Method and apparatus for binocular ranging

An apparatus for binocular ranging, comprising:

• a feature extraction unit configured to extract features from a left image and a right image captured by a binocular camera to obtain a left feature image and a right feature image;
• a cost volume generation unit configured to select one of the left feature image and the right feature image as a standard feature image and obtain a cost volume of the standard feature image;
• a confidence volume generation unit configured to obtain a confidence volume;
• a confidence map generation unit configured to obtain a confidence map by selecting a maximum value from confidence levels of all disparity values in a disparity dimension for each pixel point in the confidence volume;
• a mask map generation unit configured to obtain a mask map by mapping each pixel point having a confidence level higher than a predetermined threshold in the confidence map to 1 and mapping each pixel point having a confidence level lower than or equal to the predetermined threshold in the confidence map to 0;
• a disparity map generation unit configured to obtain a disparity map by selecting a sequence number of a maximum confidence level in the disparity dimension for the confidence levels of all disparity values in the disparity dimension for each pixel point in the confidence volume;
• a target disparity map generation unit configured to obtain a target disparity map by multiplying the mask map with the disparity map;
• and
• a ranging unit configured to estimate a distance based on the target disparity map.

An apparatus for binocular ranging, comprising:

• a feature extraction unit configured to extract features from a left image and a right image captured by a binocular camera to obtain a left feature image and a right feature image;
• a first target disparity map generation unit configured to generate a first target disparity map by using a target disparity map generation unit with the left feature image being a standard feature image and the right feature image being a reference feature image;
• a second target disparity map generation unit configured to generate a second target disparity map by using the target disparity map generation unit with the right feature image being the standard feature image and the left feature image being the reference feature image;
• a third target disparity map generation unit configured to select one of the first target disparity map and the second target disparity map as a standard target disparity map and the other one as a reference target disparity map and obtain a third target disparity map by filtering out each pixel point in the standard target disparity map that is inconsistent with its corresponding pixel point in the reference target disparity map;
• a ranging unit configured to estimate a distance based on the third target disparity map;
• and
• the target disparity map generation unit configured to generate the target disparity map, the target disparity map generation unit comprising:
• a cost volume generation sub-unit configured to select one of the left feature image and the right feature image as the standard feature image and obtain a cost volume of the standard feature image;
• and
• a confidence volume generation sub-unit configured to obtain a confidence map.