CNN-based salient features in HSI image semantic target prediction

Srivastava, Vishal; Biswas, Bhaskar

doi:10.1080/09540091.2019.1650330

Cited by 26 publications

(8 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In future work, we will increase the weather categories and weather complexity (Srivastava and Biswas, 2019). The current model can call a decision model that has been trained in the corresponding weather after the weather changing, such as calling the SunnyDay model in a clear day and calling the RainyDusk model at the rainy dusk.…”

Section: Discussionmentioning

confidence: 99%

Unmanned surface vehicle adaptive decision model for changing weather

ZhangHan

WangXinzhi

LuoXiangfeng

et al. 2021

IJCSE

View full text Add to dashboard Cite

The autonomous decision-making capability of unmanned surface vehicles (USV) is the basis for many tasks. Most of the works ignore the variability of the scene. For example, traditional decision-making methods are not adaptable to changing weather that a USV is likely to encounter. In order to solve the low adaptability problem of a USV using single decision model in changing weather, we propose an adaptive model of USV based on human memory cognitive process. The USV first stores the perceived weather features in sensory memory. Then, it combines weather characteristics with prior knowledge to classify the weather in perceptual associative memory. Finally, USV calls different decision models stored in long-term memory based on the current weather category to make the decision. Simulated experiments are carried out on USV obstacle avoidance decision task in Unity3D. Experiments show that our model performs better than using only a single decision model.

show abstract

Section: Discussionmentioning

confidence: 99%

Unmanned surface vehicle adaptive decision model for changing weather

ZhangHan

WangXinzhi

LuoXiangfeng

et al. 2021

IJCSE

View full text Add to dashboard Cite

show abstract

“…Deep Learning (DL) has become prevalent in the last few years, including RS [32][33][34], due to its unprecedented success in many application fields. This success stems from DL's ability to learn discriminative representations directly from raw data [35], often giving up the pre-processing steps and handcrafted representations typically required in traditional Machine Learning methods. Regarding deforestation detection, novel DL-based approaches have been recently proposed for deforestation detection, most of them relying on optical imagery [36][37][38].…”

Section: Introductionmentioning

confidence: 99%

Improving Deforestation Detection on Tropical Rainforests Using Sentinel-1 Data and Convolutional Neural Networks

et al. 2022

View full text Add to dashboard Cite

Detecting early deforestation is a fundamental process in reducing forest degradation and carbon emissions. With this procedure, it is possible to monitor and control illegal activities associated with deforestation. Most regular monitoring projects have been recently proposed, but most of them rely on optical imagery. In addition, these data are seriously restricted by cloud coverage, especially in tropical environments. In this regard, Synthetic Aperture Radar (SAR) is an attractive alternative that can fill this observational gap. This work evaluated and compared a conventional method based on time series and a Fully Convolutional Network (FCN) with bi-temporal SAR images. These approaches were assessed in two regions of the Brazilian Amazon to detect deforestation between 2019 and 2020. Different pre-processing techniques, including filtering and stabilization stages, were applied to the C-band Sentinel-1 images. Furthermore, this study proposes to provide the network with the distance map to past-deforestation as additional information to the pair of images being compared. In our experiments, this proposal brought up to 4% improvement in average precision. The experimental results further indicated a clear superiority of the DL approach over a time series-based deforestation detection method used as a baseline in all experiments. Finally, the study proved the benefits of pre-processing techniques when using detection methods based on time series. On the contrary, the analysis revealed that the neural network could eliminate noise from the input images, making filtering innocuous and, therefore, unnecessary. On the other hand, the stabilization of the input images brought non-negligible accuracy gains to the DL approach.

show abstract

“…Consequently, with such high spectral resolution, many small constituents can now be exposed and extracted by hyperspectral imaging instruments with very fine spectral channels for recognition, mining, classification, identification, approximation and quantification. Many applications of remote sensing like agriculture and vegetation (Thenkabail et al, 2018), semantic target prediction (Srivastava and Biswas, 2020), water quality (Mishra et al, 2017), urban classification (Li et al, 2020), target detection are greatly improved by hyperspectral sensors.…”

Section: Introductionmentioning

confidence: 99%

Hyperspectral endmember extraction using Pearson's correlation coefficient

Shah

Zaveri

2021

IJCSE

View full text Add to dashboard Cite

Hyperspectral unmixing is a source separation problem. The spectral unmixing process is simply the composition of the three-step chain (subspace identification, endmember extraction and abundance estimation). A critical step in this chain is endmember extraction which finds endmembers from the image for the estimation of abundances. In this paper, a novel framework is proposed which uses the concept of Pearson's correlation coefficient and convex geometry. The novel framework extracts endmembers from the convex set of the two bands extracted using Pearson's correlation coefficient so it is named as Pearson's correlation coefficient-based convex geometry for endmember extraction (PCGE). This PCGE framework is different from other commonly used frameworks due to only two bands convex geometry. Due to only two bands, the computation time for the proposed framework is less. The proposed framework is applied to a synthetic dataset and four popular real hyperspectral datasets. In the simulation results, the proposed framework is compared with other popular frameworks based on standard evaluation parameters (spectral angle error, spectral information divergence and normalised cross-correlation). It has been observed from the simulation results that the proposed framework outperforms popular frameworks. It has been also observed that the proposed framework takes less time than others for extracting endmembers.

show abstract

CNN-based salient features in HSI image semantic target prediction

Cited by 26 publications

References 30 publications

Unmanned surface vehicle adaptive decision model for changing weather

Unmanned surface vehicle adaptive decision model for changing weather

Improving Deforestation Detection on Tropical Rainforests Using Sentinel-1 Data and Convolutional Neural Networks

Hyperspectral endmember extraction using Pearson's correlation coefficient

Contact Info

Product

Resources

About