Optical EngineeringSampling of two-dimensional images: prevention from spectrum overlap and ghost detection
|Format||Member Price||Non-Member Price|
The aim of this work is to recall the key role of sampling theory in image sampling with 2-D solid-state image sensors. Sampling restrictions due to the actual sensor structure—especially in the case of noncontiguous sensitive pixels—are often neglected, although sampling principles are well known from a theoretical point of view. Our purpose is to present experimental results that demonstrate clearly what kind of ghost detection and spectrum overlapping may result if the spectral bandwidth of the incident image is not suited to the image sensor capabilities. These experimental results and image distortions are shown to be the direct consequences of the sampling principles that are recalled in the first part. Finally, a sequential acquisition procedure, based on image sensor scanning, is shown to fulfill sampling theory requirements and thus enlarges the bandwidth capabilities of any image sensor.