Determination of applicability of mother wavelet for coarse and fine grading of mangoes

To improve market value, any horticulture product should meet quality standards. Abnormality in fruit shape represents a defective fruit which can be rejected while exporting. This study was conducted to evaluate applicability of mother wavelet for mango shape sorting. A database of images was first formulated from mangoes with different shapes and sizes. The methodology consists of two phases. During the first phase, deformed mango fruits were separated out from well formed one using Wavelet based shape features and Support Vector Machines. The average correct classification was 86.36% for a testing set composed of 220 mangoes. The second phase involves extraction of size characteristics of well formed mangoes using statistical method. Experimental results showed that well-formed mangoes are classified by statistical size estimation method achieves 93.963% accuracy indicating that developed sorting algorithm has potential in detection and sorting of deformed mango fruits.