Segment-tree stereo matching algorithm based on improved matching costs

Matching cost aggregation is one of the oldest and still popular methods for stereo correspondence. While effective and efficient, cost aggregation methods typically aggregate the matching cost by summing/averaging over a user-specified, local support region. This is obviously only locally-optimal, and the computational complexity of the full-kernel implementation usually depends on the region size. In order to improve aggregation accuracy, we propose a segment-tree stereo matching method with improved matching costs. A reasonable weight for the matching process is assigned by introducing color and gradient multi-dimensional information components in order to overcome inaccuracies of weak texture regions. Next, similar regions are merged, whereby pixel points belonging to the same parallax consistency are merged with the corresponding generation tree. Finally, depth and color information are used for the tree reconstruction, while a color-depth weight is adopted in order to enhance the structure of the tree. Performance evaluation on 19 Middlebury data sets shows that the proposed method is comparable to previous state-of-the-art aggregation methods in disparity accuracy and processing speed. © International Association of Engineers.