The 6th International Conference on Biomass Energy (ICBE2018)

548 / 2018-09-26 11:01:01

Development of correlations for physical properties of biochar based on RGB image recognition technology

Plenary track > I.Others

With the help of image recognition technology, biochar’s characterization could be simplified and accelerated. The object of this study was to analyze the correlation of physical properties of biochar based on RGB image recognition technology and to build up appropriate models for future prediction in industry analysis and calorific value of biochar. In this study, The rice husk and maize straw were pyrolyzed at the temperature of 350, 450, 500, 550 and 600℃ to produce biochar. According to the national standards, the industrial analysis and calorific value of the two kinds of biochars were analyzed using the automatic industrial analyzer and calorimeter. Then the scanner was used to obtain the image of biochars, and the RGB values of 40 points on the scanned image were picked up randomly by the software of Colorpix. Then the R, G and B values were converted to grayscale using transformation formula. Results indicated that there was a strong linear relationship between the grayscale and the pyrolysis temperature of these two biochars, the Pearson’s r of which were both more than 0.9, so the grayscale could reflect the biomass pyrolysis temperature. By analyzing the correlation between the grayscale and industrial analysis or calorific value of the two biochars, some accurate linear or nonlinear models were built. The industrial analysis content of biochar from rice husk accorded to the Asymptotic1 nonlinear model (R2 was in 0.89~0.99), the residual square sum (RSS) of which was 1.2929, 0.2324 and 0.096. While the biochar from maize straw conformed to the linear relationship (R2 was in 0.89~0.96), the RSSs of which were 0.2314, 1.6700 and 0.6176. The image recognition technology and correlation models can be used to deduce the temperature of carbonization, industrial analysis and calorific value results of biochar quickly and accurately, and the innovative study provides a theoretical basis for rapid detection of biochar.