On the characterization of polymer powders mixing dynamics by texture analysis

Texture analysis as a tool to study the kinetics of wet agglomeration processes

“…In order of increasing complexity, we can enumerate descriptors based on the average value, the standard deviation (STD), the third moment of the intensity level histogram and those based on the gray level co-occurence matrix (GLCM). GLCM is a second order statistics which describes the spatial correlation between couples of neighbouring pixels by compiling the frequency for which different gray level intensities are found in a defined area (Gosselin at al., 2008). On the whole 11 descriptors have been considered but finally the standard deviation STD of f(z i ) has been used in this study since most of the texture descriptors were found to be equivalent to STD (the same or similar information could be .…”

Polymer powders mixing part II: Multi-component mixing dynamics using RGB color analysis

“…The camera was fixed to be normal to the free surface of the mixture since previous experiments done under several viewing positions were found to give the same information (Gosselin et al, 2008;Aït Aissa et al, 2010). Each experiment was left running for 20 min and frames of each video file were extracted using the ImageGrab30fr software every 2 s. The images were later cropped and enhanced using WCIF Image J s software.…”

“…Each experiment was left running for 20 min and frames of each video file were extracted using the ImageGrab30fr software every 2 s. The images were later cropped and enhanced using WCIF Image J s software. Further details on the experimental procedures are given in Gosselin et al (2008) and Aït Aissa et al (2010).…”

“…As described in our previous works (Gosselin et al, 2008;Aït Aissa et al, 2010), the mixing degree is calculated as:…”

“…Also, decreasing the filling ratio (volume fraction occupied by the powder inside the cylinder, f) and increasing the initial content of one color over the other were both found to increase the mixing dynamics. From the data obtained, a modification of the second order mixing model of Gosselin et al (2008) was proposed to predict the mixing curves. Finally, it was found that the rolling regime was valid only for 10 À 4 rFrr10 À 2 and 17% rfr43%.…”