Digital Image Processing

Digital Image Processing

​Histograms

​

​A histogram shows the number of pixels (y-axis) that have a particular value, for all possible pixel values (x-axis).

​

​They are a good way to visualise certain aspects of an image.

Open Ended

The image on the left has been processed to make the image on the right. Describe how the image has changed.  How have the pixel values changed?

​Brightness

​The brightness of an image can be changed by adding (or subtracting) a fixed value from every single pixel value.

​

​This shifts the histogram either left or right.

Open Ended

The image on the left has been processed to make the image on the right. Describe how the image has changed.  How have the pixel values changed?

​Contrast

​Contrast describes how different the light parts of an image are to the dark parts.

Increasing contrast therefore makes the dark pixels darker and the light pixels lighter.​

​

Increasing contrast spreads the histogram over a wider range of pixel values. The area remains constant as the number of pixels is fixed.

Open Ended

The image on the left has been processed to make the image on the right. Describe how the image has changed.  How have the pixel values changed?

​Colour Inversion

​Colour inversion switches each pixel colour for the opposite on the colour wheel.

​

​The pixel value is subtracted from the maximum possible value for every pixel.

​

​For RGB colour this happens for each of R,G and B separately.

Open Ended

The image on the left has been processed to make the image on the right. Describe how the image has changed.  How have the pixel values changed?

Edge Detection

​Edge detection colours the image black except where there is a sudden change in pixel value.

​

​Each pixel value is multiplied by four and then has the N, S, E and W pixel values subtracted.

​

​There is a one pixel boundary where the algorithm does not work.

Open Ended

The image on the left has been processed to make the image on the right. Describe how the image has changed.  How have the pixel values changed?

Median Filter

​A median filter is a method for reducing random noise in an image.

​

​Each pixel value is replaced with the median average of the surrounding pixels (N, S, E and W).

​

​Using the median average, rather than the mean, reduces the amount of blurring that occurs.

​Noisy image

​Median filtered