基于GF-1 PMS的玉米出苗情况提取

及时、准确的玉米出苗情况监测可以为农田经营管理和宏观决策提供玉米出苗期的生长信息，便于及时采取适当的科学管理措施，达到增产增收的目的。归一化植被指数（Normalized Difference Vegetation Index, NDVI）及与作物生长状态关系密切，可以用于评价玉米的出苗情况。采用地块的NDVI均值及均方差可以反映地块内玉米出苗的综合情况，以吉林省长春市九台市榆树村和解放村为例，对研究区玉米出苗情况信息进行提取，最后利用实地采集的验证数据对提取结果进行验证。研究区玉米出苗情况信息提取结果的总体精度达到80%，表明利用上述方法能够在一定程度上反映玉米的出苗情况，可以为玉米出苗情况评价提供参考依据。

Maize Emergence Condition Information Extraction Base on GF-1/PMS

Seedling stage is an important phenophases of crop growth periods. Pixel-based difference monitoring method is one of the main methods of emergence condition information extraction. However, pixel-based difference monitoring method has some defections so that it cannot effectively reveal the integrated situation of individual field blocks which is essential for field management. Mean and standard deviation of Normalized Difference Vegetation Index (NDVI) was introduced to the study of maize emergence condition information extraction. Taking the Yushu Village and Jiefang Village as cases, the information of maize emergence of field blocks was extracted. Based on mean and standard deviation of Normalized Difference Vegetation Index from GF-1 PMS data which file time is May 23, 2017, in Yushu city, maize emergence condition information were extracted, threshold nodes were build. The NDVI value of a crop at a certain time is closely related to the crop growth status and area at that moment, and to a certain extent, it can reflect the degree of crop growth. In this study, based on the NDVI value of the Yushu Village corn field blocks, the NDVI average value (NDVIMean) and the mean square error (NDVIStd) of different field blocks are counted. The pixel-based NDVI value is converted to an index that can reflect whether the corn emergence condition in the field blocks is neat and robust. The extraction of corn emergence based on field blocks should not only reflect the difference between plots, but also reflect the difference between field blocks. The two conditions of strong seedlings (NDVI higher than the threshold) and neat emergence (NDVI root mean square below the threshold) were met at the same time, the seedlings come out evenly, and otherwise it is defined as out of line. Vegetation characteristics are not obvious (NDVI is lower than a certain value) and the evaluation is not emerged. The results show that 1) GF-1 PMS data has the characteristics of wide coverage and high spatial resolution. It can meet the needs of inverting the level of corn at the plot level and has the potential for large-area and high-precision inversion of crop emergence. 2) The total precision of the extraction results of corn seedlings in this paper reached 80%, which is in line with the actual growth status of the maize in the study area. This shows that this method can reflect the seedling emergence of corn to a certain extent, and can provide scientific reference for the optimization of farmland management. 3) The combination of NDVI average and RMS can reflect the degree of seedling and regularity of corn in the field blocks to a certain extent, which makes up for the insufficiency in the past by relying on pixel-based difference monitoring methods and the data of the whole crop growth period, has the advantage of reducing the amount of data. It is concluded that the method of extracting corn seedlings proposed in this paper can reflect the comprehensive situation of corn emergence at the level of landmass, and can provide reference for the monitoring and extraction of seedlings of other crops. In the future, the method can be applied to other crops to verify the applicability of the method.