Producing accurate crop maps during the current growing season is essential for effective agricultural monitoring. Substantial efforts have been made to study regional crop distribution from year to year, but less attention is paid to the dynamics of composition and spatial extent of crops within a season. Understanding how crops are distributed at the early developing stages allows for the timely adjustment of crop planting structure as well as agricultural decision making and management. To address this knowledge gap, this study presents an approach integrating object-based image analysis with random forest (RF) for mapping in-season crop types based on multi-temporal GaoFen satellite data with a spatial resolution of 16 meters. A multiresolution local variance strategy was used to create crop objects, and then object-based spectral/textural features and vegetation indices were extracted from those objects. The RF classifier was employed to identify different crop types at four crop growth seasons by integrating available features. The crop classification performance of different seasons was assessed by calculating F-score values. Results show that crop maps derived using seasonal features achieved an overall accuracy of more than 87%. Compared to the use of spectral features, a feature combination of in-season textures and multi-temporal spectral and vegetation indices performs best when classifying crop types. Spectral and temporal information is more important than texture features for crop mapping. However, texture can be essential information when there is insufficient spectral and temporal information (e.g., crop identification in the early spring). These results indicate that an object-based image analysis combined with random forest has considerable potential for in-season crop mapping using high spatial resolution imagery.

Reliable cropland parcel data are vital for agricultural monitoring, yield estimation, and agricultural intensification assessments. However, the inherently high landscape fragmentation and irregularly shaped cropland associated with smallholder farming systems restrict the accuracy of cropland parcels extraction. In this study, we proposed an adaptive image segmentation method with the automated selection of optimal scale (MSAOS) to extract cropland parcels in heterogeneous agricultural landscapes. The MSAOS method includes three major components: (1) coarse segmentation to divide the whole images into homogenous and heterogeneous regions, (2) fine segmentation to determine the optimal segmentation scale based on average local variance function, and (3) region merging to merge and dissolve the over-segmented objects with small area. The potential cropland objects derived from MSAOS were combined with random forest to generate the final cropland parcels. The MSAOS method was evaluated over different agricultural regions in China, and derived results were assessed by benchmark cropland parcels interpreted from high-spatial resolution images. Results showed the texture features of Homogeneity and Entropy are the most important features for MSAOS to extract potential cropland parcels, with the highest separability index of 0.28 and 0.26, respectively. MSAOS-derived cropland parcels had high agreement with the reference dataset over eight tiles in Qichun county, with average F1 scores of 0.839 and 0.779 for the area-based classification evaluation (F_ab) and object-based segmentation evaluation (F_ob), respectively. The further evaluation of MSAOS on different tiles of four provinces exhibited the similar results (F_ab = 0.857 and F_ob = 0.775) with that on eight test tiles, suggesting the good transferability of the MSAOS over different agricultural regions. Furthermore, MSAOS outperformed other widely-used approaches in terms of the accuracy and integrity of the extracted cropland parcels. These results indicate the great potential of using MSAOS for image segmentation in conjunction with random forest classification to effectively extract cropland parcels in smallholder farming systems.

Soybean cultivation in China has significantly decreased due to the rising import of genetically modified soybeans from other countries. Understanding soybean’s extent and change information is of great value for national agricultural policy implications and global food security. Some previous studies have explored the quantitative relationships between crop area and spectral variables derived from remote sensing data. However, both those linear or non-linear relationships were expressed by global regression models, which ignored the spatial non-stationarity of crop spectral signature and may limit the prediction accuracy. This study presented a geographically weighted regression model (GWR) to estimate fractional soybean at 250 m spatial resolution in Heilongjiang Province, one of the most important food production regions in China, using time-series MODIS data and high-quality calibration information derived from Landsat data. A forward stepwise optimization strategy was embedded with the GWR model to select the optimal subset of independent variables for soybeans. Normalized Difference Vegetation Index (NDVI) of Julian day 233 to 257 when soybeans are filling seed was found to be the most important temporal period for sub-pixel soybean area estimation. Our MODIS-based soybean area compared well with Landsat-based results at pixel-level. Also, there was a good agreement between the MODIS-based result and census data at county level, with the coefficient of determination (R²) of 0.80 and the root mean square error (RMSE) was 340.21 km². Additionally, F-test results showed GWR model had better model goodness-of-fit and higher prediction accuracy than the traditional ordinary least squares (OLS) model. These promising results suggest crop spectral variations both at temporal and spatial scales should be considered when exploring its relationship with pixel-level crop acreage. The optimized GWR model by combining an automated feature selection strategy has great potential for estimating sub-pixel crop area at regional scale based on remote sensing time-series data.

Accurate and timely crop maps are crucial for monitoring agricultural production. Current supervised classification methods based on remote sensing rely heavily on ground-truth samples collected at a high cost, and the years without sampling highly limit classification accuracy. To address such a challenge, we proposed a time-migration method based on historical training samples collected in 2017, 2018 and 2020 to conduct supervised crop classification mapping in the target year (2021) with no ground samples. We chose Hailun City, Heilongjiang Province of northeastern China, as the study site; the major crops included corn, soybean, and rice. We reconstructed time series of Sentinel-2 data and selected spectro-temporal features to identify standard crop phenological curves. We calculated the similarity between reference and image spectra and designed label-matching rules to identify training samples through the dynamic time warping algorithm. We then used the historical samples to map the crop types of the target year. The results showed that the migration accuracy reached 95% for major crop. Using these samples as training data with a random forest to classify the target year, the overall accuracy reached 94.13%. The new sample time-migration method proposed in this study can efficiently migrate historical samples, greatly reducing the cost of ground-truth sampling.