Image scanning device and method

DETAILED DESCRIPTION According to the present invention we provide a method of reconstructing an image captured as image data in a sensor and position data in a navigation means, comprising: defining a pixel grid for all or part of the reconstructed image; using position data to identify correspondence between sensor data and the pixel grid; and determining the intensity of pixels in the pixel grid from image data obtained at a plurality of sensor positions and selected as relevant to each pixel, respectively, for which the intensity is to be determined.
This method is particularly effective when applied to reconstruction of a hand scanned image, which is received as a stream of successive sensor readings with position data for each sensor reading.
Dependence on an individual sensor reading of the scanner is reduced, and data from even irregular scanning can be assembled and processed with relatively little computational effort.
Preferably, the step of identifying correspondence between image data and the pixel grid comprises mapping of image data on to the pixel grid, whereby a region is defined in the pixel grid between the image data of boundary sensor readings after mapping of the boundary sensor readings on to the pixel grid, wherein image data defining said region are selectable as relevant to pixels within said region.
The boundary sensor readings may either be two consecutive readings in the stream, or may alternatively be any two sensor readings in the stream, in which case any sensor readings in the stream between the boundary sensor readings are also selectable as relevant to pixels within said region, the boundary sensor readings and all sensor readings between them in the stream comprising a chunk.
The "chunkwise" approach sacrifices some accuracy for a significant reduction in computation cost.
In preferred embodiments, the step of identifying correspondence between image data and the pixel grid further comprises the step of determining which pixels in the pixel grid fall within the region defined by the selected sets of sensor data