MixChannel: Advanced Augmentation for Multispectral Satellite Images

Article Environmental Sciences

Semantic segmentation of land cover from high resolution multispectral satellite images by spectral-spatial convolutional neural network

Ekrem Saralioglu et al.

Summary: The study utilized a 3D-2D CNN model to extract both spectral and spatial information from high-resolution satellite images for more accurate land cover classification, achieving a classification accuracy of 95.6%.

GEOCARTO INTERNATIONAL (2022)

Add to Collection

Article Environmental Sciences

Mapping Seasonal Agricultural Land Use Types Using Deep Learning on Sentinel-2 Image Time Series

Misganu Debella-Gilo et al.

Summary: The study utilized Sentinel-2 satellite imagery to map agricultural land in Norway into three classes: cereal crops, fodder crops, and unused areas. Results showed that the 2D CNN model achieved up to 94% overall accuracy when applied on the SITS data with a temporal resolution of 7 days.

REMOTE SENSING (2021)

Add to Collection

Article Geography, Physical

Learning deep semantic segmentation network under multiple weakly-supervised constraints for cross-domain remote sensing image semantic segmentation

Yansheng Li et al.

Summary: Due to the challenges of data shift in practical remote sensing applications, this paper proposes a method to learn DSSN for cross-domain RS image semantic segmentation using multiple weakly-supervised constraints. The research includes multiple weakly-supervised constraints, which consist of transfer invariant constraint, pseudo-label constraint, and rotation consistency constraint. Experiments demonstrate that the proposed method performs remarkably well in the cross-domain RS image semantic segmentation task.

ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING (2021)

Add to Collection

Article Engineering, Electrical & Electronic

Image Compression and Plants Classification Using Machine Learning in Controlled-Environment Agriculture: Antarctic Station Use Case

Sergey Nesteruk et al.

Summary: This article shares the experience of controlled-environment agriculture automation in the Antarctic station greenhouse facility EDEN ISS, focusing on plant classification and image compression mechanism for remote plant monitoring. By utilizing machine learning, the approach achieves 92.6% accuracy in plant classification on an unbalanced dataset, providing valuable insights for various agriculture applications including plant disease identification.

IEEE SENSORS JOURNAL (2021)

Add to Collection

Article Computer Science, Artificial Intelligence

A survey: Deep learning for hyperspectral image classification with few labeled samples

Sen Jia et al.

Summary: With the advancement of deep learning technology, there is a growing focus on building deep learning models for HSI classification with few labeled samples, utilizing methods such as transfer learning, active learning, and few-shot learning. The results of the research indicate that the fusion of deep learning methods and related techniques can effectively address the issues of small-sample sets in HSI classification.

NEUROCOMPUTING (2021)

Add to Collection

Article Engineering, Electrical & Electronic

Neural-Based Hierarchical Approach for Detailed Dominant Forest Species Classification by Multispectral Satellite Imagery

Svetlana Illarionova et al.

Summary: By utilizing multispectral satellite imagery and neural networks, the forest classification problem was addressed as an image segmentation task, represented as a hierarchical set of binary classification tasks, to achieve better results.

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING (2021)

Add to Collection

Review Environmental Sciences