Crop Mapping and Spatio–Temporal Analysis in Valley Areas Using Object-Oriented Machine Learning Methods Combined with Feature Optimization

Abstract

Timely and accurate acquisition of crop planting areas and spatial distribution are deemed essential for grasping food configurations and guiding agricultural production. Despite the increasing research on crop mapping and changes with the development of remote sensing technology, most studies have focused on large-scale regions, with limited research being conducted in fragmented and ecologically vulnerable valley areas. To this end, this study utilized Landsat ETM+/OLI images as the data source to extract additional features, including vegetation index, terrain, and texture. We employed the Random Forest Recursive Feature Elimination (RF_RFE) algorithm for feature selection and evaluated the effectiveness of three machine learning algorithms—Support Vector Machine (SVM), Random Forest (RF), and Rotation Forest (ROF)—for crop extraction. Then, based on the optimal classifiers, the main crops in the Huangshui basin for the years of 2002, 2014, and 2022 were extracted. Finally, the transfer matrix, the gravity center model, and the Standard Deviation Ellipse (SDE) model were used to analyze the spatio—temporal changes of crops over the past 20 years in the Huangshui basin. The results showed that the spectral, vegetation index, and terrain features played a crucial role in crop extraction. Comparing the performance of the classifiers, the ROF algorithm displayed superior effectiveness in crop identification. The overall accuracy of crop extraction was above 86.97%, and the kappa coefficient was above 0.824. Notably, between 2002 and 2022, significant shifts in crop distribution within the Huangshui basin were observed. The highland barley experienced a net increase in planting area at a rate of 8.34 km2/year, while the spring wheat and oilseed rape demonstrated net decreases at rates of 16.02 km2/year and 14.28 km2/year, respectively. Furthermore, the study revealed that highland barley exhibited the most substantial movement, primarily expanding towards the southeast direction.

Combinations of Feature Selection and Machine Learning Models for Object-Oriented “Staple-Crop-Shifting” Monitoring Based on Gaofen-6 Imagery

“…Comparing our results with existing studies, Wang et al [46] achieved the highest OA of 77.12% by classifying large-scale regional crop types by combining four machine learning models and two deep learning models with time-series satellite data. Fu et al [28] constructed features based on multiscale segmentation and extracted crop information for the river valley area based on Landsat imagery. The OA was 86.97% and the kappa coefficient was 0.82.…”

“…Zhang et al [27] constructed a feature selection method based on the optimal extraction cycle of features based on Sentinel-2 remote sensing images, which significantly improved the recognition accuracy of mountain rice, with the best overall classification accuracy and kappa coefficient of 86% and 0.81, respectively. Fu et al [28] demonstrated that the Random Forest Recursive Feature Elimination (RF_RFE) algorithm can provide more useful features and improve the crop identification accuracy after feature selection by 1.43% to 2.19%, 0.60% to 1.41%, and 1.99% to 2.18% in 2002, 2014, and 2022, respectively, compared to crop identification without feature optimization. Jin et al [3] used Decision Trees to construct the optimal feature space in order to achieve fine crop classification and used an object-oriented Random Forest classification algorithm to classify the multi-feature space, and the final overall accuracy and kappa coefficient were 90.18% and 0.877, which were greater than for the pixel-level and single-feature classification accuracies.…”