The problem of automatic reliability monitoring and reliability-centered maintenance is increasingly important today. In this paper, we compare the accuracy of four machine learning approaches for fault detection in a...
The problem of automatic reliability monitoring and reliability-centered maintenance is increasingly important today. In this paper, we compare the accuracy of four machine learning approaches for fault detection in a hydraulic system. The first three approaches are based on SVM classifiers with linear, polynomial and RBF kernels and the last one is a gradient boosting on oblivious decision trees. We evaluate algorithms on the synthetic dataset generated by our simulation model of the helicopter hydraulic system and show that high accuracy fault detection can be achieved.
Modern hydraulic systems should be monitored on the regular basis. One of the most effective ways to address this task is utilizing in-line automatic particle counters (APC) built inside of the system. The measurement...
详细信息
ISBN:
(纸本)9781510611047;9781510611030
Modern hydraulic systems should be monitored on the regular basis. One of the most effective ways to address this task is utilizing in-line automatic particle counters (APC) built inside of the system. The measurement of particle concentration in hydraulic liquid by APC is crucial because increasing numbers of particles should mean functional problems. Existing automatic particle counters have significant limitation for the precise measurement of relatively low concentration of particle in aerospace systems or they are unable to measure higher concentration in industrial ones. Both issues can be addressed by implementation of the CMOS image sensor instead of single photodiode used in the most of APC. CMOS image sensor helps to overcome the problem of the errors in volume measurement caused by inequality of particle speed inside of tube. Correction is based on the determination of the particle position and parabolic velocity distribution profile. Proposed algorithms are also suitable for reducing the errors related to the particles matches in measurement volume. The results of simulation show that the accuracy increased up to 90 per cent and the resolution improved ten times more compared to the single photodiode sensor.
We investigate efficient sensitivity analysis (SA) of algorithms that segment and classify image features in a large dataset of high-resolution images. Algorithm SA is the process of evaluating variations of methods a...
详细信息
ISBN:
(纸本)9781538623268
We investigate efficient sensitivity analysis (SA) of algorithms that segment and classify image features in a large dataset of high-resolution images. Algorithm SA is the process of evaluating variations of methods and parameter values to quantify differences in the output. A SA can be very compute demanding because it requires re-processing the input dataset several times with different parameters to assess variations in output. In this work, we introduce strategies to efficiently speed up SA via runtime optimizations targeting distributed hybrid systems and reuse of computations from runs with different parameters. We evaluate our approach using a cancer image analysis workflow on a hybrid cluster with 256 nodes, each with an Intel Phi and a dual socket CPU. The SA attained a parallel efficiency of over 90% on 256 nodes. The cooperative execution using the CPUs and the Phi available in each node with smart task assignment strategies resulted in an additional speedup of about 2x. Finally, multi-level computation reuse lead to an additional speedup of up to 2.46x on the parallel version. The level of performance attained with the proposed optimizations will allow the use of SA in large-scale studies.
Lane detection and tracking is essential concern in vision based autonomous vehicle navigation. This paper proposes a novel method based on probabilistic Hough transform and motion vector based analysis for detecting ...
详细信息
ISBN:
(纸本)9781538673539;9781538673522
Lane detection and tracking is essential concern in vision based autonomous vehicle navigation. This paper proposes a novel method based on probabilistic Hough transform and motion vector based analysis for detecting and tracking lane and lane departures. It addresses drawbacks of current systems under hazy situations with learning method based on previous tracking records of the system. In this method relevant lane mark features are extracted based on color variations. Therefore used HSL color model which gives higher color contrast of road surface and the markings. In order to reduce computation, processingalgorithms are applied only to the region of interest (ROI). Edges of the selected regions are extracted with Canny Edge Detection. Since lane markings are made with set of straight lines, straight lines are extracted with Probabilistic Hough Transform. Extracted lines are analyzed to detect lane. To monitor lane departures vehicle's motion is detected and tracked based on motion vector analysis with Lucas Kanade Optical Flow algorithm. Previous detection records are used to track the lane continuously in hazy situations and provide prediction on approximate lane for the roads without lane markings. Experiments were performed on set of videos taken from vehicles' front camera mounted on dashboard. According to the experiments the proposed algorithm has 81.9% of sensitivity in lane detection and 63.5% in approximate lane predictions for the roads without lane markings. Lane detection algorithm has 83% precision, 68% recall and 0.75 F1 score with threshold brightness value 140. Algorithm can further improved to guide the autonomous vehicles to take accurate decisions on selecting safety lane to drive ahead in different situations.
Underwater vision is subject to effects of underwater light propagation that act to absorb, scatter, and attenuate light rays between the scene and the imaging platform. Backscattering has been shown to have a strong ...
详细信息
ISBN:
(纸本)9781538607336
Underwater vision is subject to effects of underwater light propagation that act to absorb, scatter, and attenuate light rays between the scene and the imaging platform. Backscattering has been shown to have a strong impact on underwater images. As light interacts with particulate matter in the water column, it is scattered back towards the imaging sensor, resulting in a hazy effect across the image. A similar effect occurs in terrestrial applications in images of foggy scenes due to interaction with the atmosphere. Prior work on multi-image dehazing has relied on multiple cameras, polarization filters, or moving light sources. Single image dehazing is an ill-posed problem;proposed solutions rely on strong priors of the scene. This paper presents a novel method for underwater image dehazing using a light field camera to capture both the spatial and angular distribution of light across a scene. First, a 2D dehazing method is applied to each sub-aperture image. These dehazed images are then combined to produce a smoothed central view. Lastly, the smoothed central view is used as a reference to perform guided image filtering, resulting in a 4D dehazed underwater light field image. The developed method is validated on real light field data collected in a controlled in-lab water tank, with images taken in air for reference. This dataset is made publicly available.
