by
Phillip Clancy, Next Instruments International, Sydney, Australia
Plant breeders, seed producers and grain traders use subjective inspection methods, i.e., visual inspection, separation and counting to determine quality aspects of grains and oil seeds
Depending on the sample and the inspection standards, subjective inspection can take 15 minutes per sample or more.
On the other hand, machine vision inspection offers a rapid and more reproducible method of inspecting grains and oil seeds at a fraction of the cost of manual inspection.
Subjective vs Objective Measurements
The human eye and brain have an amazing ability to pick between two objects and decide differences in size, colour, defects and other physical parameters, however the brain cannot retain the exact image and provide a quantitative evaluation of these parameters. Not to mention, humans have different perceptions of colour and defects.
An Image Analyzer does not have the differentiation capability of a human, however an Image Analyzer can quantify the parameters, i.e., assign numbers to the parameter, which can be stored and compared to a set of standards for that parameter.
Subjective measurement is how humans measure physical parameters, where as an Image Analyzer makes an objective measurement. Another aspect of Image Analysis is that the lighting is kept constant from machine to machine where as lighting used in subjective measurements can vary.
A fluorescent lamp will show an image with more blue hues than red, as compared to a halogen lamp. Daylight is another light source but it changes depending on the time of the year, time of the day and whether it is overcast or sunny.
The image that is collected from the Image Analyzer provides a permanent record of the sample that was analyzed. A subjective measurement of a sample on the other hand has no record other than the count that was made by the inspector.
If the same image is re-analyzed by the SeedCount software, the results will be same. If a sample that has been subjectively assessed by a human is given to another human, then the results will probably be different.
There is one major problem with Image Analysis in that it is never going to be the same as the subjective measurement. The human eye can look at a seed, which has several coloured sections, and compare it with a printed standard and a subjective measurement can be made. It may not be 100 percent correct, however the eye and brain can make an assessment.
An Image Analyzer often cannot differentiate between subtle colour differences or complex shapes etc. As such, the decision to change from subjective measurements to objective measurements should not be based on the objective measurement producing the same results as the subjective measurement, which will change from person to person, but rather based on the ability of the machine to reproducibly assess samples no matter who is performing the test.
Where the machine cannot make the assessment due to the differences being too subtle or too complex, then the machine should be used to make the measurements that it can do, well and to combine it with a user option to classify the seed shown on the screen using a pull down options menu.
An operator selects a seed from the screen image and a table shows the machine assigned parameters. By clicking on a parameter, the operator can change the assessment. This hybrid system at least provides a permanent record of the assessment that can be emailed to the buyer or end user to validate the analysis.
Visit the NEXT Instruments website, HERE.
Read the full article, HERE.
Plant breeders, seed producers and grain traders use subjective inspection methods, i.e., visual inspection, separation and counting to determine quality aspects of grains and oil seeds
Depending on the sample and the inspection standards, subjective inspection can take 15 minutes per sample or more.
On the other hand, machine vision inspection offers a rapid and more reproducible method of inspecting grains and oil seeds at a fraction of the cost of manual inspection.
Subjective vs Objective Measurements
The human eye and brain have an amazing ability to pick between two objects and decide differences in size, colour, defects and other physical parameters, however the brain cannot retain the exact image and provide a quantitative evaluation of these parameters. Not to mention, humans have different perceptions of colour and defects.
An Image Analyzer does not have the differentiation capability of a human, however an Image Analyzer can quantify the parameters, i.e., assign numbers to the parameter, which can be stored and compared to a set of standards for that parameter.
Subjective measurement is how humans measure physical parameters, where as an Image Analyzer makes an objective measurement. Another aspect of Image Analysis is that the lighting is kept constant from machine to machine where as lighting used in subjective measurements can vary.
A fluorescent lamp will show an image with more blue hues than red, as compared to a halogen lamp. Daylight is another light source but it changes depending on the time of the year, time of the day and whether it is overcast or sunny.
The image that is collected from the Image Analyzer provides a permanent record of the sample that was analyzed. A subjective measurement of a sample on the other hand has no record other than the count that was made by the inspector.
If the same image is re-analyzed by the SeedCount software, the results will be same. If a sample that has been subjectively assessed by a human is given to another human, then the results will probably be different.
There is one major problem with Image Analysis in that it is never going to be the same as the subjective measurement. The human eye can look at a seed, which has several coloured sections, and compare it with a printed standard and a subjective measurement can be made. It may not be 100 percent correct, however the eye and brain can make an assessment.
An Image Analyzer often cannot differentiate between subtle colour differences or complex shapes etc. As such, the decision to change from subjective measurements to objective measurements should not be based on the objective measurement producing the same results as the subjective measurement, which will change from person to person, but rather based on the ability of the machine to reproducibly assess samples no matter who is performing the test.
Where the machine cannot make the assessment due to the differences being too subtle or too complex, then the machine should be used to make the measurements that it can do, well and to combine it with a user option to classify the seed shown on the screen using a pull down options menu.
An operator selects a seed from the screen image and a table shows the machine assigned parameters. By clicking on a parameter, the operator can change the assessment. This hybrid system at least provides a permanent record of the assessment that can be emailed to the buyer or end user to validate the analysis.
Visit the NEXT Instruments website, HERE.
Read the full article, HERE.
The Global Miller
This blog is maintained by The Global Miller staff and is supported by the magazine Milling and Grain
which is published by Perendale Publishers Limited.
For additional daily news from milling around the world: global-milling.com
No comments:
Post a Comment