Study of the Bruise Detection System for Apples Using Fluorescence Image

• Main Authors: Mu-Te Chen, 陳慕德
• Other Authors: Yi-Chich Chiu
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/39928854720673522095

碩士 === 國立宜蘭大學 === 生物機電工程學系碩士班 === 99 === The objectives of this study were to analyze and detect the apple bruise using chlorophyll fluorescent images and to develop an inspection system to clarify the different levels of the bruise of apples. To create different levels of the bruise of apple, a fruit impact system has also been developed in this study, which the force-time curve during impact processes can be automatically recorded. The graphical programming software, LabVIEW 8.5 was applied to design the man-machine interface and the control system of the bruise inspesion system. The programming software Matlab 2007a was used to develop the bruise classified algorithm based on the phenomenon of the chlorophyll fluorescence to identify the bruise levels of apples. The experimental result showed that the developed fruit impact system obtains a highly linear regression formula, with the determination coefficient of 0.9901, for the relationship between the pendulum angles and the force peak. To find a better reference for pre-inspection dark adaptation time, this study was to establish the fluorescence restitution curve of apple, the result showed that the fluorescent image of the 2nd second after excitation light compared with the dark adaptation time can obtain a better linear regression formula, with a determination coefficient of 0.724. To evaluate the performance of the bruise inspecsion system, three specified forces, 6.92±0.493kgf、8.37±0.466kgf and 10.40±0.860kgf, were applied to impact the Golden Delicious apples by the developed fruit impace system to produce the fruit bruise. The results showed that the detectable ratio are 75.0%, 75.0%, and 87.5%, respectively, with an average of 79.2%, after half an hour of impact and 87.5%, 100%, and 87.5%, respectively, with an average of 91.7% after one hour and 87.5%, 87.5%, and 100%, respectively, with an average of 91.7% after two hour of impact. The detectable ratio highly reaches 100% after four of impact. The experiments also found that if the fruits suffer for lighting lasting above five minutes by the bulb(Reflector 110V 500W, Fuji), it will obtain a lower fluorescence emission from apple. Thus it must avoid the lighting on the fruits before the inspection. Consequently, the developed bruise inspecsion system has a potential to develop an on-line inspection system in the future to ensure the fruit quality before package.