112 Prakt. Met. Sonderband 38 (2006) 
3. SUMMARY Ct 
The volume fraction of porosity in pressed and sintered steels can be quantified by image SF 
analysis. The described procedure utilises binarisation of the images (pore phase and B 
matrix phase) after setting a threshold in greyscale colour space. The bias of the ; 
measurement is caused by several sources of error (both systematic and statistical errors). 
Careful sample preparation is crucial to obtain correct results. The separation of the 
sample surface into porosity and matrix is influenced by the microscope magnification 
used and digital processing of the images taken. The magnification used is a main source 
of error, since a high magnification (500x) leads to high standard deviations in porosity and 
subsequently to broad intervals of confidence unless very numerous images are 
measured. A low magnification (100x) yields narrower intervals of confidence for a given 
number of micrographs, but the systematic error becomes more prominent because of the KL 
increasing overall pore perimeter. A magnification of 200x was determined as suitable for 
quantification, since both intervals of confidence and calculated means for porosity were Ver 
acceptable at reasonable effort. Digital pre-processing of the images with the aim of easier mit 
separation of porosity and matrix is a source of systematic error. The use of a high pass mik 
filter to produce sharp boundaries between pores and matrix was detrimental to the Kre 
results. By using an edge enhancement filter with optimized parameters the result of em 
quantification could be improved. I 
(> 
Gru 
REFERENCES bec 
[1] ENISO 2738 
[2] Drozda, M., Kaysser, W.A.: “Influence of Metallographic Treatment on Pore Analysis 
of Sintered Iron”, Praktische Metallographie, 16, 1979, p. 578 1 
[3] Weidmann, E., Nielsen, B., Arner, J.: “Metallographic Preparation of Powder 
Metallurgy Parts”, Struers Application Notes, 2005 Gu 
[4] Wojnar, L., Kurzydtowski, K., Szala, J.: "Quantitative Image Analysis”, in: ASM Te: 
Handbook Vol. 9 “Metallography and Microstructures”, 2004, p. 413 gür 
[5] Paciornik, S., de Pinho Mauricio, M.H.: “Digital Imaging”, in: ASM Handbook Vol. 9 bee 
“‘Metallography and Microstructures”, 2004, p. 386 wei 
[6] Vander Voort, G..” Influence of magnification on feature-specific image analysis sor 
measurements”, Metallography, 21, 1988, p. 327 übe 
Gre 
Rui 
Ur 
anr 
teil 
GJ 
get 
ung 
ver 
Spl 
Ho 
unc 
ges 
rele 
ve