The durability and effectiveness of protective coatings and surface treatments are heavily reliant on optimal surface preparation and providing a “clean” substrate. Despite this clear correlation, the established standards for waterjetting and particle-blasting of surfaces rely on visual interpretation and comparison with reference photos. Using new machine learning techniques coupled with advances in artificial intelligence, the depth of knowledge and experience required to accurately judge the classification of a water-jetted substrate can be greatly reduced. Determining through automation whether a surface needs to be cleaned and what standard of cleaning a job has reached will greatly reduce time requirements and will allow for the standardization of surface preparation across different operators. This work describes the development and testing of a standalone image processing technology that has been trained to accurately detect and determine the condition and quality of a substrate after waterjetting and application of a novel, anti-fouling surface treatment. The processing algorithm was trained on a variety of different SSPC/NACE standards and can provide near real-time feedback as to whether a selected standard has been met, and which areas may require further cleaning to meet a given standard.