Segment-before-Detect: Vehicle Detection and Classification through Semantic Segmentation of Aerial Images

Audebert, Nicolas; Saux, Bertrand Le; Lefèvre, Sébastien

doi:10.3390/rs9040368

Cited by 238 publications

(156 citation statements)

References 40 publications

Supporting

Mentioning

156

Contrasting

Order By: Relevance

“…Deep learning has been extensively used in the literature for a range of different applications such as vehicle detection [45,46], investigated avalanche search and rescue operations with Unmanned Areal Vehicles (UAV), change detection [47,48]. In this scheme, high level features are learned from low level ones where the features derived can be formulated for pattern recognition classification [49].…”

Section: Introductionmentioning

confidence: 99%

Hybrid Spectral Unmixing: Using Artificial Neural Networks for Linear/Non-Linear Switching

et al. 2017

View full text Add to dashboard Cite

Spectral unmixing is a key process in identifying spectral signature of materials and quantifying their spatial distribution over an image. The linear model is expected to provide acceptable results when two assumptions are satisfied: (1) The mixing process should occur at macroscopic level and (2) Photons must interact with single material before reaching the sensor. However, these assumptions do not always hold and more complex nonlinear models are required. This study proposes a new hybrid method for switching between linear and nonlinear spectral unmixing of hyperspectral data based on artificial neural networks. The neural networks was trained with parameters within a window of the pixel under consideration. These parameters are computed to represent the diversity of the neighboring pixels and are based on the Spectral Angular Distance, Covariance and a non linearity parameter. The endmembers were extracted using Vertex Component Analysis while the abundances were estimated using the method identified by the neural networks (Vertex Component Analysis, Fully Constraint Least Square Method, Polynomial Post Nonlinear Mixing Model or Generalized Bilinear Model). Results show that the hybrid method performs better than each of the individual techniques with high overall accuracy, while the abundance estimation error is significantly lower than that obtained using the individual methods. Experiments on both synthetic dataset and real hyperspectral images demonstrated that the proposed hybrid switch method is efficient for solving spectral unmixing of hyperspectral images as compared to individual algorithms.

show abstract

Section: Introductionmentioning

confidence: 99%

Hybrid Spectral Unmixing: Using Artificial Neural Networks for Linear/Non-Linear Switching

et al. 2017

View full text Add to dashboard Cite

show abstract

“…Here, two things will be done. The first, derive a practical law to calculate the area from the images taken by phantom plane (Audebert et al, 2017) at different altitudes. The second, adapting the measurements at variable altitudes to a polynomial ("Mathworks Corporations," 2017), classification of part(s) of an image as green land(s),and calculate the area automatically.…”

Section: Introductionmentioning

confidence: 99%

Multiple and Single Green Area Measurements and Classification Using Phantom Images in Comparison With Derived Experimental Law

Abu-Zaid¹

2017

ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

ABSTRACT:In many circumstances, it is difficult for humans to reach some areas, due to its topography, personal safety, or security regulations in the country. Governments and persons need to calculate those areas and classify the green parts for reclamation to benefit from it.To solve this problem, this research proposes to use a phantom air plane to capture a digital image for the targeted area, then use a segmentation algorithm to separate the green space and calculate it's area. It was necessary to deal with two problems. The first is the variable elevation at which an image was taken, which leads to a change in the physical area of each pixel. To overcome this problem a fourth degree polynomial was fit to some experimental data. The second problem was the existence of different unconnected pieces of green areas in a single image, but we might be interested only in one of them. To solve this problem, the probability of classifying the targeted area as green was increased, while the probability of other untargeted sections was decreased by the inclusion of parts of it as non-green. A practical law was also devised to measure the target area in the digital image for comparison purposes with practical measurements and the polynomial fit.

show abstract

“…To compare the performance of various approaches developed for object detection in remote sensing images, many datasets are available for researchers to conduct further investigations [3,[28][29][30][31]. These datasets promote the development of object detection methods in remote sensing imagery, but have obvious drawbacks.…”

Section: Dataset and Implementation Detailsmentioning

confidence: 99%

Deformable ConvNet with Aspect Ratio Constrained NMS for Object Detection in Remote Sensing Imagery

Wang

et al. 2017

Remote Sensing

108

View full text Add to dashboard Cite

Convolutional neural networks (CNNs) have demonstrated their ability object detection of very high resolution remote sensing images. However, CNNs have obvious limitations for modeling geometric variations in remote sensing targets. In this paper, we introduced a CNN structure, namely deformable ConvNet, to address geometric modeling in object recognition. By adding offsets to the convolution layers, feature mapping of CNN can be applied to unfixed locations, enhancing CNNs' visual appearance understanding. In our work, a deformable region-based fully convolutional networks (R-FCN) was constructed by substituting the regular convolution layer with a deformable convolution layer. To efficiently use this deformable convolutional neural network (ConvNet), a training mechanism is developed in our work. We first set the pre-trained R-FCN natural image model as the default network parameters in deformable R-FCN. Then, this deformable ConvNet was fine-tuned on very high resolution (VHR) remote sensing images. To remedy the increase in lines like false region proposals, we developed aspect ratio constrained non maximum suppression (arcNMS). The precision of deformable ConvNet for detecting objects was then improved. An end-to-end approach was then developed by combining deformable R-FCN, a smart fine-tuning strategy and aspect ratio constrained NMS. The developed method was better than a state-of-the-art benchmark in object detection without data augmentation.

show abstract

Segment-before-Detect: Vehicle Detection and Classification through Semantic Segmentation of Aerial Images

Cited by 238 publications

References 40 publications

Hybrid Spectral Unmixing: Using Artificial Neural Networks for Linear/Non-Linear Switching

Hybrid Spectral Unmixing: Using Artificial Neural Networks for Linear/Non-Linear Switching

Multiple and Single Green Area Measurements and Classification Using Phantom Images in Comparison With Derived Experimental Law

Deformable ConvNet with Aspect Ratio Constrained NMS for Object Detection in Remote Sensing Imagery

Contact Info

Product

Resources

About