当前位置:网站首页>Halcon 20.11: how to deal with the quality problem of calibration assistant

Halcon 20.11: how to deal with the quality problem of calibration assistant

2021-08-08 00:41:56 Machine vision 001

HALCON 20.11: How to deal with the quality problem of calibration assistant

stay “ The short time is three or five days ” Under Options , You can get an evaluation of each image , It includes defect image feature description and quality score percentage of problem severity . The result is 0% It indicates that the image features have great defects , and 100% It indicates that the image features have ideal quality . This can help you by deleting images with poor quality that may lead to a high error rate in the calibration process , To improve your calibration results . If you need a quality level , You can set “ Warning level ”, Defects will be listed in “ The short time is three or five days ” Next .

adopt “ Image test ” and “ Sequence test ” To detect quality problems . If you want the program to run faster , Or don't need quality feedback , Can be “ Image test ” and “ Sequence test ” Change to Quick( Perform fewer tests ) or None( No tests are performed at all ). If the defect is too serious , If no calibration mark or even calibration plate is found , The calibration button will turn gray , Unless the poor quality image is deleted from the list , Otherwise, calibration cannot be performed .

If subsequent measurements are always made on exactly the same plane , The test results on the tilt of the calibration plate may be ignored . under these circumstances , Focal length and Z Values are not individually correct , But just in their combination . The reason for this is that neither of these values itself can be determined . This leads to the result : For example, you get twice as much focus as you should ,Z The value is half the original , vice versa . Besides , The farther you place an object on the plane you calibrate , The less accurate the result is .

Be careful : Poor image quality will lead to poor calibration results , This leads to poor or incorrect measured values . However , Even if the mass fraction warning is 40% To 70% Between , Generally, acceptable results can be achieved . If necessary , Please see the following table , To get suggestions for improving image quality . When trying to improve image quality , Don't forget to check other error sources .

Quality problems

explain

Possible solutions

The calibration plate is overexposed

The calibration plate area is too bright , Cause edge offset , Center position calculation error

Turn down the lens aperture or reduce the camera exposure time or turn down the illumination brightness of the light source , Until sufficient quality is achieved

Uneven image illumination

The brightness of the calibration plate changes , It is difficult to locate the calibration plate , Less accurate

Image unevenness is often used horizontally / Results of longitudinal illumination ; Consider using diffuse lighting

The contrast is too low

The difference between the gray value of the calibration plate and the gray value of the mark of the calibration plate is not large enough

Overexposure or underexposure . In order to improve the effect , You can change the aperture or illumination brightness .

Mark diameter is too small

The diameter of the calibration plate mark found is too small

The setting should be changed or a calibration plate with a larger mark should be used

Mark is not in focus

The marker is not fully focused , Some look fuzzy , Low robustness

The depth of field must include the entire object ; Or change the focal length ; Or change the distance from the object to the camera ; Or reduce the aperture , Or use brighter lighting

Quality assessment failed

Image test failed , Even if the calibration plate can be found in the image

For rectangular marker calibration plates , Check whether any part of the image is obscured , And whether the occlusion will interrupt the black edge of the calibration plate

Tag extraction of some images failed

Calibration plate marks in some images cannot be extracted , Calibration will not be possible

Delete the image with failed tag extraction ; Replace or adjust the extraction parameters with a new image

Image quality problem detected

The image quality is below the warning level

Check the quality problem by clicking on a single image name in the list

Too few images

The number of images is lower than the recommended number

Check whether enough images have been taken according to the type of calibration plate

The calibration board image failed to cover the field of view

Some parts of the field of view are not covered by any image of the calibration plate , That is, there are areas that are not marked

Press the display button ( The display button appears in a field named Details In the column of ), All red areas not covered by the calibration plate image can be seen . Before continuing the calibration , Add missing images to the sequence

The tilt angle is not included in the sequence

The calibration plate is not tilted enough

Add more images of the calibration plate tilted in different directions ;

For rectangular alignment calibration plates , It is recommended to tilt the image twice in each quadrant , And change the tilt direction

Not all images are the same size

The image list contains images of different sizes

You changed the setting while taking the calibration image . You should delete and change those images taken before , To get useful results .

版权声明
本文为[Machine vision 001]所创,转载请带上原文链接,感谢
https://chowdera.com/2021/08/20210808004039806U.html

随机推荐