Maintenance of Machine Vision Systems for Product Quality Assessment. Part I. Addressing Changes in Lighting Conditions

Abstract

In the last decades, a growing number of image analysis applications in the process and food industries have been reported. This is because artificial vision systems can return a quick, accurate, and objective indication of the quality of the manufactured end product. However, reproducibility of the image analysis results is ensured only as long as the conditions, under which the images used for the calibration of the quality assessment model were collected, do not change during normal operation of the manufacturing process. These conditions include the status of the artificial vision illuminating system and of the camera sensor. In this paper, we specifically deal with the issues related to the aging or failure of the illuminating system. First, a strategy is developed to monitor the status of the machine vision system, in order to detect any changes in the lighting conditions. Then, two alternative strategies are proposed to avoid recalibrating the quality assessment model in case a new lighting condition is detected. The first strategy aims at determining a new combination of camera settings to compensate for the change in the light illuminating the subject, whereas the second one adapts the available model to the novel conditions. The effectiveness of the proposed strategies is tested on a pharmaceutical manufacturing case study involving quality assessment on film-coated tablets.

Automated detection of parasitized Cadra cautella eggs by Trichogramma bourarachae using machine vision

“…While human inspectors can visually inspect parts to judge the quality of workmanship, machine vision systems use advanced hardware and software components to perform these functions at higher speeds and with improved reliably and precision. Image analysis techniques have been applied in the process and food industries for several decades [18] and are being used more frequently in agriculture to discriminate between objects [19][20][21] and to count items [21,22] with different visual characteristics.…”

Automated Systems Based on Machine Vision for Inspecting Citrus Fruits from the Field to Postharvest—a Review

“…Hence, it is important to take this into account to design algorithms that are robust against failures or defects of these elements (Ottavian et al 2013(Ottavian et al & 2014. Illumination is one of the conditioning factors that most seriously affect performance.…”

Statistical process control based on Multivariate Image Analysis: A new proposal for monitoring and defect detection

“…Different ways for dealing with this problem are treated by Reis (2009), and very recently by Ottavian, Barolo andGarcía-Muñoz (2013a, 2013b). This paper presents and develops a novel Multivariate Image Analysis technique, and deeply analyzes its potential in comparison with a competing approach, in terms of control charts performance and statistical properties; trying to understand the differences between them.…”