Local climate zone mapping as remote sensing scene classification using deep learning: A case study of metropolitan China

Liu, Shengjie; Shi, Qian

doi:10.1016/j.isprsjprs.2020.04.008

Cited by 111 publications

(60 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…SL methods usually have only one hidden layer or no hidden layers; their classifiers have a simple structure and fast running speed and can be trained in a short time [34], [52]- [53]. In this study, SL algorithms with little or no parameters (e.g.…”

Section: B Sl Layer Constructionmentioning

confidence: 99%

“…The voting method obtains results via weighting or simple vote the outputs of base classifiers; some popular MCSs, including AdaBoost (AB), rotation forest, random forest (RF) and bagging (BA), are founded on such strategy [29]- [32]. However, collecting enough samples to evaluate the properties of base classifiers and set up a stable ensemble strategy is difficult due to the effect of natural [34]- [38]. It uses brain simulations to establish deep structures with multiple layers to extract high-level features progressively and solves complex classification problems.…”

Section: Introductionmentioning

confidence: 99%

“…Continuous learning from existing learnings (CLFEL) is a common thinking model and an effective way to improve learning efficiency. DL is priory for solving complex classification problems; however, most of them take RS images as material, and only a few recognises the idea of CLFEL [34]- [38]. MCS uses multiple learners, and some researches fused various DL networks to improve classification performance [43] [44].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Remote Sensing Image Classification Using Deep–Shallow Learning

Dou

Shen

et al. 2021

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Recently, classification using multiple classifier system (MCS) has been reported as an effective method to improve remote sensing (RS) image classification. Such systems provide a complementary mechanism to use multiple classifiers, which have shallow architecture to solve the same classification problem; however, the system exhibits shortcomings due to complex ensemble strategy. Deep learning (DL) has been proven to be an advanced method for complex data classification; however, how to use its advantages to overcome the shortcomings of MCS in ensemble strategy for classification accuracy improvement is worthy of study. Thus, with the multiple classifier mechanism and DL architecture, we propose a novel RS image classification framework, namely, deep-shallow learning (DSL), to improve classification accuracy. The DSL framework consists of a shallow learning (SL) layer and a DL layer. The SL layer contains various classifiers with shallow architecture, which can output different classification results for a certain input, whereas the DL layer is formed by DL networks, which can continue learning from the outputs of the SL layer. DSL simulates a human thinking model that continuously learns from the existing learnings to improve learning efficiency. In our experiment, three shallow classification algorithms, i.e. C4.5, k-nearest neighbour, and naive Bayesian, are used to train base classifiers in the SL layer, whereas a deep belief network (DBN) is used to train the DL layer. The experiment results on three different datasets indicate that DSL outperforms other methods in terms of classification accuracy by using backpropagation neural network, bagging, AdaBoost, random forest, multilayer perceptron and DBN.

show abstract

Section: B Sl Layer Constructionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Remote Sensing Image Classification Using Deep–Shallow Learning

Dou

Shen

et al. 2021

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Frame push-broom can obtain the images with larger width by changing the frame frequency of the camera, the overlap rate of adjacent frames of images can be controlled, which is of great significance for multi-temporal images matching, dynamic targets monitoring, etc. Meanwhile, the frame push-broom mode requires relatively low attitude control of the satellite, so we choose frame push-broom imaging mode as Luojia1-01 night-time light camera’s main imaging mode [18,19].…”

Section: Mission Analysismentioning

confidence: 99%

High Sensitive Night-time Light Imaging Camera Design and In-orbit Test of Luojia1-01 Satellite

Zhong

Zhang

et al. 2019

Sensors

View full text Add to dashboard Cite

Luojia1-01 satellite is the first scientific experimental satellite applied for night-time light remote sensing data acquisition, and the payload is an optical camera with high sensitivity, high radiation measurement accuracy and stable elements of interior orientation. At the same time, a special shaped hood is designed, which significantly improved the ability of the camera to suppress stray light. Camera electronics adopts the integrated design of focal plane and imaging processing, which greatly reduces the volume and weight of the system. In this paper, the design of the optical camera is summarized, and the results of in-orbit imaging performance tests are analyzed. The results show that the dynamic modulation transfer function (MTF) of the camera is better than 0.17, and the SNR is better than 35 dB under the condition of 10 lx illuminance and 0.3 reflectivity and all indicators meet the design requirements. The data obtained have been widely applied in many fields such as the process of urbanization, light pollution analysis, marine fisheries detection and military.

show abstract

“…In the deep learning era, the novel convolutional neural networks (CNNs), recurrent neural networks and graph convolutional networks started to dominate the classification of hyperspectral and remote sensing data [9], [28]- [34]. Manual feature engineering was replaced by automatic feature learning by deep networks.…”

Section: Introductionmentioning

confidence: 99%

Few-Shot Hyperspectral Image Classification With Unknown Classes Using Multitask Deep Learning

Liu

Shi

Zhang

2021

IEEE Trans. Geosci. Remote Sensing

Self Cite

195

View full text Add to dashboard Cite

Current hyperspectral image classification assumes that a predefined classification system is closed and complete, and there are no unknown or novel classes in the unseen data. However, this assumption may be too strict for the real world. Often, novel classes are overlooked when the classification system is constructed. The closed nature forces a model to assign a label given a new sample and may lead to overestimation of known land covers (e.g., crop area). To tackle this issue, we propose a multitask deep learning method that simultaneously conducts classification and reconstruction in the open world (named MDL4OW) where unknown classes may exist. The reconstructed data are compared with the original data; those failing to be reconstructed are considered unknown, based on the assumption that they are not well represented in the latent features due to the lack of labels. A threshold needs to be defined to separate the unknown and known classes; we propose two strategies based on the extreme value theory for few-shot and manyshot scenarios. The proposed method was tested on real-world hyperspectral images; state-of-the-art results were achieved, e.g., improving the overall accuracy by 4.94% for the Salinas data. By considering the existence of unknown classes in the open world, our method achieved more accurate hyperspectral image classification, especially under the few-shot context.

show abstract

Local climate zone mapping as remote sensing scene classification using deep learning: A case study of metropolitan China

Cited by 111 publications

References 57 publications

Remote Sensing Image Classification Using Deep–Shallow Learning

Remote Sensing Image Classification Using Deep–Shallow Learning

High Sensitive Night-time Light Imaging Camera Design and In-orbit Test of Luojia1-01 Satellite

Few-Shot Hyperspectral Image Classification With Unknown Classes Using Multitask Deep Learning

Contact Info

Product

Resources

About