Corner detection device and corner detection method

Claim 2

A corner detection device for detesting from a retrieval scope of an input signal each corner position of an area of a square configuration where a signal level is set to a level higher than a level of a reference signal level, the device comprising: memory means for memorizing said input signal; and corner detection means for detecting a point where the signal level of said input signal is at first set to a level higher than the level of said reference signal by reading from said retrieval scope upper and lower limits in a horizontal direction sequentially the input signal which is memorized in said memory means, and at the same time reading from the retrieval scope left and right ends in a vertical direction sequentially, and detecting a point where the signal level of said input signal is at first set to a level higher than the level of said reference signal by reading the input signal memorized in said memory means in a diagonal direction sequentially at a predetermined angle from each angle, and detecting four points whose positions differ from the detected points.

Claim 3

The corner detection device according to claim 1, wherein said corner detection means examines distances between said detected points, and judges that the positions differ when the distances exceed a predetermined reference value.

Claim 4

The corner detection device according to claim 1, wherein said corner detection means judges whether or not the detected points are correct by selecting two points which form sides of the area out of said four points, and by examining whether the signal level only on one side is set to a level higher than the level of said reference signal out of both sides of a straight line connecting the two points.

Claim 5

The corner detection device according to claim 1, wherein said corner detection means detects a point which is set to a level higher than the reference signal level by reading in a horizontal direction and in a vertical direction followed by determining the second retrieval scope on the basis of the detected points and performs the reading in said diagonal direction in the second retrieval scope.

Claim 6

The corner detection device according to claim 1, wherein said corner detection means detects the diagonal line of said area from the detected points in a case in which the four points whose positions differ cannot be detected by reading in said horizontal direction and in said vertical direction as well as reading in said diagonal direction, the means detecting points which are not detected by reading in the diagonal direction at an angle of the diagonal line.

Claim 7

The corner detection device according to claim 5, wherein said corner detection means detects a third point which is not detected by reading in a diagonal direction at an angle of a straight line connecting two points which are detected in a case in which said diagonal line cannot be detected, the means detecting a fourth point which cannot be detected by detecting said diagonal line from the detected three points and reading in the diagonal direction at an angle of the diagonal line.

Claim 8

A corner detection method for detecting from a retrieval scope of an input signal each corner position of an area of a square configuration where a signal level is set to a level higher than a level of a reference signal level, the method comprising: a memory step of memorizing said input signal in a predetermined memory means; a first detection step of reading said input signal memorized in said memory means sequentially in a horizontal direction from an upper limit to a lower limit of said retrieval scope and detecting a point where the signal level of said input signal is set to a level higher than said reference signal level at first by reading sequentially in a vertical direction from the right end and the left end of said retrieval scope; a second detection step of detecting a point where the signal level of said input signal is set to a level higher than the level of said reference signal level at first by reading said input signal which is memorized in said memory means sequentially in the diagonal direction at a predetermined angle from each angle of said retrieval scope; and a third detection step of detecting said corner position by detecting the four points whose positions differ from points detected in said first and second detection steps.

Claim 9

A corner detection method for detecting from a retrieval scope of an input signal each corner position of an area of a square configuration where a signal level is set to a level higher than a level of a reference signal level, the method comprising: a memory step of memorizing said input signal in a predetermined memory means; a first detection step of reading said input signal memorized in said memory means sequentially in a horizontal direction from an upper limit to a lower limit of said retrieval scope and detecting a point where the signal level of said input signal is set to a level higher than said reference signal level at first by reading sequentially in a vertical direction from the right end and the left end of said retrieval scope; a second detection step of detecting a point where the signal level of said input signal is set to a level higher than the level of said reference signal level at first by reading said input signal which is memorized in said memory means sequentially in the diagonal direction at a predetermined angle from each angle of said retrieval scope; and a third detection step of detecting said corner position by detecting the four points whose positions differ from points detected in said first and second detection steps.

Claim 10

The corner detection method according to claim 7, wherein said third detection step examines a distance between the points which are detected in said first and second detection steps to judge that the positions differ when the distance exceeds a predetermined reference value.

Claim 11

The corner detection method according to claim 7, wherein said third detection step judges whether or not the detected points are correct by selecting two points which form sides of said area out of said detected four points, and by examining whether the signal level only on one side is set to a level higher than the level of said reference signal out of both sides of a straight line connecting the two points.

Claim 12

The corner detection method according to claim 7, wherein said first detection step detects a point which is set to a level higher than the level of the reference signal by reading in the horizontal direction and in the vertical direction followed by determining the second retrieval scope on the basis of the detected point while said second detection step performs reading in said diagonal direction in said second retrieval scope.

Claim 13

The corner detection method according to claim 7, wherein said third detection step detects the diagonal line of said area from the detected points in a case in which the four points whose positions differ cannot be detected by reading in said horizontal direction and in said vertical direction in said first detection step as well as reading in said diagonal direction in said second detection step, the step detecting the point which cannot be detected by reading in the diagonal direction at an angle of the diagonal line.

