|کد مقاله||کد نشریه||سال انتشار||مقاله انگلیسی||ترجمه فارسی||نسخه تمام متن|
|84024||158857||2016||7 صفحه PDF||سفارش دهید||دانلود کنید|
• We studied an alternative e-nose based on gas chromatographic system for fruit classification.
• Combination of partition column – gas sensor could generate a separable chromatogram profile.
• PCA based feature extraction can be used to improve the identification accuracy.
An alternative model of electronic nose systems by applying a combination of partition column with gas sensor was investigated for fruit classification and identification. The principle of physical and chemical separation in chromatography analysis known as an interaction between stationary phase material and compounds is able to profile the flavor sample; thus it is potentially implemented to substitute function of the sensor array on the conventional electronic nose. The electronic nose consists of a sample handling with combination of solenoid valves, a packed partition column coupled with a gas sensor as detector operated under a controlled temperature and data analysis software by using a neural network. The system was tested to classify three different flavors of fruit, i.e. durian, jackfruit, and mango. The result showed that it can generate reliable and repeatable chromatograms, from which, a unique pattern among samples can be extracted. Therefore, the patterns are able to be clearly classified with the neural network. The experiment showed that it can recognize the three different flavors with the level of accuracy of 82%.
Journal: Computers and Electronics in Agriculture - Volume 121, February 2016, Pages 429–435