摘要

Based on the feature of performance detection of a seedmeter, an intelligent detecting method for seed position was proposed in this paper. Two cameras with orthogonal optical axis were applied to monitoring seed position. Mass center of seed was regarded as detecting attribute, and the image coordinate of mass center calculated by method of moment was input into BP. Further more, two neural networks were constructed to synchronously process left and right images, so object position was gained quickly. The experiments showed that seed position could be obtained and the law of seed dropping was also monitored.

  • 出版日期2001

全文