Search

The scale and severity of forest disturbances across the globe are increasing due to climate change and human activities. Remote sensing analysis using time series data is a powerful approach for detecting large-scale forest disturbances and describing detailed forest dynamics. Various large-scale forest disturbance detection algorithms have been proposed, but most of them are only suitable for detecting high-magnitude forest disturbances (e.g., fire, harvest). Conversely, more continuous, subtle, and gradual lower-magnitude forest disturbances (e.g., thinning, pests, and diseases) have been subject to less focus. Deep learning (DL) can distinguish subtle differences in information within time series data, offering new opportunities to capture forest disturbances in a complete and detailed way. This study proposes an approach for analyzing forest dynamics across large areas and long time periods by combining DL time series classification and prior knowledge constraint. The approach consists of two stages: (1) an improved self-attention model used for time series classification to identify sequences with forest disturbance characteristics; (2) developed skip-disturbance recovery index (S-DRI) characterizing the temporal context, using prior knowledge constraint to identify forest disturbance years in time series with disturbance characteristics. In this study, the year of forest disturbances in five study areas located in the United States, Canada, and Poland from 2001 to 2020 was mapped. A total of 3082 manually interpreted test data with different disturbance causal agents (such as fire, harvest, conversion, hurricane, and pests) were sampled from five research areas for validation. Our approach was also evaluated against two forest disturbance benchmark datasets derived from LandTrendr and the Global Forest Change (GFC) dataset. The results demonstrate that our approach achieved an overall accuracy of 87.8%, surpassing the accuracy of LandTrendr (84.6%) and the Global Forest Change dataset (81.4%). Furthermore, our approach demonstrated lower omission rates (ranging from 10.0% to 67.4%) in detecting subtle to severe causal agents of forest disturbance, in comparison to LandTrendr (with a range of 18.0% to 81.6%) and GFC (with a range of 15.0% to 88.8%). This study, which involved mapping large-scale and long-term forest disturbance in multiple regions, revealed that our approach can be applied to new areas without a requirement for complex parameter adjustments. These results demonstrate the potential of our approach in generating comprehensive and detailed forest disturbance data, thus providing a new and effective method in this domain.

Multi-temporal remote sensing imagery has been regarded as an effective tool to monitor cropland. But optical sensors often miss key stages for crop growth because of clouds, which poses challenges to many studies. The synergistic of SAR and optical data is expected to lift this problem, especially in areas with persistent cloud cover. However, due to the different characteristics of optical and SAR sensors, it is difficult to build a relationship between the two with most existing methods, let alone construct the long-time correlations to fill optic observation gaps using SAR data. Inspired by deep learning, this study presents a novel strategy to learn the relationship between optical and SAR time series based on the sequence of contextual information. To be specific, we extended the conventional CNN-RNN to build Multi-CNN-Sequence to Sequence (MCNN-Seq) model, and formulate the correlation between the optic and SAR time series sequences. We verified the MCNN-Seq model and found that the accuracy of the predicted optical image was determined by crop types and phenological stages, both in the spatial and temporal domain, respectively. For several crops, such as onion, winter wheat, corn, and sugar beet, our predictions are fitting well with R2 0.9409, 0.9824,0.9157, and 0.9749, respectively. Compared to CNN and RNN, the simulation accuracy achieved by the MCNN-Seq model is much better in terms of R2 and RMSE. In general, results demonstrate that deep learning models have the potential to synergize SAR and optical data and provide replaceable information when the optical data has a long data gap due to the persistent clouds.

Surface albedo is the ratio of the light-flux reflected in all directions of the ground and the total flux incident. It is a very important surface parameters. In this paper, atmospheric correction algorithm based on dark pixel method was introduced, and corrected surface reflectance in study area was obtained after pre-processing of the CCD data of HJ-1 Satellite. The surface albedo of study area was combined by the inversion formula which created by Liang Shunlin to converts albedo from narrowband to broadband, and then a correlation analysis of the results and accuracy given. The result presents that: the retrieval accuracy is reliable basically.