A parallel implementation of a proposed stereo vision algorithm for three-dimensional scene reconstruction is presented. The algorithm firstly estimates the disparity map of a scene from a pair of rectified stereo images using an adaptive template matched filter. Next, the estimated disparity is utilized to retrieve the three-dimensional information of the scene, by considering the stereo camera's intrinsic parameters. The proposed algorithm is implemented on an embedded graphics processing unit by exploiting massive parallelism for high-rate image processing. The performance of the proposed algorithm is evaluated in real-life scenes captured on a laboratory experimental platform in terms of accuracy and processing speed.
A real-time system for restoration of images degraded by haze is presented. First, a transmission function estimator is automatically constructed using genetic programming. Next, the resultant estimator is employed to compute the transmission function of the scene by processing an input hazy image. Finally, the estimated transmission function and the hazy image are used in a restoration model based on atmospheric optics to obtain a haze-free image. The proposed method is implemented in a laboratory prototype for high-rate image processing. The performance of the proposed approach is evaluated in terms of objective metrics using synthetic and real-world images.
Stereo matching is challenging due to the presence of perspective distortions and noise. Commonly, stereo matching algorithms utilize local matching techniques to determine the correspondence of two pixels of the same point in a scene. This work presents a stereo matching algorithm based on locally-adaptive windows and correlation filtering. The proposed algorithm estimates the disparity of each pixel by matching an adaptive sliding-window obtained from the left image in the right image of the stereo pair. Computer simulation results obtained with the proposed algorithm are analyzed and discussed by processing pairs of stereo images.
Image restoration consists in retrieving an original image by processing captured images of a scene which are degraded by noise, blurring or optical scattering. Commonly restoration algorithms utilize a single monocular image of the observed scene by assuming a known degradation model. In this approach, valuable information of the three dimensional scene is discarded. This work presents a locally-adaptive algorithm for image restoration by employing stereo vision. The proposed algorithm utilizes information of a three-dimensional scene as well as local image statistics to improve the quality of a single restored image by processing pairs of stereo images. Computer simulations results obtained with the proposed algorithm are analyzed and discussed in terms of objective metrics by processing stereo images degraded by optical scattering.
In image restoration problems it is commonly assumed that image degradations are linear. In real-life this assumption is not always satisfied causing linear restoration methods fail. In this work, we present the design of an image restoration filtering based on genetic programming. The proposed filtering is given by a secuence of basic mathematical operators that allows to retrieve an undegraded image from an image degraded with noise. Computer simulations results obtained with the proposed algorithm in terms of objective metrics are analyzed and discussed by processing images degraded with noise. The obtained results are compared with those obtained with existing linear filters.