Target object fast ranging method based on binocular vision

Abstract

The invention provides a target object fast ranging method based on binocular vision, comprising the following steps: (1) performing camera calibration; (2) acquiring images and detecting a target object; (3) calculating a search area; (4) selecting a start position of searching; (5) extracting grayscale features of the central region of the target object; (6) calculating a matched degree threshold; (7) performing searching for matching; and (8) calculating the parallax and the distance. According to the position of the target object in one picture, the method quickly searches for the matchedtarget object in another picture, which reduces the computational complexity and effectively improves the ranging speed.

Description

Technical field [0001] The invention relates to the field of computer vision, in particular to a method for fast ranging of a target object based on binocular vision. Background technique [0002] Binocular vision is widely used in robotics, aviation, military, medical and industrial fields. Binocular vision can calculate the actual distance of the target object by simulating human vision principles: first use the camera to obtain two pictures of the same target object at different angles; then find the relationship between the two picture pixels to match, and calculate the matching pixels To obtain the disparity value of the target object, finally use the disparity value of the target object to calculate the actual distance between the target object and the camera according to the principle of similar triangles. The specific principle refers to "Learning OpenCV" (Bradski G, Kaebler A, Yu Shiqi, Liu Ruizhen translation. Tsinghua University Press, 2009). [0003] Current matching ...

AI Technical Summary

Problems solved by technology

If the window is too small to include the feature points in the "search window", there will be a mismatch

Images