The Elman nonlinear dynamic forecast model for judging water content of the freah tohacco leaf by ANN method and Matlab aoft was built Thee width ,area und mean values aynthetical judgment indexes that covered biS weight in the three-demarcation AHP method were taken aa the Elman network s input in computer vision to overcome the shortcoming of the ainde index. The output conclusion of error analysis in Elman model indicated that the mathematic model could be uaed aa the forecast model for water content. The forecasting precision was more thsn 90 ％. The predicting result achieved its intended goal. Based on thia , the aoft for detectinS water content syatem was developed.%在Matlab中利用ANN方法建立了鲜烟叶含水量Elman非线性动态预测数学模型,利用计算机视觉技术,选取在改进AHP法中占较大权重综合评判指标叶片宽、面积以及均值特征量,作为网络的输入,克服了用单一指标进行评判的缺陷,通过实验数据对Elman模型进行验证和误差分析,结果表明该含水量模型可对鲜烟叶含水量进行预测,预测精度大于90%,预测结果达到预期目标,在此基础上进行了含水量检测系统软件的开发.
展开▼
