Remote Sensing Applications in Sugarcane Cultivation: A Review

A large number of studies have been published addressing sugarcane management and monitoring to increase productivity and production as well as to better understand landscape dynamics and environmental threats. Building on existing reviews which mainly focused on the crop’s spectral behavior, a comprehensive review is provided which considers the progress made using novel data analysis techniques and improved data sources. To complement the available reviews, and to make the large body of research more easily accessible for both researchers and practitioners, in this review (i) we summarized remote sensing applications from 1981 to 2020, (ii) discussed key strengths and weaknesses of remote sensing approaches in the sugarcane context, and (iii) described the challenges and opportunities for future earth observation (EO)-based sugarcane monitoring and management. More than one hundred scientific studies were assessed regarding sugarcane mapping (52 papers), crop growth anomaly detection (11 papers), health monitoring (14 papers), and yield estimation (30 papers). The articles demonstrate that decametric satellite sensors such as Landsat and Sentinel-2 enable a reliable, cost-efficient, and timely mapping and monitoring of sugarcane by overcoming the ground sampling distance (GSD)-related limitations of coarser hectometric resolution data, while offering rich spectral information in the frequently recorded data. The Sentinel-2 constellation in particular provides fine spatial resolution at 10 m and high revisit frequency to support sugarcane management and other applications over large areas. For very small areas, and in particular for up-scaling and calibration purposes, unmanned aerial vehicles (UAV) are also useful. Multi-temporal and multi-source data, together with powerful machine learning approaches such as the random forest (RF) algorithm, are key to providing efficient monitoring and mapping of sugarcane growth, health, and yield. A number of difficulties for sugarcane monitoring and mapping were identified that are also well known for other crops. Those difficulties relate mainly to the often (i) time consuming pre-processing of optical time series to cope with atmospheric perturbations and cloud coverage, (ii) the still important lack of analysis-ready-data (ARD), (iii) the diversity of environmental and growth conditions—even for a given country—under which sugarcane is grown, superimposing non-crop related radiometric information on the observed sugarcane crop, and (iv) the general ill-posedness of retrieval and classification approaches which adds ambiguity to the derived information.

Download Full-text

Deep learning para la clasificación de usos de suelo agrícola con Sentinel-2

Revista de Teledetección ◽

10.4995/raet.2020.13337 ◽

2020 ◽

pp. 35

Author(s):

M. Campos-Taberner ◽

F.J. García-Haro ◽

B. Martínez ◽

M.A. Gilabert

Keyword(s):

Remote Sensing ◽

Deep Learning ◽

Near Infrared ◽

Spatial Information ◽

Learning Algorithm ◽

Support Vector ◽

Sensing Applications ◽

Learning Techniques ◽

Remote Sensing Applications ◽

Sentinel 2

<p class="p1">The use of deep learning techniques for remote sensing applications has recently increased. These algorithms have proven to be successful in estimation of parameters and classification of images. However, little effort has been made to make them understandable, leading to their implementation as “black boxes”. This work aims to evaluate the performance and clarify the operation of a deep learning algorithm, based on a bi-directional recurrent network of long short-term memory (2-BiLSTM). The land use classification in the Valencian Community based on Sentinel-2 image time series in the framework of the common agricultural policy (CAP) is used as an example. It is verified that the accuracy of the deep learning techniques is superior (98.6 % overall success) to that other algorithms such as decision trees (DT), k-nearest neighbors (k-NN), neural networks (NN), support vector machines (SVM) and random forests (RF). The performance of the classifier has been studied as a function of time and of the predictors used. It is concluded that, in the study area, the most relevant information used by the network in the classification are the images corresponding to summer and the spectral and spatial information derived from the red and near infrared bands. These results open the door to new studies in the field of the explainable deep learning in remote sensing applications.</p>

Download Full-text

THE USE OF DEEP LEARNING IN REMOTE SENSING FOR MAPPING IMPERVIOUS SURFACE: A REVIEW PAPER

ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences ◽

10.5194/isprs-archives-xlvi-4-w3-2021-199-2022 ◽

2022 ◽

Vol XLVI-4/W3-2021 ◽

pp. 199-203

Author(s):

S. Mahyoub ◽

H. Rhinane ◽

M. Mansour ◽

A. Fadil ◽

Y. Akensous ◽

...

Keyword(s):

Remote Sensing ◽

Deep Learning ◽

Urban Expansion ◽

Learning Approaches ◽

Deep Convolutional Neural Networks ◽

Comprehensive Overview ◽

Classification Image ◽

Sensing Applications ◽

Wide Range ◽

Remote Sensing Applications

Abstract. In recent years, deep convolutional neural networks (CNNs) algorithms have demonstrated outstanding performance in a wide range of remote sensing applications, including image classification, image detection, and image segmentation. Urban development, as defined by urban expansion, mapping impervious surfaces, and built-up areas, is one of these fascinating issues. The goal of this research is to explore at and summarize the deep learning approaches used in urbanization. In addition, several of these methods are highlighted in order to provide a comprehensive overview and comprehension of them, as well as their pros and downsides.

Download Full-text