JP2516580B2 - Image motion detector - Google Patents

Description

The present invention relates to an image motion detection device used for image stabilization of an image pickup device.

[Conventional technology]

When an image of a subject is picked up by using the image pickup device, if the image pickup device is directly held by a hand or mounted on a moving body such as a vehicle, shaking occurs during image pickup on the screen. Therefore, a hand shake correction device that detects the shake of the screen from the video signal and corrects the position of the image based on the shake information has been put into practical use. Such a camera shake correction device requires an image motion detection device that detects the motion of the image on the screen.

The operation principle of the conventional image motion detection device will be described.

FIG. 3 is a block diagram of a conventional image motion detection device. In the figure, the representative point storage circuit 1 divides the current field of the screen currently imaged by the imaging device into a plurality of areas,
A video signal corresponding to a specific representative point of each area is stored as a modern front point signal. This circuit also supplies to the correlation calculation circuit 2 the representative point signal of the previous field scanned one field before the current field. The correlation calculation circuit 2 performs a correlation calculation between the previous representative point signal and the video signal of the current field, and compares the difference between the previous representative point signal and the video signal of the current field, and the output thereof is the motion vector detection circuit 3
Given to. The motion vector detection circuit 3 analyzes how the video signal of the corresponding representative point of the previous field and the current field changes the address on the representative point storage circuit 1 from the calculation result of the correlation calculation circuit 2 to determine the image motion. It detects a vector.

The conventional image motion detection device configured as described above is
When an interlaced video signal is supplied, the video signal of the image obtained by dividing the video signal of one field before into a plurality of areas, setting a representative point in each area, and sampling at this representative point The above correlation calculation is performed between the point signal and the video signal in each area of the current field corresponding to the point signal. After that, the motion of the image is detected by obtaining an image motion vector indicating in what direction and in what direction the image of the current field has moved with respect to the image of the previous field.

[Problems to be Solved by the Invention]

However, the above-mentioned configuration causes the following problems. This will be described with reference to FIG.

FIG. 4 is a state diagram showing an A field for scanning odd scanning lines and a B field for scanning even scanning lines of a video signal in one frame. In the figure, the scanning lines m1 to m9 indicated by solid lines are the scanning lines of the A field,
Scan lines n1 to n9 indicated by broken lines indicate scan lines of the B field, and the A field and the B field form a frame. Here, when the same image that is stationary on the image pickup screen is scanned in the A field and the B field, it is understood that a vertical pixel shift occurs by D as shown in FIG. As a result, when the correlation calculation is performed between the previous representative point signal of the A or B field and the respective modern surface point signals in each area of the video signal of the B or A field corresponding thereto, as shown in FIG. , B, correlation calculation is performed with the vertical pixel shift D due to interlacing occurring between both fields. Therefore, there is a problem that the detected motion vector necessarily includes an error in the vertical direction on the image pickup screen.

The present invention solves such a conventional problem, and an object of the present invention is to improve the accuracy of image motion detection in a direction perpendicular to a scanning line on an imaging screen.

[Means for solving the problem]

The invention of claim 1 of the present application holds the video signal at the representative point of each region of the current field as a modern front point signal in the video signal divided into a plurality of regions, and the previous representative at the representative point of each region of the previous field. A representative point storage circuit that outputs a point signal, a correlation calculation circuit that performs correlation calculation processing between the previous representative point signal from the representative point storage circuit and the corresponding video signal of the current field, and the calculation processing result of the correlation calculation circuit An image motion detection device used in an interlaced video signal processing device comprising a motion vector detection circuit for detecting a motion vector of an image from the motion vector detection circuit, the vertical component of the motion vector detected by the motion vector detection circuit. In addition, a motion vector correction circuit for performing vector correction by adding or subtracting the vertical scanning position shift amount between the current field and the previous field is provided. Is shall.

Further, the invention of claim 2 of the present application is characterized in that the scanning position shift amount of the motion vector correction circuit is set to 1/2 of the line interval in one field.

[Action]

According to the inventions of claims 1 and 2 having such characteristics, the image in the imaging screen is obtained by performing the correlation calculation processing between the previous representative point signal in the video signal of the previous field and the video signal of the current field. The motion vector of is detected. In such an image motion detection device, a fixed vector for correcting the shift between the scanning positions of the current field and the previous field is created by the motion vector correction circuit, and this fixed vector is combined with the above-described motion vector. An image motion vector with no vertical error is obtained.

〔Example〕

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of an image motion detecting device according to an embodiment of the present invention. It should be noted that components having the same functions as those in FIG. 3 showing the conventional example are designated by the same reference numerals, and detailed description thereof will be omitted. In the figure, the representative point storage circuit 1 divides the current field into a plurality of areas, stores and holds a video signal corresponding to the representative point of each area as a modern front point signal, and scans one field before the current field. This is a representative point storage circuit for outputting the previous representative point signal of the preceding field. The correlation calculation circuit 2 is a correlation calculation circuit for performing a correlation calculation between the preceding representative point signal and the video signal in each area of the current field, and specifically, it is the same as the conventional operation. The motion vector detection circuit 3 is a motion vector detection circuit that detects an image motion vector from the calculation result of the correlation calculation circuit 2, and these are the same as the above-mentioned conventional example. Now, in the present embodiment, a motion vector correction circuit 4 for correcting the image motion vector obtained by the motion vector detection circuit 3 is provided.

