A Systematic Review of the Use of Deep Learning in Satellite Imagery for Agriculture

2022-10-03

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

AI Summary

Deep learning (DL) applications in agricultural satellite image analysis are hindered by scarce labeled data, ambiguous task alignment, and insufficient exploitation of spatial characteristics. Method: We systematically review 193 peer-reviewed studies and propose (i) an input-modality taxonomy for remote sensing imagery; (ii) the first DL application atlas tailored to agricultural remote sensing tasks; and (iii) a data-gap diagnostic framework. Contribution/Results: We classify 12 agricultural remote sensing tasks by DL applicability, identifying seven—particularly crop monitoring and yield prediction—as critically constrained by high-quality annotated data scarcity. We introduce novel data curation strategies, including multi-source collaborative annotation and temporal consistency enhancement. Furthermore, we establish a task–algorithm–data triadic mapping to standardize and enhance reproducibility in agricultural remote sensing DL research.

Abstract

Agricultural research is essential for increasing food production to meet the needs of a rapidly growing human population. Collecting large quantities of agricultural data helps to improve decision-making for better food security at various levels: From international trade and policy decisions, down to individual farmers. At the same time, deep learning has seen a wave of popularity across many different research areas and data modalities. Satellite imagery has become available in unprecedented quantities, driving much research from the wider remote sensing community. The data hungry nature of deep learning models and this huge data volume seem like a perfect match. But has deep learning been adopted for agricultural tasks using satellite images? This systematic review of 193 studies analyzes the tasks that have reaped benefits from deep learning algorithms, and those that have not. It was found that while land use/land cover research has embraced deep learning algorithms, research on other agricultural tasks has not. This poor adoption appears to be due to a critical lack of labeled datasets for these other tasks. Thus, we give suggestions for collecting larger datasets. In addition, satellite images differ from ground-based images in a number of ways, resulting in a proliferation of interesting data interpretations unique to satellite images. So, this review also introduces a taxonomy of data input shapes and how they are interpreted in order to facilitate easier communication of algorithm types and enable quantitative analysis.

Problem

• Deep Learning
• Agricultural Image Analysis
• Spatial Characteristics

Innovation

• Deep Learning
• Agricultural Satellite Imagery
• Spatial Resolution Utilization