JPH04196775A - Still picture forming device - Google Patents

Abstract

PURPOSE:To eliminate the need for accurate specification of relation of position between pictures for image pickup by increasing number of picture elements sequentially based on a moving vector between the pictures. CONSTITUTION:A moving vector calculation circuit 102 calculates a moving vector from a 1st field data and a 2nd field data deviated by one field from the 1st field data. Then a movement compensation circuit 103 compensates the movement based on the moving vector obtained and the picture element subjected to movement compensation is compensated between a picture element of the 1st field and an interpolated picture element. Then a line interpolation circuit 104 applies line interpolation to the 1st picture data. Then the interpolation value is corrected at an interpolation value correction circuit 106 by using a data subjected to movement compensation with respect to the interpolated value. Since it is not required to make image pickup while specifying the relation of position accurately, a picture with high resolution is simply obtained.

Description

[Detailed description of the invention] Industrial applications The present invention relates to a high-resolution still image forming apparatus.

2. Description of the Related Art In conventional cameras, in order to obtain a still image with high resolution using an image sensor with a small number of pixels, there is a method of temporally shifting the image sensor and synthesizing a plurality of frames.

FIG. 7 shows a configuration diagram for obtaining a still image with high resolution by moving the image sensor in parallel in the horizontal and vertical directions. Light incident on the camera is imaged onto an image sensor 702 by a lens 701.

The imaged light is converted into an electrical signal by the image sensor 702,
It is sent to the memory 703 in accordance with the synchronization signal output by the synchronization signal generator 705.

Further, the image sensor 702 is translated in parallel in the horizontal and vertical directions to fill in the gaps between pixels. Here, the image sensor 702
Since the parallel movement is performed by minute displacement, it is driven by a driving element 704 such as a piezoelectric element. For example, to obtain three times the resolution in the horizontal direction and two times the resolution in the vertical direction, movement is performed three times in the horizontal direction and two times in the vertical direction, a total of six times. Here, the driving time is during the period during which the signal of the image sensor 702 is transferred to the memory.

FIG. 8 shows a diagram in which the image sensor 702 is moved three times in the horizontal direction and twice in the vertical direction, a total of six times, and expanded in the memory. Here, the ◯ mark is image data that was first captured by the image sensor, and the △ and ◇ marks are images that were captured when the image sensor was moved in the horizontal direction. Also, the * mark is image data captured when moving in the vertical direction, and the square mark and ◆ mark are images captured when moving in the horizontal and vertical directions. Then, by outputting the data in the memory as is, a high-resolution image can be obtained.

By moving the image sensor in parallel as described above, a high-resolution image can be obtained using an image sensor with a small number of pixels.

Problems to be Solved by the Invention Although it is possible to obtain high-resolution images by moving the image sensor in parallel, the above method of directly storing it in memory makes it difficult to accurately determine the position of the image sensor in parallel. Must be confirmed.

The present invention provides high resolution (
It is an object of the present invention to provide a still image forming device that can obtain an image with the number of pixels (at least four times the number of pixels of an image sensor).

Means for Solving the Problems 1 In order to achieve the above object, the still image forming device of the present invention pairs two images each out of a plurality of images whose entire screen is slightly shifted, and calculates a motion vector between the images. a motion vector calculation means for calculating the motion vector, a motion compensation means for compensating the motion based on the motion vector, an interpolation means for interpolating between pixels of a reference image, and a motion vector calculation means for calculating the motion based on the motion vector; and interpolation value correction means for correcting interpolation data.

Effect: With the above configuration, when obtaining high-resolution image data by using multiple pieces of image data with a small number of pixels, the number of pixels is sequentially increased based on the motion vector between images.
There is no need to accurately define the positional relationship between images when capturing images.

Example Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of a still image forming apparatus showing an embodiment of the present invention.

In FIG. 1, multiple images with the entire screen slightly shifted are captured by an image sensor, and one field delay device 1
01, processing is performed in pairs of two sheets shifted by one field.

First, the motion vector calculation circuit 102 calculates a motion vector from the second field data that is shifted by one field from the first field data. The calculation method is
The pattern matching method is used as shown in FIG.

The pattern matching method is a method in which a part of the first image is cut out, compared with the second image, and the shift (displacement) when they match is used as the motion vector of the image.

