The Classification Method Study of Crops Remote Sensing with Deep Learning, Machine Learning, and Google Earth Engine

Abstract

The extraction and classification of crops is the core issue of agricultural remote sensing. The precise classification of crop types is of great significance to the monitoring and evaluation of crops planting area, growth, and yield. Based on the Google Earth Engine and Google Colab cloud platform, this study takes the typical agricultural oasis area of Xiangride Town, Qinghai Province, as an example. It compares traditional machine learning (random forest, RF), object-oriented classification (object-oriented, OO), and deep neural networks (DNN), which proposes a random forest combined with deep neural network (RF+DNN) classification framework. In this study, the spatial characteristics of band information, vegetation index, and polarization of main crops in the study area were constructed using Sentinel-1 and Sentinel-2 data. The temporal characteristics of crops phenology and growth state were analyzed using the curve curvature method, and the data were screened in time and space. By comparing and analyzing the accuracy of the four classification methods, the advantages of RF+DNN model and its application value in crops classification were illustrated. The results showed that for the crops in the study area during the period of good growth and development, a better crop classification result could be obtained using RF+DNN classification method, whose model accuracy, training, and predict time spent were better than that of using DNN alone. The overall accuracy and Kappa coefficient of classification were 0.98 and 0.97, respectively. It is also higher than the classification accuracy of random forest (OA = 0.87, Kappa = 0.82), object oriented (OA = 0.78, Kappa = 0.70) and deep neural network (OA = 0.93, Kappa = 0.90). The scalable and simple classification method proposed in this paper gives full play to the advantages of cloud platform in data and operation, and the traditional machine learning combined with deep learning can effectively improve the classification accuracy. Timely and accurate extraction of crop types at different spatial and temporal scales is of great significance for crops pattern change, crops yield estimation, and crops safety warning.

Landscape Analysis and Ecological Risk Assessment during 1995–2020 Based on Land Utilization/Land Coverage (LULC) and Random Forest: A Case Study of the Fushun Open-Pit Coal Area in Liaoning, China

“…According to the spectral characteristics of different ground objects, NDVI, NDISI/NDBI/NDWI, visible light RGB bands, and red bands were selected as evaluation parameters, and 10,000 training points in each period were established to classify images in each period. Based on the classification results, Fragstate4.2 was used to calculate the main landscape pattern index within the study area, and the landscape pattern and ecological characteristics were analyzed to comprehensively evaluate the landscape pattern evolution and ecological risk index evaluation from 1995 to 2020 [7,46] (Figure 2).…”

“…(2) The supervised classification method is fast and efficient, but the accuracy is relatively low. After years of continuous development of machine learning technology and deep learning methods, the automatic extraction and classification of features has become a new direction for artificial intelligence [6][7][8][9]. Random forest belongs to supervised classification and is representative of the Bagging model of integrated learning in machine learning technology [10].…”

Extraction of Lilium davidii var. unicolor Planting Information Based on Deep Learning and Multi-Source Data

“…The image data were downloaded from the Copernicus Open Access Hub (https://scihub.copernicus.eu/dhus/#/home) (accessed on 7 and 9 August 2021), and the Sentinel-1 image was preprocessed for orbit correction, thermal noise removal, radiometric calibration, and terrain correction using the SNAP (version number: 9.0) software from the European Space Agency to obtain VV and VH backscatter coefficient maps. The nearest-neighbor method [20] was adopted to resample the Sentinel-1/2 image to 10 m for cropping and embedding. DEM data from GDEMV3 with a resolution of 30 m from the Geospatial Data Cloud Platforms (https://www.gscloud.cn) (accessed on 7 August 2021) were resampled to 10 m for slope, aspect, and elevation extraction.…”

Collecting and Disseminating Ground Truth Data for Generation of Agriculture Statistics

“…Output from both approaches is a single-layer thematic image with discrete classes representing features from the input raster data. Modern machine learning-based approaches, such as random forest [17][18][19], and deep neural network structures (convolutional neural network [CNN], deep neural network [DNN], recurrent neural network [RNN]) [20][21][22], are widely used for image classification. Besides, the satellite imagery's spatial and temporal resolution should always be considered when performing image classification for agriculture mapping.…”