SusanCornersDetector: Each constructor results in different SupportedFormats

[justingruenberg]

The parameter-less constructor adds in Format8bppIndexed, Format24bppRgb, Format32bppRgb, Format32bppArgb to the SupportedFormats array, which the other constructor does not. Finding corners from an object created with the second constructor will throw an exception on 8bpp images (at least).

framework/Sources/Accord.Imaging/AForge.Imaging/SusanCornersDetector.cs

Lines 143 to 149 in a5a2ea8

	public SusanCornersDetector()
	{
	base.SupportedFormats.UnionWith(new[] { PixelFormat.Format8bppIndexed,
	PixelFormat.Format24bppRgb,
	PixelFormat.Format32bppRgb,
	PixelFormat.Format32bppArgb });
	}

Inside the InnerProcess the image gets converted to 8bpp if it isn't already:

framework/Sources/Accord.Imaging/AForge.Imaging/SusanCornersDetector.cs

Lines 178 to 186 in a5a2ea8

	if (image.PixelFormat == PixelFormat.Format8bppIndexed)
	{
	grayImage = image;
	}
	else
	{
	// create temporary grayscale image
	grayImage = Grayscale.CommonAlgorithms.BT709.Apply(image);
	}

The second constructor should call the parameterless constructor.