A New Model for Transfer Learning-Based Mapping of Burn Severity

Zheng, Zhong; Wang, Jinfei; He, Yongjun; Liao, Chunhua; Gao, Yanghua; Yang, Shiqi

doi:10.3390/rs12040708

Cited by 12 publications

(6 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These techniques compute the features from pre and post-fire acquisitions to determine the severity of burned areas. Machine learning methods are also adopted to predict the damage severity using pre and post-wildfire satellite imagery such as Support Vector Regressor (SVR) [82] and Random Forest [83].…”

Section: Post-fire Mapping Based Satellite Remote Sensing Imagerymentioning

confidence: 99%

Wildfire Segmentation Using Deep Vision Transformers

Ghali

Akhloufi

Jmal³

et al. 2021

Remote Sensing

View full text Add to dashboard Cite

In this paper, we address the problem of forest fires’ early detection and segmentation in order to predict their spread and help with fire fighting. Techniques based on Convolutional Networks are the most used and have proven to be efficient at solving such a problem. However, they remain limited in modeling the long-range relationship between objects in the image, due to the intrinsic locality of convolution operators. In order to overcome this drawback, Transformers, designed for sequence-to-sequence prediction, have emerged as alternative architectures. They have recently been used to determine the global dependencies between input and output sequences using the self-attention mechanism. In this context, we present in this work the very first study, which explores the potential of vision Transformers in the context of forest fire segmentation. Two vision-based Transformers are used, TransUNet and MedT. Thus, we design two frameworks based on the former image Transformers adapted to our complex, non-structured environment, which we evaluate using varying backbones and we optimize for forest fires’ segmentation. Extensive evaluations of both frameworks revealed a performance superior to current methods. The proposed approaches achieved a state-of-the-art performance with an F1-score of 97.7% for TransUNet architecture and 96.0% for MedT architecture. The analysis of the results showed that these models reduce fire pixels mis-classifications thanks to the extraction of both global and local features, which provide finer detection of the fire’s shape.

show abstract

Section: Post-fire Mapping Based Satellite Remote Sensing Imagerymentioning

confidence: 99%

Wildfire Segmentation Using Deep Vision Transformers

Ghali

Akhloufi

Jmal³

et al. 2021

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…e geographical distribution of re susceptibility was an urgent need to implement adequate management actions [3][4][5][6]. It was of importance to perform studies on modeling the susceptibility of forest res [7,8].…”

Section: Introductionmentioning

confidence: 99%

Modeling the Susceptibility of Forest Fires Using a Genetic Algorithm: A Case Study in Mountain Areas of Southwestern China

Zheng

Gao

Zhang

et al. 2022

Scientific Programming

Self Cite

View full text Add to dashboard Cite

Modeling fire susceptibility in fire-prone areas of forest ecosystems was essential for providing guidance to implement prevention and control measures of forest fires. Traditional models were developed on the basis of random selection of absence data (i.e., nonfire data from unburned areas), which could bring uncertainties to modeling results. Here, a new model with the genetic algorithm for Rule-set Production (GARP) algorithm and 10 environmental layers was proposed to process presence-only data in the susceptibility modeling of forest fires in Chongqing city. To do this, 70% of 684 fire occurrence data (479) during the period of 2000–2018 were applied to train the proposed model. And, 30% of these fire occurrence data (205) and the same amount of no-fire data (205) were emerged as validation dataset. The results showed that, for some environmental layers (i.e., distance to the nearest road, land cover, precipitation, distance to the nearest settlement, aspect, relative humidity, elevation, wind speed, and temperature), their P values were less than 0.05, indicating that these 9 environmental layers have significant influence on the spatial distribution of fire susceptibility in Chongqing city. On the contrary, with a higher P value (i.e., 0.126), the slope layer has an insignificant effect on fire susceptibility in the study area. Furthermore, the results of receiver operating characteristic analysis (ROC) showed that the proposed model has a good performance with an AUC value of 0.869, an accuracy value of 0.732, a sensitivity value of 0.59, a specificity value of 0.873, a positive predictive value of 0.823, and a negative predictive value of 0.681. This study revealed the validity of the proposed model in modeling the susceptibility of forest fires.

show abstract

“…However, forest images show large intra-class variation and often show significant spectral resemblance to non-forest images. Thus, annotating forest images requires domain expertise, impeding the acquisition of large scale labeled datasets and application of supervised methods in forest mapping [2].…”

Section: Introductionmentioning

confidence: 99%

Progressive Unsupervised Deep Transfer Learning for Forest Mapping in Satellite Image

Nouman¹,

Saha

Shahzad

et al. 2021

2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW)

View full text Add to dashboard Cite

Automated forest mapping is important to understand our forests that play a key role in ecological system. However, efforts towards forest mapping is impeded by difficulty to collect labeled forest images that show large intraclass variation. Recently unsupervised learning has shown promising capability when exploiting limited labeled data. Motivated by this, we propose a progressive unsupervised deep transfer learning method for forest mapping. The proposed method exploits a pre-trained model that is subsequently fine-tuned over the target forest domain. We propose two different fine-tuning mechanism, one works in a totally unsupervised setting by jointly learning the parameters of CNN and the k-means based cluster assignments of the resulting features and the other one works in a semi-supervised setting by exploiting the extracted knearest neighbor based pseudo labels. The proposed progressive scheme is evaluated on publicly available EuroSAT dataset using the relevant base model trained on BigEarth-Net labels. The results show that the proposed method greatly improves the forest regions classification accuracy as compared to the unsupervised baseline, nearly approaching the supervised classification approach.

show abstract

A New Model for Transfer Learning-Based Mapping of Burn Severity

Cited by 12 publications

References 43 publications

Wildfire Segmentation Using Deep Vision Transformers

Wildfire Segmentation Using Deep Vision Transformers

Modeling the Susceptibility of Forest Fires Using a Genetic Algorithm: A Case Study in Mountain Areas of Southwestern China

Progressive Unsupervised Deep Transfer Learning for Forest Mapping in Satellite Image

Contact Info

Product

Resources

About