| 摘要 |
In tomosynthesis, plural sources 1, 2, 3, 4, form shadowgraph images P1, P2, P3, P4 of an object with features A, B from different perspective directions. In synthesising a selected sectional image the images P1 ... P4 are superposed with appropriate shift to cause corresponding image point locations defining the selected layer to overlap and reinforce one another. The present invention reduces the effect of background image points from other layers by a process of selection or enhancement logic which is applied to each set of overlapping local image values relating to a corresponding image location in the selected section, to allocate to that location a related image value or a substantially zero value depending on whether the image detail is assessed as lying in the selected section or not. In one embodiment, Figure 4, the images P1 ... P4 are each scanned in 5 lines Z1 ... Z5 to form a corresponding video signal 9 which is applied to a parallel storage arrangement with different shift delays O, DELTA t1, DELTA t2, DELTA t3. The overlapping signal values at each location Pn1 ... Pn4 are processed by an analyser 10a which causes a control stage 10b to transfer an image signal to an output register Sn for display only when a signal exceeds a threshold in all or most component images. Alternatively (Fig. 7 not shown) images P1 ... P4 can be multiplied together using a Titus tube, polarised light beam and associated television equipment. <IMAGE> |
