A microwave reflectometer technique for classifying a rubber cup lump

摘要

Rubber cup lumps, the product from rubber trees (which is a global economic plant), are usually graded by human inspectors for two classes, wet or dry, for selling in a rubber yard market. Grading rubber cup lumps using human inspection is unreliable and doubtful for the rubber sellers. In this paper, we, therefore, propose a technique to classify rubber cup lumps using a microwave reflectometer operating at 2.45 GHz which is more robust and reliable. Since the moisture content in a rubber cup lump is related to the magnitude of the reflected microwave from the cup lump, automatic classifying based on the reflection coefficient is proposed. A probe sensor based on a circular waveguide was designed and implemented. The sensor is optimum for rubber cup lump sizes of 500-1500 cm(3). A classification algorithm was developed with the magnitude function of the reflection coefficients formulated by measuring 100 samples of the two classes of rubber. To validate the algorithm and the sensor, 300 rubber cup lumps were tested and the proposed technique achieved 88.67% overall accuracy.