Linear sensor imaging method and apparatus

Abstract

A first linear sensor array produces, at a first time of occurrence, a first image of a segment of an object as the object moves across the sensor at a variable speed. The image and its time of occurrence are accumulated in a buffer. As the object continues to move across the sensor, a successive plurality of the segment images seen by the first sensor are accumulated in the buffer, together with their respective times of occurrence. A second linear sensor array spaced a distance d from the first array produces, at a second time of occurrence, a second image of another segment of the moving object. The second image is compared with the accumulated first images to determine which of the accumulated first images is most similar to the second image. The object's speed S can then be estimated as an appropriately weighted function of d/(t.sub.j -t.sub.m) [i.e. S.varies.d/(t.sub.j -t.sub.m)], where t.sub.j is the time of occurrence of a selected one of the second images and t.sub.m is the time of occurrence of the first image most similar to the selected second image. By continually monitoring the object's speed in this fashion, one may select and transfer into an image buffer those image segments which together form the highest quality image of the object, independently of the object's speed of motion past the sensor and without the need for expensive, bulky stepper motors, mechanical motion encoders, or the like.