Python Digital Image Processing (11): Automatic Image threshold segmentation

Image Thresholding Segmentation is a widely used segmentation technique, which uses the difference between the target region and its background in the image, and considers the image as a combination of two kinds of regions (target region and background region) with different gray levels, and chooses a reasonable threshold value. To determine whether each pixel in the image should belong to the target area or the background area, resulting in a corresponding two-value image.

In the Skimage library, the function of threshold segmentation is placed in the filters module.

We can manually specify a threshold value to achieve segmentation. You can also have the system automatically generate a threshold value, the following methods are used to automatically generate thresholds.

1, Threshold_otsu

Threshold Segmentation method based on Otsu, function call format:

skimage.filters. Threshold_otsu (image, nbins=256)

The parameter image refers to a grayscale image, which returns a threshold value.

 fromSkimageImportdata,filtersImportMatplotlib.pyplot as Pltimage=Data.camera () Thresh= Filters.threshold_otsu (image)#returns a threshold valueDST = (Image <= Thresh) *1.0#segmentation based on threshold valuesPlt.figure ('Thresh', Figsize= (8,8)) Plt.subplot (121) Plt.title ('Original Image') plt.imshow (Image,plt.cm.gray) Plt.subplot (122) Plt.title ('binary Image') plt.imshow (Dst,plt.cm.gray) plt.show ()

The return threshold is 87 and is split according to 87:

2, Threshold_yen

Use the same method as above:

The return threshold is 198, divided as:

3, Threshold_li

Use the same method as above:

Thresh = Filters.threshold_li (image)

Returns the threshold value of 64.5, divided as:

4, Threshold_isodata

How to calculate the threshold value:

threshold = (image[image <= threshold].mean () +image[image > Threshold].mean ())/2.0

Use the same method as above:

Thresh = Filters.threshold_isodata (image)

The return threshold value is 87, so the split effect is the same as Threshold_otsu.

5, Threshold_adaptive

The calling function is:

skimage.filters. threshold_adaptive (image, block_size, method= ' Gaussian ')

Block_size: Block size, which refers to the size of the adjacent area of the current pixel, usually an odd number (such as 3,5,7 ... ）

Method: Methods used to determine adaptive thresholds, with ' mean ', ' generic ', ' Gaussian ' and ' median '. Default to Gaussian when omitted

This function accesses the image directly after a threshold, not the threshold value.

 fromSkimageImportdata,filtersImportMatplotlib.pyplot as Pltimage=Data.camera () DST=filters.threshold_adaptive (image, 15)#returns a threshold imagePlt.figure ('Thresh', Figsize= (8,8)) Plt.subplot (121) Plt.title ('Original Image') plt.imshow (Image,plt.cm.gray) Plt.subplot (122) Plt.title ('binary Image') plt.imshow (Dst,plt.cm.gray) plt.show ()

You can change the size and method values of the block_size to see more effects. Such as:

Dst1 =filters.threshold_adaptive (image,31,'mean'=filters.threshold_adaptive (image,5,'median')

These two effects are as follows:

Python Digital Image Processing (11): Automatic Image threshold segmentation