Find the error in the image

Hoang,

I checked the image that you sent, and noticed that you want to detect the missing parts in the "star" shape on clothes. Is this correct?
If yes, the details depend on whether you just want to find issues with "start" shapes on cloths or you want to detect issues with any shape.

In general, I think you need first to define a reference image that contains the correct shape with the expected dimensions and colors, so that you can compare with.
Also you might need to follow a standard when taking pictures for the star shape so that the location, size and dimensions values of the shape in all captured images are very close together.

After this, you can try using LEADTOOLS Image processing functions to load each captured image (that contains a star shape) and convert it to pure black and white using the following C# code (I am attaching the result image):

using (Leadtools.Codecs.RasterCodecs _codecs = new Leadtools.Codecs.RasterCodecs())
{
using (Leadtools.RasterImage _image = _codecs.Load(@"..\..\BFolder\B.png"))
{
//This step is to make the backgraound as white as possbile
Leadtools.ImageProcessing.Color.ChangeContrastCommand _ChangeContrast = new Leadtools.ImageProcessing.Color.ChangeContrastCommand();
_ChangeContrast.Contrast = 1000;
_ChangeContrast.Run(_image);
//Convert the image to pure black and white
Leadtools.ImageProcessing.Color.IntensityDetectCommand _intensityDetect = new Leadtools.ImageProcessing.Color.IntensityDetectCommand(255,
255,
Leadtools.RasterColor.Black,
Leadtools.RasterColor.White,
Leadtools.ImageProcessing.Color.IntensityDetectCommandFlags.Master);
_intensityDetect.Run(_image);
//Convert the image to 1-bit black and white
Leadtools.ImageProcessing.ColorResolutionCommand _colorRes = new Leadtools.ImageProcessing.ColorResolutionCommand(Leadtools.ImageProcessing.ColorResolutionCommandMode.InPlace, 24,
Leadtools.RasterByteOrder.Bgr | Leadtools.RasterByteOrder.Gray,
Leadtools.RasterDitheringMethod.None,
Leadtools.ImageProcessing.ColorResolutionCommandPaletteFlags.Fixed,
null);
_colorRes.Run(_image);
_codecs.Save(_image, @"..\..\BFolder\TestBImage.png", Leadtools.RasterImageFormat.Png, 1);
}
}

And then you can either read the image as segments (you can read it line by line for example) and compare the number of black pixels in each line (or segment) with the predefined reference image (or numbers) .
This way you might be able to find any missing parts or issues in images.
You can find more details here:

https://www.leadtools.com/sdk/image-processing/functions

Note that I didn't try this suggestion myself, but I think it's worth a try.

2 Replies

Know the answer? Post it — somebody with the same question will find it here.

Guest User7y ago
Accepted answer
Your file doesn't download, please post your code here.
Maen Badwan7y ago
Hoang,

I checked the image that you sent, and noticed that you want to detect the missing parts in the "star" shape on clothes. Is this correct?
If yes, the details depend on whether you just want to find issues with "start" shapes on cloths or you want to detect issues with any shape.

In general, I think you need first to define a reference image that contains the correct shape with the expected dimensions and colors, so that you can compare with.
Also you might need to follow a standard when taking pictures for the star shape so that the location, size and dimensions values of the shape in all captured images are very close together.

After this, you can try using LEADTOOLS Image processing functions to load each captured image (that contains a star shape) and convert it to pure black and white using the following C# code (I am attaching the result image):
using (Leadtools.Codecs.RasterCodecs _codecs = new Leadtools.Codecs.RasterCodecs())
{
using (Leadtools.RasterImage _image = _codecs.Load(@"..\..\BFolder\B.png"))
{
//This step is to make the backgraound as white as possbile
Leadtools.ImageProcessing.Color.ChangeContrastCommand _ChangeContrast = new Leadtools.ImageProcessing.Color.ChangeContrastCommand();
_ChangeContrast.Contrast = 1000;
_ChangeContrast.Run(_image);

//Convert the image to pure black and white
Leadtools.ImageProcessing.Color.IntensityDetectCommand _intensityDetect = new Leadtools.ImageProcessing.Color.IntensityDetectCommand(255,
255,
Leadtools.RasterColor.Black,
Leadtools.RasterColor.White,
Leadtools.ImageProcessing.Color.IntensityDetectCommandFlags.Master);

_intensityDetect.Run(_image);


//Convert the image to 1-bit black and white
Leadtools.ImageProcessing.ColorResolutionCommand _colorRes = new Leadtools.ImageProcessing.ColorResolutionCommand(Leadtools.ImageProcessing.ColorResolutionCommandMode.InPlace,                                                                                                                                  24,
Leadtools.RasterByteOrder.Bgr | Leadtools.RasterByteOrder.Gray,
Leadtools.RasterDitheringMethod.None,
Leadtools.ImageProcessing.ColorResolutionCommandPaletteFlags.Fixed,
null);

_colorRes.Run(_image);

_codecs.Save(_image, @"..\..\BFolder\TestBImage.png", Leadtools.RasterImageFormat.Png, 1);
}
}
And then you can either read the image as segments (you can read it line by line for example) and compare the number of black pixels in each line (or segment) with the predefined reference image (or numbers) .
This way you might be able to find any missing parts or issues in images.
You can find more details here:
https://www.leadtools.com/sdk/image-processing/functions

Note that I didn't try this suggestion myself, but I think it's worth a try.