Abstract—With the development of precision agriculture and agricultural insurance, there are increasing demands for the fine-scale crops planting information in large areas. This paper presents a comprehensive approach for crop type identification and planting area estimation at a farmland plot scale by collaboratively utilizing the high-and-moderate spatial resolution satellite imagery. The proposed method roughly contains four steps: firstly, by implementing an image segmentation and a following manual editing, the objects of farmland plot with exact boundary are extracted from the high spatial resolution imagery; Secondly, with the effective-data processing technology and spectral indices calculation based on the multi-temporal moderate resolution images, the cloudlessly fragmentary effective data which served as the source of properties for plot objects is obtained; thirdly, the specific NDVI time-series and phenological parameters for each farmland plot are further derived from these effective data; Lastly, based on the multi-dimensional feature space of plot objects, the crop types and corresponding planting areas are mapped using the Random Forest Classifier. This approach has been tested for several crops in Sihong County, Jiangsu Province, China. The results showed that, this method can map the distribution of wheat, rice and corn at a farmland plot scale with relatively high accuracy. The user accuracy of wheat, rice and corn reached to 98.62%, 97.05% and 97.74%, respectively, and the overall accuracy was 95.36% with a Kappa coefficient of 0.936. The area accuracy of these three crops also amounted to 94.18%, 93.37% and 91.23%, respectively. This experiment illustrated the effectiveness and usefulness of the proposed method, and was referential to finely planting information extraction for other crops.

 

Index Terms—remote sensing, crop identification, farmland plot scale

Cite: Huang Qiting, Luo Jiancheng, and Dong Wen, "Crops Planting Information Retrieval at Farmland Plot Scale Using Multi-Sources Satellite Data," Journal of Advanced Agricultural Technologies, Vol. 4, No. 2, pp. 96-103, June 2017. Doi: 10.18178/joaat.4.2.96-103