Claim 14

The corner detection method according to claim 11, wherein said third detection step detects the third point which cannot be detected by reading in the diagonal direction at an angle of a straight line connecting the two points which are detected when said diagonal line cannot be detected, the step detecting the fourth points which cannot be detected in reading in the diagonal direction at an angle of the diagonal line by detecting said diagonal line from three points which are detected.

Claim 15

A video image processing device for synthesizing a second video image with a first video image, comprising: coordinate detection means for detecting the coordinates of four corners of a rectangular area which is'set to said first video image by retrieving in plural directions, at least, in the horizontal direction, in the vertical direction and in the oblique direction, the all of pixels of said first video image; three-dimensional conversion means for computing a three-dimensional conversion matrix used in a three-dimensional conversion processing performed on said second video image based on the coordinates of four corners which are detected by said coordinate detection means, and performing the three-dimensional conversion processing on said second video image using the computed three-dimensional conversion matrix so that the size of the rectangular area which is set to said first video image coincides with the size of the second video image which is three-dimensionally converted; and synthesizing means for synthesizing said first video image with the second video image which is three-dimensionally converted by said three-dimensional conversion means.

Claim 16

A video image processing device for synthesizing a second video image with a first video image, comprising: coordinate detection means for detecting the coordinates of four corners of a rectangular area which is'set to said first video image by retrieving in plural directions, at least, in the horizontal direction, in the vertical direction and in the oblique direction, the all of pixels of said first video image; three-dimensional conversion means for computing a three-dimensional conversion matrix used in a three-dimensional conversion processing performed on said second video image based on the coordinates of four corners which are detected by said coordinate detection means, and performing the three-dimensional conversion processing on said second video image using the computed three-dimensional conversion matrix so that the size of the rectangular area which is set to said first video image coincides with the size of the second video image which is three-dimensionally converted; and synthesizing means for synthesizing said first video image with the second video image which is three-dimensionally converted by said three-dimensional conversion means.

Claim 17

The video image processing device according to claim 13, wherein the rectangular area which is set to said first video image corresponds to an area which has a specified color of said first video image.

Claim 18

The video image processing device according to claim 14, further including, key signal generation means for generating a first key signal corresponding to the rectangular area which is set to said first video image, said video image processing device wherein, said synthesizing means synthesizes said first video image with said second video image which is three-dimensionally converted based on said first key signal.

Claim 19

The video image processing device according to claim 15, wherein said three-dimensional conversion means further performs said three-dimensional conversion processing on the second key signal for keying said second video image.

Claim 20

The video image processing device according to claim 16, wherein said three-dimensional conversion means has memory means for memorizing said second video image, and read address generation means for supplying a read address corresponding to said conversion matrix to said memory means.

Claim 21

The video image processing device according to claim 17, wherein said read address generation means computes said read address based on the inverse matrix of said three-dimensional conversion matrix.

Claim 22

The video image processing device according to claim 18, wherein said three-dimensional conversion matrix is shown by a matrix in three lines by three columns composed of eight conversion parameters as follows: ##EQU18##20.

Claim 23

The video image processing device according to claim 19, wherein said three-dimensional conversion means obtains eight parameter values of said three-dimensional conversion matrix utilizing that the following equation is satisfied: ##EQU19## where (X 1 , Y 1 ) (X 2 , Y 2 ) (X 3 , Y 3 ) (X 4 , Y 4 ) show the coordinates of four corners in the area set to the first video image, (X' 1 , Y' 1 ) (X' 2 , Y' 2 ) (X' 3 , Y' 3 ) (X' 4 , Y' 4 ) show the coordinates of four corners of the second video image, and K 1 , K 2 , K 3 , K 4 show parameters in simultaneous coordinate system.

Claim 24

A video image processing method for synthesizing a second video image with a first video image, comprising the steps of: detecting the coordinates of four corners of a rectangular area which is set to said first video image by retrieving in plural directions, at least, in the horizontal direction, in the vertical direction and in the oblique direction, the all of pixels of said first video image; computing a three-dimensional conversion matrix used in a three-dimensional processing performed on said second video image based on the detected coordinates of the four corners, performing the three-dimensional conversion processing on said second video image using the computed three-dimensional conversion matrix so that the size of the rectangular area set to said first video image coincides with the size of the second video image which is three-dimensionally converted; and synthesizing said first video image with said second video image which is three-dimensionally converted.

Claim 25

A video image processing method for synthesizing a second video image with a first video image, comprising the steps of: detecting the coordinates of four corners of a rectangular area which is set to said first video image by retrieving in plural directions, at least, in the horizontal direction, in the vertical direction and in the oblique direction, the all of pixels of said first video image; computing a three-dimensional conversion matrix used in a three-dimensional processing performed on said second video image based on the detected coordinates of the four corners, performing the three-dimensional conversion processing on said second video image using the computed three-dimensional conversion matrix so that the size of the rectangular area set to said first video image coincides with the size of the second video image which is three-dimensionally converted; and synthesizing said first video image with said second video image which is three-dimensionally converted.