The recent success of deep learning has lead to growing interest in applying these methods to signal processing problems. This paper explores the applications of deep learning to synthetic aperture radar (SAR) image f...
详细信息
ISBN:
(数字)9781510609044
ISBN:
(纸本)9781510609037;9781510609044
The recent success of deep learning has lead to growing interest in applying these methods to signal processing problems. This paper explores the applications of deep learning to synthetic aperture radar (SAR) image formation. We review deep learning from a perspective relevant to SAR image formation. Our objective is to address SAR image formation in the presence of uncertainties in the SAR forward model. We present a recurrent auto-encoder network architecture based on the iterative shrinkage thresholding algorithm (ISTA) that incorporates SAR modeling. We then present an off-line training method using stochastic gradient descent and discuss the challenges and key steps of learning. Lastly, we show experimentally that our method can be used to form focused images in the presence of phase uncertainties. We demonstrate that the resulting algorithm has faster convergence and decreased reconstruction error than that of ISTA.
Real time non-rigid reconstruction pipelines are extremely computationally expensive and easily saturate the highest end GPUs currently available. This requires careful strategic choices to be made about a set of high...
详细信息
Real time non-rigid reconstruction pipelines are extremely computationally expensive and easily saturate the highest end GPUs currently available. This requires careful strategic choices to be made about a set of highly interconnected parameters that divide up the limited compute. At the same time, offline systems, prove the value of increasing voxel resolution, more iterations, and higher frame rates. To this end, we demonstrate a set of remarkably simple but effective modifications to these algorithms that significantly reduce the average per-frame computation cost allowing these parameters to be increased. Specifically, we divide the depth stream into sub-frames and fusion-frames, disabling both model accumulation (fusion) and non-rigid alignment (model tracking) on the former. Instead, we efficiently track point correspondences across neighboring sub-frames. We then leverage these correspondences to initialize the standard non-rigid alignment to a fusion-frame where data can then be accumulated into the model. As a result, compute resources in the modified non-rigid reconstruction pipeline can be immediately re-purposed to increase voxel resolution, use more iterations or to increase the frame rate. To demonstrate the latter, we leverage recent high frame rate depth algorithms to build a novel “twin” sensor consisting of a low-res/highfps sub-frame camera and a second low-fps/high-res fusion camera.
The intrinsic parameters of the camera are mainly composed of the focal length, coordinate of the main point and the tilting factor. The focal length is an important parameter for the camera to obtain the three-dimens...
详细信息
The intrinsic parameters of the camera are mainly composed of the focal length, coordinate of the main point and the tilting factor. The focal length is an important parameter for the camera to obtain the three-dimensional information from the two-dimensional image, which determines the shooting angle of the lens. It can be obtained by the camera calibration, such as the Zhang's calibration method, but the process is tedious and takes a long time. In this paper, a fast method to extract focal length of camera is proposed. To improve and simplify the camera self-calibration algorithm, parallel particle swarm optimization (PSO) algorithm is applied to solve the improved Kruppa equation, making the pole computation converted to the fitness function of cost function. In addition, in the process of PSO calculation, parallel computing experiments are carried out on the multi-core computer platform, and the advantages of parallel computing are fully used. The experimental results show that the precision of the method's focal length is close to Zhang's calibration method, and the time of the algorithm is greatly reduced compared with other calibration algorithms.
The vision characteristics-based on image detection techniques are functional method, which are appropriate for many industrial and agricultural applications. The color detection method is such a technique for non-des...
详细信息
ISBN:
(纸本)9783319495682;9783319495675
The vision characteristics-based on image detection techniques are functional method, which are appropriate for many industrial and agricultural applications. The color detection method is such a technique for non-destructive grading of the grape after the harvest process, which have been widely used in a variety of fruits' quality inspection and grading aspects based on vision computing method. However, the shapes of the entire bunch of grapes are obviously different with each other, which will influence the accuracy of non-destructive detection. This work develops a method for image detection and color grading of red and black grape samples, which uses vision computing technology to extract the effective color features of red and black grape samples, the experimental results show that the effective region extraction and color detection algorithms are feasible for color detection of grape.
The proceedings contain 27 papers. The topics discussed include: cognitively inspired artificial bipedal humanoid gait generation;personal identification using forearm vein patterns;k-same-net: neural-network-based fa...
ISBN:
(纸本)9781538608500
The proceedings contain 27 papers. The topics discussed include: cognitively inspired artificial bipedal humanoid gait generation;personal identification using forearm vein patterns;k-same-net: neural-network-based face deidentification;articulation acoustic kinematics in ALS speech: a longitudinal study case;voice pathology detection using deep learning: a preliminary study;analysis of disfluencies for automatic detection of mild cognitive impartment: a deep learning approach;automatic imageprocessing based dental image analysis using automatic Gaussian fitting energy and level sets;automatic event handling during robot assisted eye surgery;automatic identification of botanical samples of leaves using computer vision;approach of kinematic control for a nonholonomic wheeled robot using artificial neural networks and genetic algorithms;impact of target variable distribution type over the regression analysis in wind turbine data;image reconstruction of the human forearm by electrical impedance tomography;assessment of the spectral quality of fused images using the CIEDE2000 distance;a superpixel-based approach based on consensus for delineating agricultural plots;automated segmentation of powdery mildew disease from cherry leaves using imageprocessing;and control architecture to reproduce the knee and ankle movement using a transfemoral prosthesis.
暂无评论