Digital Holography as Computer Vision Position Sensor with an Extended Range of Working Distances.

ABSTRACT

Standard computer vision methods are usually based on powerful contact-less measurement approaches but applications, especially at the micro-scale, are restricted by finite depth-of-field and fixed working distance of imaging devices. Digital holography is a lensless, indirect imaging method recording the optical wave diffracted by the object onto the image sensor. The object is reconstructed numerically by propagating the recorded wavefront backward. The object distance becomes a computation parameter that can be chosen arbitrarily and adjusted to match the object position. No refractive lens is used and usual depth-of-field and working distance limitations are replaced by less restrictive ones tied to the laser-source coherence-length and to the size and resolution of the camera sensor. This paper applies digital holography to artificial visual in-plane position sensing with an extra-large range-to-resolution ratio. The object is made of a pseudoperiodic pattern allowing a subpixel resolution as well as a supra field-of-observation displacement range. We demonstrate an in-plane resolution of 50 nm and 0.002deg. in X, Y and θ respectively, over a working distance range of more than 15 cm. The allowed workspace extends over 12×10×150mm3. Digital holography extends the field of application of computer vision by allowing an extra-large range of working distances inaccessible to refractive imaging systems.