Structure from Infrared Stereo Images

Hajebi, Kiana; Zelek, John

doi:10.1109/crv.2008.9

Cited by 25 publications

(21 citation statements)

References 70 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Among infrared sensors, LWIR spectrum has the largest difference compared with the visible spectrum, and considering their price advantage, uncooled LWIR cameras show great potential to enhance the robustness of visual odometry. Though economically accessible, uncooled cameras have their own shortcomings as mentioned in [12], among which the following three are worth to mention. (1) High Noise: there is fixed noise in the image due to camera self-emission.…”

Section: Related Workmentioning

confidence: 99%

“…The first type of methods treat the multi-spectral setups as a monocular sensor by ignoring the parallax and superimpose the LWIR image directly onto the visible image to obtain stereo correspondences [4], [19], [28]. However, not only do these methods introduce errors due to the parallax that shouldn't The IR images of a human head in different wavelengths [12]. The difference between the far infrared (LWIR) image and the near infrared (SWIR) image shows the independence of reflectance in the former.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Multi-Spectral Visual Odometry without Explicit Stereo Matching

Dai

Sun

Hovakimyan

et al. 2019

2019 International Conference on 3D Vision (3DV)

View full text Add to dashboard Cite

Multi-spectral sensors consisting of a standard (visible-light) camera and a long-wave infrared camera can simultaneously provide both visible and thermal images. Since thermal images are independent from environmental illumination, they can help to overcome certain limitations of standard cameras under complicated illumination conditions. However, due to the difference in the information source of the two types of cameras, their images usually share very low texture similarity. Hence, traditional texture-based feature matching methods cannot be directly applied to obtain stereo correspondences. To tackle this problem, a multi-spectral visual odometry method without explicit stereo matching is proposed in this paper. Bundle adjustment of multi-view stereo is performed on the visible and the thermal images using direct image alignment. Scale drift can be avoided by additional temporal observations of map points with the fixed-baseline stereo.Experimental results indicate that the proposed method can provide accurate visual odometry results with recovered metric scale. Moreover, the proposed method can also provide a metric 3D reconstruction in semi-dense density with multi-spectral information, which is not available from existing multi-spectral methods.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Multi-Spectral Visual Odometry without Explicit Stereo Matching

Dai

Sun

Hovakimyan

et al. 2019

2019 International Conference on 3D Vision (3DV)

View full text Add to dashboard Cite

show abstract

“…Moreover, it gives an initial guess for the D 3 position of every point. For the first step of the method, a reasonable amount of stable and tractable matched points is obtained through a specific method for infrared images based on the phase congruency model [24,25] which is combined with more classical feature detectors. With few and low-textured infrared images, the result would be limited to a sparse D 3 reconstruction.…”

Section: Introductionmentioning

confidence: 99%

Automated thermal 3D reconstruction based on a robot equipped with uncalibrated infrared stereovision cameras

Sentenac

Bugarin

Ducarouge

et al. 2018

Advanced Engineering Informatics

View full text Add to dashboard Cite

In many industrial sectors, Non Destructive Testing (NDT) methods are used for the thermomechanical analysis of parts in assemblies of engines or reactors or for the control of metal forming processes. This article suggests an automated multi-view approach for the thermal D 3 reconstruction required in order to compute D 3 surface temperature models. This approach is based only on infrared cameras mounted on a Cartesian robot. The low resolution of these cameras associated to a lack of texture to infrared images require to use a global approach based first on an uncalibrated rectification and then on the simultaneous execution, in a single step, of the dense D 3 reconstruction and of an extended self-calibration. The uncalibrated rectification is based on an optimization process under constraints which calculates the homographies without prior calculation of the Fundamental Matrix and which minimizes the projective deformations between the initial images and the rectified ones. The extended self-calibration estimates both the parameters of virtual cameras that could provide the rectified images directly, and the parameters of the robot. It is based on two criteria evaluated according to the noise level of the infrared images. This global approach is validated through the reconstruction of a hot object against a reference reconstruction acquired by a D 3 scanner.

show abstract

“…It is also used in [6,7,23,24] in addition with elastic surface-net algorithm [26] or the marching cubes algorithm [27] to extract mesh. Kiana Hajebi presented in her work [28], a sparse disparity map reconstruction were produced based on features-based stereo matching technique from infrared images. A number of constraints (epipolar geometry) and assumptions (image brightness, proximity .…”

Section: Related Workmentioning

confidence: 99%

Hardware Implementation and Validation of 3D Underwater Shape Reconstruction Algorithm Using a Stereo-Catadioptric System

et al. 2016

View full text Add to dashboard Cite

Abstract:In this paper, we present a new stereo vision-based system and its efficient hardware implementation for real-time underwater environments exploration throughout 3D sparse reconstruction based on a number of feature points. The proposed underwater 3D shape reconstruction algorithm details are presented. The main concepts and advantages are discussed and comparison with existing systems is performed. In order to achieve real-time video constraints, a hardware implementation of the algorithm is performed using Xilinx System Generator. The pipelined stereo vision system has been implemented using Field Programmable Gate Arrays (FPGA) technology. Both timing constraints and mathematical operations precision have been evaluated in order to validate the proposed hardware implementation of our system. Experimental results show that the proposed system presents high accuracy and execution time performances.

show abstract

Structure from Infrared Stereo Images

Cited by 25 publications

References 70 publications

Multi-Spectral Visual Odometry without Explicit Stereo Matching

Multi-Spectral Visual Odometry without Explicit Stereo Matching

Automated thermal 3D reconstruction based on a robot equipped with uncalibrated infrared stereovision cameras

Hardware Implementation and Validation of 3D Underwater Shape Reconstruction Algorithm Using a Stereo-Catadioptric System

Contact Info

Product

Resources

About