Non-destructive detection method of watermelon maturity based on BMV features

Maturity of watermelons is hard to evaluate by non-destructive methods. In this paper, a watermelon maturity non-destructive detection method based on band magnitude vector (BMV) features of acoustic impulse response was introduced. BMV feature of audios was presented for melon mature detection. A simple acoustic collection platform was built for testing the correlation of BMV and ripeness stages, and it was compared with four acoustic features that already were used in watermelon non-destructive detection. Then, the influence of acoustic response and BMV feature was researched in different impacting strength. BMV features from two varieties of melons were detected for ripeness stages by PNN algorithm. Experimental results showed that the correlation of BMV and ripeness was the highest in all of the features, the influence of impacting strength was little for BMV, and the method had high accuracy for detecting maturity of two melon varieties.