For example, if the pixels used for the first pattern are 8 x 8 pixels and the search area for the second image is 24 x 24, motion vectors can be detected within a range of ±8 in the horizontal direction and ±8 in the vertical direction. . Then, when the absolute value of each difference is taken, the minimum point is found as shown in FIG. Here, if the displacement (motion vector) at the minimum point exists between pixels, for example, if the minimum point exists between b and d, then the displacement at the intersection X of the straight line connecting ab and cd is The displacement at the minimum point may be used.

Then, the motion is compensated by the motion compensation circuit 103 based on the determined motion vector. The motion-compensated pixels are compensated between the first field pixel (O mark) and the interpolation pixel (x mark), as indicated by the Δ mark shown in FIG.

Next, the line interpolation circuit 1
04 performs line interpolation. In Figure 4, if the 0 mark is a pixel in the first field and the It may be determined using the average value of (marked with ○) or an even-order filter.

Next, the interpolated value is corrected by the interpolated value correction circuit 10B using the motion-compensated data for the interpolated interpolated value.

In Fig. 4, if O mark is the first pixel, X mark is the interpolated pixel, and Δ mark is the motion compensated pixel, the data of the interpolated pixel (x mark) is Correction is performed using the data of the pixels (marked with △).

In FIG. 5, if a+ is the upper pixel data, a2 is the lower pixel data, and b is the data interpolated by a++ a2, %C1 is the motion-compensated pixel data, then the corrected interpolated data d+ is d+= (C+-b+)X (la-I+)/ls+b
Determine by +. Here, 111 is the distance between the pixels bl and a2, and 11 is the distance between the pixels bl and C5.

Next, using the interpolation value corrected image, two more images are paired to increase the number of pixels. As previously mentioned,
Motion vector calculation and motion compensation are performed in the same way, but in the subsequent processing, interpolation processing is not line interpolation, but
Perform pixel interpolation.

The pixel interpolation method is horizontal as shown in Figure 6.

Performed vertically and diagonally. Further, the interpolation data is corrected at three points near the motion-compensated pixel.

Then, while the combined data is sequentially paired, the number of pixels is sequentially increased through interpolation and interpolation value correction.

Effects of the Invention As described above, according to the present invention, motion vectors are calculated and motion compensation is performed by pairing two images of a plurality of image data whose entire screen is slightly shifted.

By performing interpolation processing and interpolation data correction, it is not necessary to accurately define the positional relationship between images to capture images, and thus it is possible to easily obtain high-resolution images.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a still image forming apparatus in an embodiment of the present invention, FIG. 2 is a schematic diagram showing a pattern and its search range, and FIG. 3 is a diagram showing the difference between displacement (motion vector) and pattern. A characteristic diagram showing the relationship with the absolute value, FIG. 4 is a schematic diagram showing the relationship between line interpolated pixels and motion compensated pixels, FIG. 5 is a characteristic diagram showing the interpolation value correction method, and FIG.
The figure is a schematic diagram showing the relationship between pixels subjected to pixel interpolation and pixels subjected to motion compensation, and FIG. 7 is a block diagram showing the configuration of a still image forming apparatus that obtains a plurality of images by shifting conventional image sensors. FIG. 8 is a schematic diagram showing the state inside the memory 703 in FIG. 7. 101...1 field delay device, 102...Motion vector calculation circuit, 103...Motion compensation circuit,
104...Line interpolation circuit, 105...Pixel interpolation circuit, 106...Interpolated value correction circuit. Name of agent: Patent attorney Akira Okaji and two others Fig. 2 Fig. 3 f4 tate M 9 v

Claims (2)

[Claims] (1) An imaging means that captures a plurality of images whose entire screen is slightly shifted, and a motion vector that pairs two images out of the data of the plurality of images and calculates a motion vector between the two images. a calculating means; a motion compensating means for compensating for motion based on the motion vector; an interpolating means for interpolating between pixels of a reference image; and correcting interpolated data of the interpolating means using motion-compensated pixel data. A still image forming device comprising interpolation value correction means. (2) The still image forming device according to claim 1, wherein in the interpolation value correction circuit, when the motion-compensated pixel is close to the interpolation pixel, a large weight is applied to the pixel, and when the pixel is far from the interpolation pixel, the correction is performed with a small weight. .