There are several challenges in creating an electronic archery scoring system using computer vision techniques. Variability of light, reconstruction of the target from several images, variability of target configuration, and filtering noise were significant challenges during the creation of this scoring system. This paper discusses the approach used to determine where an arrow hits a target, for any possible single or set of targets and provides an algorithm that balances the difficulty of robust arrow detection while retaining the required accuracy.