JPH0458682A - Image pickup device - Google Patents

Abstract

PURPOSE:To reduce the deterioration in the resolution in an image pickup element whose adjacent picture elements in the vertical direction are deviated in the horizontal direction with deviation in 1/2 picture element with each other and to attain natural interpolation of contour by providing an interpolation means, a selection means and a contour detection means. CONSTITUTION:A video image formed on an image pickup element 2 through a lens 1 is stored in a frame memory 4 via an A/D converter 3. A data D corresponding to each point represented in a coordinate is read from the frame memory 4 and calculated respectively by a vertical correlation matrix 6, a horizontal correlation matrix 7 and a vertical horizontal correlation matrix 8. A contour detection circuit 5 applies a prescribed calculation and when the result is 0 or over, a selector switch 9 selects an output of the vertical correlation matrix 6, and when the output is 0 or below, it is discriminated that the contour in the vertical direction is strong and the output of the horizontal correlation matrix 7 is selected. When the output is 0, an output of the vertical horizontal correlation matrix 8 is selected by the selector switch 9. The output of the selector switch 9 is written again in the frame memory 4. The data is unchanged since the data passes through an interpolation matrix at a point indicated by a picture element 12 actually in existence.

Description

Detailed Description of the Invention - (Industrial Application Field) The present invention relates to an imaging device having an imaging device in which vertically adjacent pixels are arranged horizontally shifted by 1/2 pixel from each other. It concerns interpolation.

[Conventional technology]

Conventionally, in an imaging device using an imaging device in which vertically adjacent pixels 12 are arranged horizontally shifted by 1/2 pixel from each other, as shown in FIG. When interpolating a virtual pixel (hereinafter referred to as a virtual pixel) 13 at a virtual position using surrounding effective pixels 12, the interpolation method is shown in FIGS. 3(b), (e), and (d). Among the available methods, the structure is fixed to one method, that is, interpolation using a matrix.

[Problem to be solved by the invention]

In the conventional configuration described above, since interpolation is performed using a single matrix, there is a problem as shown in -1 below.

a. When performing interpolation using a vertical (horizontal) correlation matrix, the resolution in the vertical (horizontal) direction deteriorates.

b. In the case of the interpolation method using the vertical phase open 7-trix, an appropriate output cannot be obtained for the contour portion in the horizontal direction, resulting in an unnatural contour.

Similarly, in the case of an interpolation method using a horizontal correlation matrix,
A similar phenomenon appears in the vertical contour.

The present invention has been made under these circumstances, and it is an object of the present invention to provide an imaging device of this type that is capable of natural interpolation of contours with little deterioration in resolution.

[Means to solve the problem]

In the present invention, in order to achieve the above object, an imaging device is configured as shown in (1) below.

(1) Vertically adjacent pixels are horizontally 1/1/2
An imaging device having an imaging element arranged two pixels apart and interpolating a virtual pixel surrounded by the pixel; An imaging device comprising a selection means for making a selection, and an outline detection means for detecting an outline of an image near a virtual pixel to be interpolated and instructing the selection means to select an interpolation mode that matches the outline.

[Effect]

With the configuration (1) above, the interpolation method is selected that matches the outline of the image near the virtual pixel to be interpolated.

(Example〕 The present invention will be explained in detail below with reference to Examples.

FIG. 1 is a configuration diagram of an "imaging device" which is an embodiment of the present invention. In the figure, 1 is an optical lens, 2 is an image sensor in which vertically adjacent pixels are shifted horizontally by 1/2 pixel, 3 is an A-D (analog-digital) converter, and 4 is a frame. Memory, 5 is a contour detection circuit, 6
1 is an interpolation matrix A, 7 is an interpolation matrix B, 8 is an interpolation matrix C19 is a selection switch, and 10 is a DA (digital-to-analog) converter 511 is an imaging signal processing block.

FIG. 3 shows the pixel arrangement of the image sensor 2 and its readout path in the configuration diagram of FIG. 1, and the specific method of interpolation matrices A, B, and C shown in FIG. It is.

