Technical Abstract: A high speed, non-destructive, low cost sorting machine has been developed to separate pistachio product streams. An optical method was used to differentiate pistachio kernels from in-shell nuts, with recognition rates of 97.9% for kernels (2.1% false negatives) and 99.3% for in-shell nuts (0.7% false positives). By adjusting the conditions of the method, these values can be manipulated to favor better recognition of one over the other depending on the more favorably desired stream. Recent research using a commercially available sorting system has reported separation of kernels from in-shell nuts with false positives of 0.1% and false negatives of 1.2 %. While these rates are slightly better than those reported here, the cost of this system is on the order of hundreds of dollars, as compared to commercially available sorting systems, which are in the range of tens of thousands of dollars.