One challenge in the development of structural health monitoring technology is the necessity to distinguish benign effects from those caused by damage. For ultrasonic guided waves systems, this is a problem of particular importance. Guided waves create complex, multi-modal, and dispersive wave fields which reflect off specimen boundaries as well as damage. Direct time-domain comparisons with a known baseline can be used to overcome these complexities, but fail to discriminate damage from benign environmental effects. Although many environmental changes affect guided waves, variations in temperature are often the most dominant. This paper proposes a computationally efficient temperature compensation technique based on the scale-invariant correlation coefficient. Using experimental measurements, we compare the performance of the scale-invariant correlation coefficient with two other compensation strategies: the local peak coherence and optimal signal stretch methods. We demonstrate the scale-invariant correlation coefficient to be robust, effective, and computational efficient.