There are provided a depth information acquisition method and device. The method includes: determining a relative geometric position relationship between a ToF camera and left and right cameras in a binocular camera, and internal parameters; collecting the depth map generated by the ToF camera and the images of two cameras; converting the depth map into a binocular disparity value between corresponding pixels in the images of the two cameras; mapping, by using the converted binocular disparity value, any pixel in the depth map generated by the ToF camera to corresponding pixel coordinates of the images of the two cameras to obtain a sparse disparity map; and performing calculation on all pixels in the depth map generated by the ToF camera to obtain a dense disparity map, thereby obtaining more accurate and denser depth information; or inversely calibrating collected depth map by the ToF camera with the sparse disparity map.