Feasibility of Detecting Sugarcane Varieties by Electronic Nose Technique in Sugarcane Syrup

Sugar cane is one of the most important industrial plants which is the first source of sugar production in Iran (about 40-50%). The sugar industry plays a key role among the various industries of the country with the daily supply of energy to the citizens. In addition to household consumption, sugar is of particular importance in the food industry because of its sweetening and volume properties. Sugarcane content is often in the range of 10-15% and in some cases up to 17%. Various factors such as variety and date of planting or harvesting of last year are important to start harvesting. On the other hand, Sugarcane cannot be stored in the factory and its sugars factors are decomposed quickly by storage and sugar cane weight decreases due to loss of moisture. It would be better if the factory consumes more fresh sugarcane. Therefore, an electronic nose instrument was used to test the variety of sugarcane syrup and its association with the odors emitted from it to identify the variety of sugarcane for harvesting time. Four sugarcane varieties (CP57, CP69, IRC99-02, and CP48) were selected from the sugarcane sample fields. Linear discriminant analysis (LDA), principal component analysis (PCA) and neural networks (ANN) were used to detect the different sugarcane varieties. The results showed that all three methods had high accuracy in variety classification. But the LDA and PCA methods performed better than the ANN method. So that, the classification accuracy of sugarcane varieties was 98.33%, 97% and 96.7%, respectively. The results showed the high ability of the olfactory machine to diagnose between the sugarcane varieties, which can be used as a rapid and low cost instrument in the sugarcane industry.