but can have a detrimental effect on the value of that crop. A system to monitor the needs of each plant or smaller groups of plants, then, would be helpful in catering water application to each plant’s needs. Stem water potential (SWP) has been widely accepted as a good indicator of the level of stress in a plant, and it can be measured fairly precisely, although invasively, using a pressure chamber device. This device, however, is bulky and cumbersome, and it requires an individual to spend 5-10 minutes at a given plant. Using this device to monitor fields of more than several dozen plants, then, becomes time consuming and impractical. Leaf temperature, when adjusted for several environmental conditions, can also give a good indication of SWP, and can be measured proximally as opposed to the invasive nature of the pressure chamber. Recent methods have employed a hand-held sensor to measure leaf temperature and other environmental variables like wind speed, air temperature and humidity. A hand-held sensor would greatly improve the efficiency of measurements compared to the pressure chamber, but it would still be impractical for fields of commercial size, as it still requires manual measurements at each plant. The next step in speed and efficiency would be to employ a remotely monitoring system. Satellite and aircraft imagery is already being used to monitor crop stress on large scales, but these applications typically have resolutions on the scale of acres, and are therefore too imprecise to really monitor individual plants or even small groups of plants. A smaller unmanned aerial vehicle (UAV) could employ the same methods as satellites and larger aircraft, but relatively inexpensively and at a scale catered to the needs of a given field for precision monitoring. A small UAV was retrofitted with an inexpensive infrared (IR) temperature sensor and a multispectral camera. The camera and sensor were installed on an aluminum frame and aligned such that the sensor’s field of view (FOV) covered a particular region in the camera FOV. The UAV was flown above two orchard crops, almonds and walnuts, continually recording both images and temperatures. The camera and sensor were synchronized in such a manner that each image had a temperature assigned to it. The pixel contents of each image were classified in to four different classes: sunlit leaves, shaded leaves, sunlit soil, and shaded soil. In post-processing, a discriminant analysis technique was applied to the images determined to contain mostly vegetation in the IR sensor’s FOV to determine the proportion of each class in the FOV of the IR sensor. Assuming that the measured temperature could be described as a weighted sum of each class, a linear system of equations was set up to solve for the temperature of each class using several aerial measurements of the same tree. Using this technique, the temperature of the vegetation can be solved for and used in determining SWP. Initial results indicate a good correlation between the temperatures estimated from the aerial data using the UAV and the temperature of those classes sampled on the ground immediately following each flight. With leaf temperatures ranging from 15 to 30° C between nine different flights, the proposed technique was able to estimate the temperature of the sunlit and shaded leaves to within several degrees of the sampled temperature in most cases, with an r-squared value of 0.96 for the entire set.