Week 9
October 3, 2017
Today, we did not have our class for DSP lab because of an unfortunate circumstance that happened to our instructor. (Get well soon sir.) Nevertheless, the submission of our major activities were not postponed. We placed the hard copies of our Major Activity 1, which is the journal formatting, and Major Activity 2, which is about the morphological extraction of the coffee beans from the photos given to us, to Sir Arboleda’s table on their faculty. And we have sent the soft copies through e-mail. Last weekend until yesterday, I have worked for the major activity 2 which somehow required a meticulous effort because we needed to extract parameters from the 30 samples from each of the 19 photos. I used the downloaded photos from the laptop so the pixels of the photos would be uniform with each other and it would use the maximum size. The size of the photos can change when downloaded from phone and aside from this, Sir Arboleda told us not to resize the photos because the area and the perimeter would also change in dimensions. Upon extracting the features using Matlab, some of the photos displayed after using the vislabels command showed a lot of numbers as labels which are treated as noise. To reduce the noise, my classmates and I discovered that the value for c on the Matlab command can be changed. So Instead of c=255-b, I also used c=250-b, c=280-b, and c=285-b on other photos. After this modification, the labels of the photos were already clear enough to distinguish so their area and perimeter could be identified. However, there are coffee beans which really have 2 numbers placed on it so both labels were checked using the g(n) command where n is the label of the coffee bean. When g(n) is entered, the area and perimeter of the coffee bean would show up. One label indicated a very small value for area and the other one shown an objective number of pixels so the latter is preferred to be used. So the data were recorded on an excel file. The properties, equivalent diameter and roundness were obtained by using their formulas which were executed on the excel file and by utilizing the values of area and perimeter. After the tables on the excel file were filled with the data, I transferred them to Microsoft Word to make them ready for printing. Also, I placed my name, section, and instructor’s name on the document.