by Discount Bluehost


Published: Monday, 23 December 2013

Exercise 1.



Image: joints.pgm

Task: Extract and measure the extent of corrosion along the joints. The image consists of composite material. The composites are connected through their border (joints). With the advance of erosion, these connections weaken. The task is to propose an automatic erosion measurement. This measurement consists of segmenting the joints (the black marker is not part of the joints) and then sub-segmenting the corroded parts of these segments. The corroded parts consists of the thin black areas along the joints.

Operators: genball, gradmorph, segmentheight, watershed, segmentarea, closing, dilation, area

Exercise 2.




Image: pex1.pgm, pex2.pgm

Task: Correct the gradient illumination. During the image acquisition the different areas of images have been acquired with different background illumination. This means that some areas of the image are darker then others. This darkness is not a qualitative property of the image, so it needs to be filtered. The resulting image should retain the features of the original image but not the background gradient.

Hint: The background represents a low frequency variation in the image. The background can therefore be extracted by some low frequency filter, such as gaussian blur or morphological opening.

Operators:gaussianfilter, minmax, normalize, opening

Exercise 3.



Image: trois003.pgm

Task: Correct the frequency noise. The image represents a distinct frequency noise. Correct the noise using the fourier decomposition of the image.

Operators: any2complex, float2byte, dilatball, drawline, byte2float, inverse, mult (*), complex_modulus, fft

Exercise 4.



Image: carbures.pgm

Task: Segment the carbids on the image, and generate the histogram of their volumes.

Operators:openball, closing, fill, center, attribute, colormap.regions

Exercise 5.



Image: mortier_2d.pgm

Task: The image has a central radial illumination gradient. This means that the background illumination is changing in function of the distance from the center. In this exercise we suppose that the illumination is linear from the center, and we are would like to correct the background illumination to get a uniform image. The problem is similar to that of Exercise 2., but here we have more knowledge about the nature of the gradient. Therefore we can make a more accurate correction.

Operators: drawball, border, average, identifyline, math.pow

Exercise 6.

Image: mortier_3d.pgm

Task: Segment the three components of the image (cavities, gray filler, dark filler and binder)

Operators:areaopening, areaclosing, medianfilter, asf, inverse, normalize


Hits: 2071

Add comment

Security code