| dc.description.abstract | This study was aimed to measure the degrees of roasting of coffee to achieved a quantitative method for the valuation degrees of roasting, developed the automatic coffee roasting machine, determined the duration of coffee roasting on various degrees of roast processed by coffee roast machine based on the models created from analysis, and integrated a control system based on minicomputer Raspberry Pi. The samples tend to cover all degrees of roast, they were obtained from series of roasting processes which periodically harvested, with three variation of roast temperature (115oC, 125oC, and 140oC) and two variation of weight (2kg and 5 kg). The roasted coffee image data from ccd camera were extracted using digital image processing to generate six main image variables, namely: hue, saturation, intensity, red, green, and blue. The same method used to the image of eight standards digress of roast. Each variable from those two data were paired and compared to get the relationship between them at every degree of roast. The linier regression and correlation method were used to identify and analyzed the data. Based on the analysis, the R average variable tend to have the greatest relationships and successfully predicted series of roast duration that covered all degrees of roast at every roast temperature. The automatic coffee roasting machine based on Raspberry Pi also worked as expected. | en_US |
