Mapping Plastic Greenhouses with Two-Temporal Sentinel-2 Images and 1D-CNN Deep Learning

Abstract

Plastic greenhouses (PGs) are widely built near cities in China to produce vegetables and fruits. In order to promote sustainable agriculture, rural landscape construction, and better manage water resources, numerous remote sensing methods have been developed to identify and monitor the distribution of PGs, of which many map PGs based on spectral responses and geometric shapes. In this study, we proposed a new fine- and coarse-scale mapping approach using two-temporal Sentinel-2 images with various seasonal characteristics and a one-dimensional convolutional neural network (1D-CNN). Having applied this approach in a pilot area study, the results were summarized as follows: (1) A time-series analysis of Sentinel-2 images showed that the reflectance of greenhouses changes during crop growth and development. In particular, the red-edge and near-infrared bands undergo a significant increase and then decrease during the whole crop growth period. Thus, two critical period images, containing a substantial difference in greenhouse reflectance, were sufficient to carry out an accurate and efficient mapping result. (2) The 1D-CNN classifier was used to map greenhouses by capturing subtle details and the overall trend of the spectrum curve. Overall, our approach showed higher classification accuracy than other approaches using support vector machines (SVM) or random forests (RF). In addition, the greenhouse area identified was highly consistent with the existing surfaces observed in very high-resolution images, with a kappa co-efficient of 0.81. (3) The narrow band feature differences (red-edge and near infrared narrow bands) in two-temporal Sentinel-2 images played a significant role in high-precision greenhouse mapping. The classification accuracy with narrow band features was much better than the maps produced without narrow band features. This scheme provided a method to digitize greenhouse precisely and publish its statistics for free, which enable advanced decision support for agriculture management.

Sentinel-2 Recognition of Uncovered and Plastic Covered Agricultural Soil

“…Temporary covers in the considered period were limited to scattered parcels spread across the region, requiring high precision and sensitivity of the detection method. This setting has rarely been addressed in literature as typical study areas are extensively covered [27,64,65]. Research has emphasised the impact of diverse aspects, such as underlying soil or vegetation, the presence of dew on the cover, and the type of plastic material [24,28].…”

Improving Soil Water Content and Surface Flux Estimation Based on Data Assimilation Technique

“…The rainfall is insufficient to support the winter wheat production, as about two-thirds of the annual rainfall occurs from July to September. Thus, agricultural production of the wheat-growing season relies primarily on surface ditch irrigation, with the irrigation water drawn mainly from the Yellow River [33].…”

Tracking Spatio-Temporal Dynamics of Greenhouse-Led Cultivated Land and its Drivers in Shandong Province, China

“…Meanwhile, a number of unsupervised methods based on the novel AG-extraction indices, such as the vegetable land extraction index (VI) (Zhao et al, 2004), moment distance index (MDI) (Aguilar et al, 2016), plastic-mulched landcover index (PMLI) (Lu et al, 2014), plastic greenhouses index (PGI) (Yang et al, 2017) and greenhouses detection index (GDI) (González-Yebra et al, 2018), have been proposed to distinguish GCL from other land use types. In order to improve the robustness of such unsupervised methods, previous studies also have adopted the supervised approaches, such as support vector machine (SVM) (Bektas Balcik et al, 2020), random forest (RF) (Lin et al, 2021), artificial neural network (ANN) (Carvajal et al, 2006) and convolutional neural network (CNN) (Sun et al, 2021), to extract the spatial distribution of GCL. Despite the fact that all of these researches performed well and produced a number of accurate GCL maps in various locations and years, only a few studies used the resulting GCL maps to detect spatio-temporal dynamics and driving forces of GCL (Arcidiacono and Porto, 2010;Picuno et al, 2011;Yu et al, 2017;Ou et al, 2020).…”