Automatic Dishware Inspection: Applications and Comparisons of Two New Methods

Commercial dishwashing systems currently involve human loading, sorting, inspecting, and unloading dishes and silverware pieces before and after washing in hot and humid environments. In such difficult working conditions, leading to high turn-over of low-paid employees, automation is desirable, especially in large-scale kitchens of hospitals, navy ships, schools, hotels and other dining facilities. Our project is a part of developing an integrated machine vision sorting and inspecting system for mixed dish pieces and silverware exiting a flight-type commercial dishwashing machine, coupled with automatic loading and unloading. We propose two new methods for automatically inspecting dish cleanliness, namely adaptive thresholding and maximum saliency map. On the first method, a new technique using partitioning and adaptive thresholding, combined with global thresholding are introduced. On the second method, we propose a new normalization technique. Both algorithms are fast, simple, and produce results invariant with lighting conditions and dish rotation about the camera-dish axis. Algorithms are written and tested by MatlabÒ R14 and Image Processing Toolbox V5.0 to 110 dish images taken in different lighting condition using different position of 51 separate dishes (either clean or dirty) of our dish set, in which 77 images are from dirty dishes with 799 dirty points in these dishes. The adaptive thresholding method produces 95.0% and 96.5% accuracies in discriminating clean from dirty dishes and dirty spot detection, respectively. While the maximum saliency map method produces 100% accuracies in discriminating clean from dirty dishes and 93.5% accuracies in and dirty spot detection.