In recent years,there have been many studies on the recognition of strawberry maturity,but there are still problems such as low recognition accuracy and expensive experimental instruments.These factors make their meth...In recent years,there have been many studies on the recognition of strawberry maturity,but there are still problems such as low recognition accuracy and expensive experimental instruments.These factors make their methods difficult for farmers to use.To solve these problems,we developed a fast,non-destructive,accurate and convenient method for strawberry maturity identification using smartphones.In this paper,strawberry maturity is divided into three levels:mature,nearly-mature and immature.Considering the actual strawberry harvest process and postharvest handling,we focus on the differentiation between the mature and the nearly-mature ones to help farmers reduce possible damage in transit and improve profitability.We obtained the images of strawberries with different maturities at 535 nm and 670 nm wavelengths through a smartphone and got absorbance data by image processing based on the region of interest.The absorbance data were used to establish three maturity recognition models—i.e.,multivariate linear,multivariate nonlinear and SoftMax regression classifier.The results showed that the multivariate nonlinear model had the highest identification accuracy(which is over 94%)in the greenhouse.Therefore,this method has considerable potential as a means for rapid recognition of strawberry maturity.展开更多
基金This research was supported by the National Training Program of Innovation and Entrepreneurship for Undergraduates(No.201710019174).
文摘In recent years,there have been many studies on the recognition of strawberry maturity,but there are still problems such as low recognition accuracy and expensive experimental instruments.These factors make their methods difficult for farmers to use.To solve these problems,we developed a fast,non-destructive,accurate and convenient method for strawberry maturity identification using smartphones.In this paper,strawberry maturity is divided into three levels:mature,nearly-mature and immature.Considering the actual strawberry harvest process and postharvest handling,we focus on the differentiation between the mature and the nearly-mature ones to help farmers reduce possible damage in transit and improve profitability.We obtained the images of strawberries with different maturities at 535 nm and 670 nm wavelengths through a smartphone and got absorbance data by image processing based on the region of interest.The absorbance data were used to establish three maturity recognition models—i.e.,multivariate linear,multivariate nonlinear and SoftMax regression classifier.The results showed that the multivariate nonlinear model had the highest identification accuracy(which is over 94%)in the greenhouse.Therefore,this method has considerable potential as a means for rapid recognition of strawberry maturity.
