论文部分内容阅读
土壤盐渍化是干旱、半干旱农业区主要的土地退化问题,像沙漠化一样也是当今世界重要的环境和社会经济问题。以新疆渭干河—库车河三角洲绿洲作为研究区,利用实测的盐碱土光谱数据和地下水埋深、地下水矿化度和表层土壤矿化度等因子构建了基于BP神经网络的盐碱土盐分反演模型。该模型的最终输出结果与期望得到的结果相差不大,在输出的误差中,最大误差为0.92%,最小误差为0.34%,得到的误差在允许范围之内,可以满足实践的需要。通过运行检验程序,得到的模型精度是80.77%。研究表明,应用该方法来反演盐碱土盐分信息是可行的。
Soil salinization is a major problem of land degradation in arid and semi-arid agricultural areas and, like desertification, is an important environmental and socio-economic issue in the world today. Taking the Weigan River and Kuqa River delta oasis in Xinjiang as the study area, the salinity-alkaline soil salt index based on the BP neural network was established based on the measured saline-alkali soil spectral data and the groundwater depth, the salinity of groundwater and the salinity of the surface soil. Play model. The final output of the model is not much different from the expected result. The maximum error of the output error is 0.92% and the minimum error is 0.34%. The obtained error is within the allowable range to meet the practical needs. By running the test program, the accuracy of the model obtained is 80.77%. The research shows that it is feasible to use this method to retrieve salinity-saline information.