Nonaliasing real-time spatial transform image processing system
First Claim
1. A method of mapping in a 3D coordinate system of 2D input images of views of objects stored in a data base to output images in a moveable rectangular viewing plane having a recatngular column and row pixel matrix, said system having a main axis extending normal to said viewing plane and having a viewing point thereon in spaced relation to said viewing plane, said method comprising the steps of:
- representing each of said 2D images as a rectangular matrix of pixel intensity values arranged in columns and rows of pixels and having corner coordinates;
establishing a rectangular pixel matrix of each of said stored 2D images;
computing four corner coordinates for an output image in said viewing plane corresponding to said corner coordinates of any selected one of said 2D images;
making tow orthogonal passes wherein said first pass is over said input image to form an intermediate image having columns and rows of pixels and corner coordinates determined by said computed four coordinate corners of said output image, to form said output image with columns and rows of pixels and corner coordinates determined by said computed four coordinate corners of said output image;
making line transformations in said first pass wherein said input image columns are converted to said intermediate image columns with said intermediate image columns having lengths derived from said input image columns establishing respective ratios defining the lengths of said input image columns relative to said intermediate image columns;
making line transformations in said second pass wherein said rows of said intermediate image are converted to said output image rows on said viewing screen with said output image rows having lengths derived form said intermediate image rows establishing respective ratios defining the lengths of said intermediate image rows relative to said output image rows;
calculating and forming for each of said line transformations from said ratios, the number of input pixels required respectively from said columns of said input image and said rows of said intermediate image to create output pixels respectively for said intermediate image columns and for said final image rows;
mapping aid input pixels sequentially and fractionally to said output pixels for each of said line transformations; and
summing and averaging said intensity values of said input pixels corresponding to each said output pixel and respective accruing composite values so obtained to said output pixels.
1 Assignment
0 Petitions
Accused Products
Abstract
The disclosure relates to a computer controlled imaging system involving a digital image processing and display system which has the ability to compose and construct a display scene from a library of images with sufficient processing speed to permit real-time or near real time analysis of the images by a human operator or a hardware/software equivalent thereof. An unique method of mapping a digital input image to a viewing plane in both linear and perspective modes is disclosed.
171 Citations
9 Claims
-
1. A method of mapping in a 3D coordinate system of 2D input images of views of objects stored in a data base to output images in a moveable rectangular viewing plane having a recatngular column and row pixel matrix, said system having a main axis extending normal to said viewing plane and having a viewing point thereon in spaced relation to said viewing plane, said method comprising the steps of:
-
representing each of said 2D images as a rectangular matrix of pixel intensity values arranged in columns and rows of pixels and having corner coordinates; establishing a rectangular pixel matrix of each of said stored 2D images; computing four corner coordinates for an output image in said viewing plane corresponding to said corner coordinates of any selected one of said 2D images; making tow orthogonal passes wherein said first pass is over said input image to form an intermediate image having columns and rows of pixels and corner coordinates determined by said computed four coordinate corners of said output image, to form said output image with columns and rows of pixels and corner coordinates determined by said computed four coordinate corners of said output image; making line transformations in said first pass wherein said input image columns are converted to said intermediate image columns with said intermediate image columns having lengths derived from said input image columns establishing respective ratios defining the lengths of said input image columns relative to said intermediate image columns; making line transformations in said second pass wherein said rows of said intermediate image are converted to said output image rows on said viewing screen with said output image rows having lengths derived form said intermediate image rows establishing respective ratios defining the lengths of said intermediate image rows relative to said output image rows; calculating and forming for each of said line transformations from said ratios, the number of input pixels required respectively from said columns of said input image and said rows of said intermediate image to create output pixels respectively for said intermediate image columns and for said final image rows; mapping aid input pixels sequentially and fractionally to said output pixels for each of said line transformations; and summing and averaging said intensity values of said input pixels corresponding to each said output pixel and respective accruing composite values so obtained to said output pixels.
