Build Proof of Concept Computer Vision Application focused on particular model

Need to create a small application that can detect a particular model in an image. The model is generally in the same place at all times. Complication seems minimal. Long term goal, would like to use machine learning to increase accuracy.