农作物遥感估产已经是广泛应用的技术 ,但由于西南地区复杂的地貌类型以及遥感信息源等因素的限制 ,农作物估产方面的研究起步较晚 ,因而研究山区的农作物遥感估产对西南地区有重要的理论价值和现实意义。安宁河谷为一南北走向的山间盆地 ,是四川省第二大粮食生产基地。本文运用遥感 (RemoteSensing)、地理信息系统 (GIS)和全球定位系统(GPS)———简称 3S ,采取全数字式判读方式提取冬小麦的播种面积 ,进行野外点、线采样 ,并利用数理统计方法 ,建立了该地区的冬小麦估产模型

Southwest China is a mountainous region. Elevation and landform (e.g.slope) are two key factors to limit the land usemode, furthermore, to influence food variety and yield. Food Yield-Estimation (FYE) by Remote Sensing is an advanced and mature technology. Because of the complicated landform and fragmentary surface of Southwest China and the limitation of Remote Sensing source, the study of FYE lags behind other regions. It is of great significance to the establishment of the FYE model in mountainousareas of SouthwestChinaboth in theoretical value and practical aspect. Anning river basin is a SN-trended intermontane basin,and the second largest food production base in Sichuan province. The authors makeuse of the combination of Remote Sensing, GIS and GPS to place point and line field samples, get the sowed area of winter wheatand establish the winter wheat yield-estimation model. The establishmentof the yield of the winter wheat is decided by factors of two aspects. One is the sowed area;the other is the per unit area yield model of the winter wheat. The acquisition of sowed area of the winter wheat is depended on the agrarian area interpreted by remote sensing and the percentage of the winter wheat. The establishment of the unit yield model is depended on the choice of the formula and the parameters used to make up the formula. There are severalkinds of factors to influence the growth condition and decide the unit yield of the winter wheat, such as climate (include temperature, precipitation, moisture, sun shine etc.), biological factor, planting method, soil potentiality, landform and so on. In these factors, the elevation and the gradient are two factors that can reflect the natural features of the winter wheat in Southwest China. The winter wheat's planting position and yield are influenced by elevation andslope. In fact, the remote sensing factor is not really a factor to decide the yield, but only the compositive reflection of other factors on the sensor. Zone division is anotherkey factor to decide the choice of the parameters of the formula and influencethe precision of the model, because different zones have different major factors to influence the yield. The selection of the formula is dependedon the mathematicmethod, the knowledge of the region's natural and economic conditions and the accumulation of experience. The precision of the model is decided by the representative of the formula and the parameters.