Abstract: Crop remote sensing classification and recognition is not only the basis for extracting the planting range and area of crops, but also the basis for monitoring the growth, yield, disasters and other related information of crops. Rape and wheat are the main winter crops in Jiangsu Province. In order to solve the problem of rapid and accurate extraction of main summer crops in Jiangsu Province, taking Yixing city as an example, this paper proposed a summer harvest crop extraction method based on time series images and key phenological features. Based on the different characteristics of different agricultural crops in the process of growth and development, combined with the phenological information of crops, the time series image data expressing the characteristics of each growth stage of crops were used to extract different crops. The extraction results showed that the overall accuracy was 92.65% and the Kappa coefficient was 0.86, indicating that the technology of extracting summer harvest crops in Jiangsu Province was feasible, and the extraction results could provide reference for the estimation of crop planting area, and offer technical support for crop fine monitoring and the monitoring of non-grain production of cultivated land.

Key words: time series image, key phenological characteristics, crop extraction, wheat, rape