Projection ray intersecting location-based multicolour pseudo-random coded projected active vision method

3D reconstruction has been recognised as an important cue for tasks like active computer vision and image understanding. However, most existing techniques need to calibrate the intrinsic and extrinsic parameters of the projector, making the calculation very complex. This paper presents an active vision method of multicolour pseudo-random encoded projected for recovering the 3D shape of surface from one calibrated image, based on projection ray intersecting location. This method combines geometric and photometric information in order to reconstruct 3D shape only calibrating camera parameters and projection rays. Based on the thought of ray intersecting location, by projecting pseudo-random coded pattern onto surface of object, using a projected ray of a feature point at projector side, and an imaging ray of the same feature point received by camera, the 3D reconstruction is implemented by seeking the intersection point of the projected ray and the imaging ray. We show a number of examples to demonstrate the accuracy of the method. Experimental results on the real world images show that the proposed method reconstructs the 3D shape of objects very efficiently.