-
-
2. A method of linearly mapping a rectangular digital input image having columns and rows and corner coordinates, and represented as a pixel matrix of intensity values, to any predetermined quandrilateral having columns and rows and corner coordinates, comprising the steps of:
-
performing sequentially two orthogonal scanning passes of said input image wherein said first pass forms from said input image an intermediate image having columns and rows of pixels and corner coordinates and said second pass forms an output image having columns and rows of pixels and corner coordinates from said intermediate image; performing line transformations during said first pass whereby said columns of said intermediate image correspond to said columns of said input image and have lengths and positions determined from said coordinates of said quadrilateral; performing line transformations during said second pass whereby said rows of said output image correspond to said rows of said intermediate image and have lengths and positions determined from said coordinates of said quadrilateral; calculating for each of said line transformations (1) the vertical ratios of the respective lengths of said columns of said input image relative to said columns of said intermediate image and (2) the horizontal ratios of the respective lengths of said rows of said intermediate image relative to said rows of said output image; calculating a pixel ratio for each of aid line transformations from said vertical and horizontal ratios to determine respectively, for each of said line transformations, the number of said input image pixels required to create said intermediate image pixels and the number of said intermediate image pixels required to create said output image pixels and calculating, for each of said created intermediate and output pixels, an average intensity value from the intensity values of said input and intermediate pixels respectively; and mapping said input pixels sequentially and fractionally to said intermediate pixels and mapping and intermediate pixels sequentially and fractionally to said output pixels for each of said line transformations.
-
-
3. A method of mapping a digital input image to a viewing screen which is moveable in a 3D coordinate system and wherein said image comprises columns and rows of pixels representing a matrix of pixel intensity values in said system, said system having an observation point on a main axis thereof and a viewing screen having a format with columns and rows extending normal to said axis and spaced from said observation point, said system defining a vertical reference plane parallel to said main axis and extending normal to said viewing screen, comprising the steps of:
-
projecting sequentially and orthogonally said columns of said input image to said reference plane to form first pass object lines in said reference plane; processing each line of said first pass object lines by projecting divergent lines form said observation point to said each line of said first pass object lines to define an intermediate vertically extending image column line in relation to said viewing screen having equal line segments of a length at least equal to one pixel length, said divergent lines intersecting said each line of said first pass object lines and resulting in corresponding first pass line segments thereof between said divergent lines having respective lengths corresponding to the number of said pixels separating adjacent one of said divergent lines; calculating pixel ratios of said first pass line segments relative to corresponding ones of said line segments of said intermediate image column line; mapping pixels respectively from said first pass line segments to corresponding ones of said line segments of said intermediate image column line; projecting divergent planes from said observation point through said rows of said viewing screen having separation of at least one pixel length in said screen each of said divergent planes intersecting said screen to form a final image row line and intersecting said input image to form a second pass object line; processing each said second pass object line in each of said divergent planes by projecting second pass divergent lines form said observation point through said columns of said screen having a separation of at least one pixel in said screen, said second pass divergent lines intersecting each said second pass object line to form second pass line segments thereof and intersecting said screen to form horizontally extending correctly positioned final image rows, said second pass line segments having respective lengths corresponding to the number of said pixels separating adjacent ones of said second pass divergent lines; calculating pixel ratios of said second pass line segments relative to corresponding line segments of said final image row; and mapping pixels respectively from said second pass line segments to corresponding line segments of each said final image row.
