Research On Wheat Area Extraction Based On MODIS-NDVI Time Series

Wheat is one of the major food crops in northern China. As food resources, It is not only the important resources to the national economy and people’s livelihood, but also to the vital national resource strategy. Wheat production has the following characteristics:wide scope of planting area, strong seasonal changes, regional differences, low economic benefit per unit area. The traditional manual method to obtain the field survey of crop is space segment information. This method is very difficult in terms of the efficiency of technical means and the social economic point of view. It’s of great significance that how to accurately and rapidly access to wheat distribution information in North China can apply to the guidance of wheat production and ensure national food security. With the rapid development and application of technology of remote sensing, remote sensing has a unique advantage in Vegetation dynamic monitoring and forecasting. Its characteristic are less time-consuming, high efficiency, and a wide range of time-continuous high repeatability, which is traditional crop monitoring methods can not matched. Therefore the use of remote sensing technology can solve this problem.The area extraction of crop planting is a key in crop dynamic monitoring. Because of the influence of crop planting structure and farming in our country, such as the complexity of the crop species, caused by geographical differences between "synonyms spectrum" and "foreign body with spectrum" phenomenon of the existence, It is difficult to distinguish between crop spatial distribution characteristics from only a single image spectral information. Therefore, according to planting structure of Chinese main food crop and crop’s phenological period, using time series spectral information of remote sensing data,which can be good to combine the seasonal variation of crop growth mechanism with the spectral characteristics of vegetation.Finally, through a certain method to solve identification of crop and area of extraction.In this study, Using MOD13Q1 of MODIS land cover data products to acquire 23 phase normalized difference vegetation index(NDVI) of 2010, by remote sensing technology extracting wheat planting area of Haihe basin is the research content.How to improve the accuracy of extracted wheat area and the quality of remote sensing data is focus. Main works are showed by the following aspects:(1)Decode Haihe basin area MODIS-NDVI data product quality.The quality of MODIS-NDVI remote sensing data will affect plant area of wheat acreage. As a result, the stduy consider the quality of remote sensing data from a quantitative point of view. In accordance with MODIS-NDVI data the quality file of QA datasets, use relevant methods to evaluate MODIS-NDVI data quality.(2)The reconstruction of MODIS-NDVI time series. The main means are different time series filtering methods reconstructed MODIS-NDVI comparison, according to the characteristics of the Haihe basin crops, through comparison analysis above methods. Finally, the application of the weighted S-G filtering to Haihe basin MODIS-NDVI to reconstruct and assess accuracy(3)Establish extraction model of wheat planting area in Haihe basin based on remote sensing data. Since the significantly different of conditions from north to south hydrothermal in study area, a greater impact on crop growth result in the same kind of crop phenology of dislocation. So using MODIS-NDVI time series of remote sensing data to set a single threshold value to extract wheat acreage in Haihe basin is not feasible. According to the meteorological biological theory, temperature is a key of affecting crop growth. This study chooses temperature factors in wheat planting area partition, can ensure the accuracy of extraction of wheat.(4)Accuracy evaluation of wheat planting area’s extraction. The study obtained precision of 87% of Haihe basin wheat acreage after precision evaluation.It conforms to the precision requirements of the large area crop extraction.