A stereo matching method estimates the disparity value between two correspondences in both stereo images. The disparity value represents the depth information of objects obtained from stereo images which have two different viewpoints. In many papers, the stereo matching method is tested under limited disparity conditions. These conditions follow configurations of test images. However, the disparity range in real applications is not known for those conditions and configurations. In this case, we have to check all pixels in the scan line to find the correspondence. Therefore, it is a time consuming task. Thus, we propose a fast disparity estimation method using the limited search range. The proposed method limits the disparity search range with the fast motion-search algorithm and local image characteristics.
Stereo Matching algorithms reconstruct a depth map from a pair of stereoscopic images. Stereo Matching algorithms are computationally intensive. Implementing efficient stereo matching algorithms on embedded systems is very challenging. This paper compares implementation efficiency and output quality of the state of the art dense stereo matching algorithms on the same multicore embedded system. The three different classes of stereo matching algorithms are local methods, semi-global methods and global methods. This paper compares three algorithms of the literature with a good trade-off between complexity and accuracy : Bilateral Filtering Aggregation (BFA, Local Method), One Dimension Belief Propagation (BP-1D, Semi Global Methods) and Semi Global Matching (SGM, Semi Global Methods). For the same input data the BFA, BP-1D and SGM were fully optimized and parallelized on the C6678 platform and run at respectively 10.7 ms, 4.1 ms and 47.1 ms.
In this paper, we propose a new method for accelerating stereo matching in autonomous vehicles using an upright pinhole camera model. It is motivated by that stereo videos are more restricted when the camera is fixed on the vehicles driving on the road. Assuming that the imaging plane is perpendicular to the road and the road is generally flat, we can derive the current disparity based on the previous one and the flow. The prediction is very efficient that only requires two multiplications per pixel. In practice, this model may not hold strictly but we still can use it for disparity initialization. Results on real datasets demonstrate the our method reduces the disparity search range from 128 to 61 with only slightly accuracy decreasing.
Stereo matching methods estimate a disparity value of the object as depth information. In general, most stereo matching methods are tested under ideal radiometric conditions. However, those ideal conditions cannot exist in real life. Adaptive normalized cross correlation (ANCC) is a method that is robust to radiometric variation. It estimates significantly accurate disparity values in the illumination variant condition. However, it has a high complexity problem in the cost computation because of the block matching-based method and the bilateral filtering process. In this paper, we propose a pixel-based ANCC using hue and gradient information to improve the computation complexity problem. The results show that the cost computation time is reduced even though error rates corresponding to the exposure and illumination changes have larger variations than those of the ANCC result.