Code release for 360monodepth. With our framework we achieve monocular depth estimation for high resolution 360° images based on aligning and blending perspective depth maps.
Thank you for your great work.when I use depth map and rgb to genetate point clouds,in one scene,there are many points,how I process all point clouds to genetate a Integral point clound.In the website, I see a example, x = np.array([0, 0, 0)
x_prime = r @ x + c,but the final point cloud is chaotic.
Make sure that the rotation and C you are using refer to cam2world rotation and translation. Otherwise, it could be that the depthmap looks great but it is not geometrically accurate
Thank you for your great work.when I use depth map and rgb to genetate point clouds,in one scene,there are many points,how I process all point clouds to genetate a Integral point clound.In the website, I see a example, x = np.array([0, 0, 0) x_prime = r @ x + c,but the final point cloud is chaotic.