介电特性在番茄和苹果品种识别中的应用

Present study on identifying species of fruits and vegetables are mostly limited to image processing method，which has the shortcoming of high investment. In order to study simple, cheap and low cost fruit identification method, dielectric properties of tomatoes, such as FA-189, 402 and Guocui species as well as apples, such as Fuji and Hongxing species were studied. Applied , the ratio of tomatoes’ relative dielectric constants under 10kHz and 100kHz, and , relative dielectric constant under 10kHz, as indexes in identifying tomato species. It was pointed that the sample was 402 when , it was FA-189 when and , and it was Guocui when and . The average identification ratio for tomato species was 81%. Meantime, BP neural network was also used in identifying tomato species. Its average identification ratio was 80% under the network of 2-11-3. Based on difference between relative dielectric properties, , and resistivity, , of Fuji apples and Hongxing apples, used as indexes in identifying apple species. The sample was Fuji apple if , or it was Hongxing apple. The average identification ratio of this method was 91%. This study provides a new method for identifying fruit’s species automatically and offers a new approach for applying dielectric properties in agricultural engineering.