-
-
4. A method for the linear position mapping of input pixels from an input pixel line in an input image to output pixels in an output pixel line in an output image, comprising the steps of:
-
(a) Determining the corresponding starting and stopping points for said lines and the size factor defined as the ratio of the length of said input line to the length of said output line, said ratio also being equal to the fractional number of consecutively arranged ones of said input pixels required to form the first one of said output pixels; (b) beginning at the start of said input pixel line and sequentially feeding consecutively available pixels therefrom until a current number at least equal to said ratio is reached; (c) feeding a pixel to said output pixel line and finding the current fractional pixel difference between said current number and said ratio; (d) returning to said input pixel line and, starting with the next available pixel, continuing to sequentially feed consecutive available pixels therefrom, the number of which are added to said current difference until a new current number at least equal to said ratio is reached; and (e) returning to step (c) to continue the process until there are no more available pixels in said input pixel line. - View Dependent Claims (6)
-
-
5. A method for linear position mapping of input pixels from an input pixel line in an input image to output pixels in an output pixel line in an output image, comprising the steps of:
-
(a) determining the corresponding starting and stopping points for said lines and the size factor defined as the ratio of the length of said input line to the length of said output line, said ratio also being equal to the fractional number of consecutively arranged ones of said input pixels required to form one of said output pixels; (b) beginning at the start of said input pixel line and sequentially feeding consecutively available pixels therefrom; (c) calculating a sum by incrementing an initial value by one upon the said feeding of each said consecutive available pixel until a current sum at least equal to said ratio is reached; (d) feeding a pixel to said output pixel line and finding the current difference between said current sum and said ratio when the current sum is at least equal to said ratio; (e) returning to said input pixel line and, starting with the next available pixel, continuing to sequentially feed consecutive available pixels therefrom and incrementing said current difference until a new current sum at least equal to said ratio is reached ; and (f) returning to step (d) to continue the process until there are no more available pixels in said input pixel line.
-
-
7. A method for nonlinear position mapping of input pixels from an input pixel line in a input image to output pixels in an output pixel line in an output image, comprising the steps of:
-
(a) determining (1) the corresponding starting and stopping points for said lines, (2) a size and factor function, and (3) the current size factor ratio of the fractional number of consecutively arranged ones of said input pixels required to form;
one of said output pixels lines, said ratio being at the origin of said lines;(b) beginning at the start of said input pixel line and sequentially feeding consecutive available pixels therefrom until a current sum at least equal to said current ratio is reached; (c) feeding a pixel to said output pixel line and finding a new current size factor ratio from said size factor function at that point and finding a current difference between said current sum and said current ratio; (d) returning to said input pixel line and, starting with the next available pixel, continuing to sequentially feed consecutive available pixels therefrom which are added to said current difference until anew current sum at least equal to said current size factor ratio is reached; and (e) returning to step (c) to continue the process until there are not more available pixels in the input line. - View Dependent Claims (8)
-
-
9. A method of mapping a digital input image to a viewing screen which is moveable in a 3D coordinate system and wherein said image comprises columns and rows of pixels representing a matrix of pixel intensity values in said system, said system having an observation point on a main axis thereof and viewing screen having a format with columns and rows of pixels extending normal to said axis and spaced from said observation point, said system defining a vertical and horizontal reference plane parallel to said main axis and extending normal to said viewing screen, comprising the steps of:
-
projecting sequentially and orthogonally said columns of said input image to said vertical reference plane to form first pass object lines in said vertical reference plane; processing each line of said first pass object lines by projecting divergent lines form said observation point to said each line of said first pass object lines to define an intermediate vertically extending image column line in relation to said viewing screen having equal line segments of a length at least equal to one pixel length, said divergent lines intersecting said each line of said first pass object lines and resulting in corresponding first pass line segments thereof between said divergent lines having respective lengths corresponding to the number of said pixels separating adjacent ones of said divergent lines; calculating pixel ratios of said first pass line segments relative to corresponding ones of said line segments of said intermediate image column line; mapping pixels respectively from said first pass line segments to corresponding ones of said line segments of said intermediate image column line; projecting divergent planes form said observation point through said rows of said viewing screen having a separation of at least one pixel length in said screen, each of said divergent planes intersecting said screen to form a final image row line and intersecting said input image to form a second pass object line; projecting sequentially and orthogonally said second pass object lines to said horizontal reference plane to form constructional second pass object lines in said horizontal reference plane; processing each said constructional second pass object line by projecting second pass divergent lines from said observation point through said columns of said screen having a separation of at least one pixel in said screen, said second pass divergent lines intersecting each said constructional second pass object line to form second pass line segments thereof and intersecting said screen to form horizontally extending correctly positioned final image rows, said second pass line segments having respective lengths corresponding to the number of said pixels separating adjacent ones of said second pass divergent lines; calculating pixel ratios of said second pass line segments relative to corresponding line segments of said final image row; and mapping pixels respectively form said second pass line segments to corresponding line segments of each said final image row.
-
Specification