In FIG. 3(a), 12 is an actually existing pixel, ]
3 is a virtual pixel to be interpolated, and 14 is a signal readout path. 15 is a two-dimensional coordinate, and when it is imported into the frame memory 4, it has a one-to-one relationship with the address in the memory 4.
The following measures shall be taken. FIG. 3(b) shows a vertical correlation interpolation method using upper and lower signals, which has the characteristic of lowering the resolution in the vertical direction. Figure 3(e) is on the left;
This is a horizontal correlation interpolation method that uses the right signal, which has the characteristic of reducing resolution in the horizontal direction. Figure 3(d)
is a vertical-horizontal correlation interpolation method that uses upper, lower, left, and right signals, and has the characteristic of reducing resolution in both vertical and horizontal directions. This figure shows the specific circuit configurations of the contour detection circuit 5 and the interpolation matrices A, B, and C shown in 6.7.8.

In the configurations shown in FIGS. 1, 2, and 3, the optical lens 1
The image formed on the image sensor 2 is converted into an electrical signal, further converted into digital data by the AD converter 3, and stored in the frame memory 4. At this time, the coordinate data corresponding to the virtual pixel 13 to be interpolated shown in FIG. 3 is set to 0. The accumulated data is read out sequentially to perform interpolation. Now, consider interpolation of a point indicated by (x, y) coordinates corresponding to the virtual pixel 13 to be interpolated. In Fig. 2, from frame memory 4, the positions # (x, 3'-1), (x, y+1), (x, y)
, (x-1, y), (x+1°y), D (X, y-1), D (
x, y+1), D (x.

y), D (x-1, y), D (x+1. y)
are read out and calculated using a vertical correlation matrix 6, a horizontal correlation matrix 7, and a vertical/horizontal correlation matrix 8, respectively. Also, in the contour detection circuit 5, D(x, y-1) −
D(x, y+]) I-l D(x-1, y) -D(
x+1. y) 1 is performed, and if the result is 0 or more, it is determined that the contour in the horizontal direction is stronger than in the vertical direction in the vicinity of the virtual pixel to be interpolated, and the selection switch 9 is
The output of the vertical correlation matrix 6 is selected.

Further, when the output of the contour detection circuit 5 is 0 or more, it is determined that the contour in the vertical direction is stronger than in the horizontal direction, and the output of the horizontal correlation matrix 7 is selected by the 1 selection switch 9.

Also, if the output of the contour detection circuit 5 is equal to 0, the horizontal
It is determined whether the contour edges in the vertical direction are at the same level, and the selection switch 9 selects the output of the vertical/horizontal 7 trix 8. In each case, the output of the selection switch 9 is written into the frame memory 4 again.

Next, consider a point corresponding to the actually existing pixel 12. As shown in FIG. 3(a), the pixels above, below, left, and right of the actually existing pixel 12 are all interpolated virtual pixels 13, and in the frame memory 4, the data is It is 0. Follow-up Figure 2 Contour detection circuit 5
The output of the vertical/horizontal correlation matrix 8 is always 0, and as a result, the output of the selected vertical/horizontal correlation matrix 8 always holds 1-element data, so the point indicated by the actually existing pixel 12 passes through the interpolation matrix. This will not change the data.

The data obtained by performing the interpolation as described above is stored again in the full scale memory 4, converted to analog by the D-A converter 10, processed by the imaging signal processing block 11, and then processed into the image signal processing block 11. Output as a signal.

In addition, in the above embodiment, the readout of the image sensor is performed as shown in FIG.
As shown in a), the present invention is not limited to kneading in a zigzag manner, but can also be carried out, for example, with reading in the horizontal direction.

Further, in the embodiment, the outputs of the interpolation circuits 6 to B are switched, but it goes without saying that the interpolators may be switched by switching the software of the microcomputer.

(Effects of the Invention) As explained above, according to the present invention, the contour of the image in the vicinity of the virtual pixel to be corrected is detected, and the interpolation mode that matches the contour is selected to perform interpolation. Images with natural contours can be obtained with little deterioration in resolution.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention, FIG. 2 is a circuit diagram of essential parts of the embodiment, and FIG. 3 is a diagram explaining the pixel arrangement, readout path, and interpolation method of the image sensor. 2-m-imaging element 5--contour detection circuit 6.7.8-m-interpolation matrix 9--selection switch 12-1 pixel 13-=virtual pixel

Claims (1)

[Claims] (1) Vertically adjacent pixels are horizontally 1/1/2
An imaging device having an imaging element arranged two pixels apart and performing interpolation of a virtual pixel surrounded by the pixel, an interpolation means having a plurality of types of interpolation modes, and selecting the plurality of types of interpolation modes. and contour detection means for detecting the contour of an image in the vicinity of a virtual pixel to be interpolated and instructing the selection means to select an interpolation mode that matches the contour.