A Flexible New Technique for Camera Calibration
IEEE Transactions on Pattern Analysis and Machine Intelligence
Distinctive Image Features from Scale-Invariant Keypoints
International Journal of Computer Vision
Stereo Localization Using Dual PTZ Cameras
CAIP '09 Proceedings of the 13th International Conference on Computer Analysis of Images and Patterns
Stereo Localization Based on Network's Uncalibrated Camera Pairs
AVSS '09 Proceedings of the 2009 Sixth IEEE International Conference on Advanced Video and Signal Based Surveillance
Stereo rectification of uncalibrated and heterogeneous images
Pattern Recognition Letters
Machine learning for high-speed corner detection
ECCV'06 Proceedings of the 9th European conference on Computer Vision - Volume Part I
| Hi-index | 0.00 |
To improve the localization accuracy and robustness of the moving 3D-target under the nature scenes, we propose a new target localization method through combining MSER (Maximally Stable Extremal Region) detector with SIFT (Scale Invariant Feature Transform) descriptor into the dual-PTZ-cameras stereo vision system. Firstly, stereo vision rectification is performed on the right-and-left images captured from the dual-PTZ-cameras with different focal lengths using designed Look-up-table(LUT )and BP neural network. Secondly, more high quality affine invariant features are extracted from the rectified images to perform initial matching using affine invariant feature detector and descriptor. Thirdly, erroneous correspondences is detected by RANSAC. Then, robust features matching under the multi-view-point and multi-focal-length is achieved. The localization experimental results of the moving 3-D target in a complex environment show that the proposed method has good localization accuracy and robustness.