Method for on-line machine vision measurement, monitoring and control of product features during on-line manufacturing processes
First Claim
Patent Images
1. A method of monitoring a process producing a characterizing product, the method comprising the steps of:
- sequentially capturing multivariate images of the characterizing product, each image consisting of an image array of pixel elements of measured intensity values in at least three wavelength ranges defining the dimensions for the image array;
creating a feature vector from said image array;
performing a regression method to correlate the feature vector with a characterizing feature of the product; and
,creating an output of said characterizing feature for continuous monitoring of the process.
4 Assignments
0 Petitions
Accused Products
Abstract
A method for extracting feature information from product images using multivariate image analysis based on Principal Component Analysis (PCA) which is used to develop predictive models for feature content and distribution on the imaged product. The imaging system is used to monitor product quality variables in an on-line manufacturing environment. It may also be integrated into a closed-loop feedback control system in automated systems.
-
Citations
43 Claims
-
1. A method of monitoring a process producing a characterizing product, the method comprising the steps of:
-
sequentially capturing multivariate images of the characterizing product, each image consisting of an image array of pixel elements of measured intensity values in at least three wavelength ranges defining the dimensions for the image array; creating a feature vector from said image array; performing a regression method to correlate the feature vector with a characterizing feature of the product; and
,creating an output of said characterizing feature for continuous monitoring of the process. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)
-
-
17. A method of monitoring a process in which a seasoning is added to a food product under variable operating conditions, comprising the steps of:
-
shining a light onto the food product; sequentially capturing multivariate images of the food product in the visible spectrum, each image consisting of an image array having a plurality of pixel elements, each pixel element having intensity values of the colours red, green, and blue defining three dimensions for the image array; applying a multi-way principal component analysis to the image array to reduce the dimensions of the image array to a t1–
t2 score space;creating a feature vector from said t1–
t2 score space;performing a regression method to correlate the feature vector with a coating property on the food product; and
,creating an output of the coating property for continuous monitoring and feed back control of the operating conditions of the process.
-
-
18. A method of monitoring a process in which a seasoning is added to a snack food product under variable operating conditions, the method comprising the steps of:
-
shining a light onto a food product; sequentially capturing multivariate images of the food product in the visible spectrum, each image consisting of an image array having a plurality of pixel elements, each pixel element having intensity values of the colours red, green, and blue defining three dimensions for the image array; applying a multi-way principal component analysis to the image array to reduce the dimensions of the image array to a first t1–
t2 score space;selecting a mask to inscribe an area A in said first t1–
t2 score space which excludes background from the image array;projecting the multivariate images to said first t1–
t2 score space and removing pixels which lie outside said area A to create pure product images;applying a multi-way principal component analysis to the pure product images to reduce the dimensions of the pure product images to a second t1–
t2 score space;creating a feature vector from said second t1–
t2 score space;performing a regression method to correlate the feature vector with a coating property of the food product; and
,creating an output of the coating property for continuous monitoring and feed back control of the operating conditions of the process. - View Dependent Claims (19, 20)
-
-
21. A method of monitoring a process in which a fuel is consumed to produce a flame under changing operating conditions, comprising the following steps of:
-
sequentially capturing multivariate images of the flame in the visible spectrum, each image consisting of an image array having a plurality of pixel elements, each pixel element having intensity values of the colours red, green, and blue defining three dimensions for the image array; applying a multi-way principal component analysis to the image array to reduce the dimensions of the image array to a t1–
t2 score space;selecting a mask to inscribe an area A in said t1–
t2 space corresponding to a flame luminous region;calculating statistical features of said flame luminous region to create a feature vector; performing a regression method to correlate the feature vector and at least one selected process measurements with a characterizing feature of the process; and
,creating an output of said characterizing feature for continuous monitoring of the operating conditions of the process. - View Dependent Claims (22, 23, 24)
-
-
25. A method for product feature monitoring and quality control in an on-line environment comprising the steps of:
-
exposing a food product to an electromagnetic radiation source; imaging at least one product feature of the food product; extracting background features from the product feature image; creating an image score space summarizing information in the product feature image; correlating the product feature image score space to a training sample score space by statistical regression; and
,predicting the occurrence of the product feature on the food product. - View Dependent Claims (26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43)
-
Specification