FIG. 2 is a block diagram showing a specific configuration of the motion vector correction circuit 4 shown in FIG. In the figure, the adder 4a is a circuit that adds a certain signal to the signal input to it, and the subtractor 4b is a circuit that subtracts a certain signal from the signal that is also input. In the switch circuit 4c, if the previous field is an A field that scans odd scan lines,
It is a switch circuit that switches the outputs of the adder 4a and the subtractor 4b depending on whether the B field scans an even scan line.
The motion vector correction circuit 4 corrects an error caused when the image motion vector obtained by the motion vector detection circuit 3 is detected by ignoring a vertical image shift between fields due to interlaced scanning. It is provided. The motion vector detection circuit 3 outputs the vertical scanning position shift amount of the A and B fields depending on whether the current field scanned by the imaging device is an even field (B field) or an odd field (A field). It is added or subtracted to the vertical component of the motion vector of the image and combined. Here, the correction amount to be subtracted or added is the size of the scanning position shift amount D shown in FIG. 4 (in this embodiment, half the size of the scanning line of one field).

The operation of the image motion detecting device of the present embodiment configured as described above will be described below.

The representative point storage circuit 1 divides the current field currently displayed on the imaging screen into a plurality of areas, stores a video signal corresponding to a specific representative point in each area as a modern front point signal, and stores it for at least one field period. Hold. Then, the previous representative point signal corresponding to the specific representative point stored in the previous field one field before the current field is output to the correlation calculation circuit 2. The correlation calculation circuit 2 directly receives each previous representative point signal of the previous field output from the representative point storage circuit 1 and the corresponding video signal in each area of the current field, and performs a correlation calculation process between the two signals. I do. In the correlation calculation process, for example, the previous representative point signal stored in the representative point storage circuit 1 is processed in the memory and subtracted from the video signal of the current field to analyze the cross-correlation between the two signals. However, here, the apparent shift of the pixels detected by shifting in the vertical direction between the fields due to the interlaced scanning is not taken into consideration.

Based on the calculation processing result obtained from the correlation calculation circuit 2, the motion vector detection circuit 3 detects an image motion vector indicating in which direction and in what direction the pixels of the current field have moved with respect to the image of the previous field. To do. here,
Among many pieces of image motion vector information obtained in a plurality of regions in each field for which correlation calculation processing is performed, these pieces of image motion vector information are classified into vector information indicating the same direction and the same movement amount, and The moving direction and the moving amount of the image indicated by a large amount of the same type of vector information are processed as an image motion vector for the entire screen of the photographing apparatus. A method of subtraction or addition in the motion vector correction circuit 4 will be described with reference to FIG. 4. When the current field of the currently input video signal is the A field, the signal subjected to correlation calculation processing in the correlation calculation circuit 2 is A. As shown in FIG. 4 from the vertical component of the motion vector of the image detected by the motion vector detection circuit 3, since it becomes the previous representative point signal obtained from the video signal of the field and the video signal of the B field one field before. The indicated scan position deviation amount D is subtracted. This subtraction is a motion vector detected in the image scanned in the A field with respect to the image scanned in the B field before, even though the position of the pixel to be correlated moves upward (the direction of subtraction). The purpose is to correct that it is not reflected in. At this time, the switch circuit 4c is a subtractor
Try to choose 4b.

Conversely, the current field of the currently input video signal is B
In the case of a field, the scanning position shift amount D shown in FIG. 4 is added from the vertical component of the motion vector detected by the motion vector detection circuit 3. This addition is a motion vector detected in the image scanned in the B field with respect to the image scanned in the A field previously, even though the position of the pixel to be correlated is moved downward (direction of addition). The purpose is to correct that it is not reflected in.
At this time, the switch circuit 4c selects the adder 4a.
The above operation is performed every time the field on the screen of the image pickup apparatus changes, and the image motion vector corrected by such subtraction or addition is output as the final image motion vector.

As described above, according to the present embodiment, the motion vector correction circuit 4 corrects the image motion vector detected by the motion vector detection circuit 3, so that the image pickup screen inevitable by the conventional image motion detection device is not obtained. The error in the vertical direction of the image motion vector can be reduced.

Further, the configuration of the motion vector correction circuit 4 is divided into three parts of an adder 4a, a subtractor 4b, and a switch circuit 4c as shown in FIG. 2, but it is not limited to this. It may be a motion vector correction circuit that combines a motion vector of an image with a fixed vector having a fixed direction and size. Further, the same effect can be obtained by applying the present invention to not only the shake correction of the image pickup operation but also the adaptive aperture correction using the image motion vector, the adaptive noise reduction, and the like.

〔The invention's effect〕

As described in detail above, according to the present invention, when detecting an image motion vector in an image pickup screen for a video signal in an image pickup apparatus which is scanned by an interlace system, a vertical direction between fields by interlaced scanning is detected. The error of the image motion vector, which is inevitably caused by the pixel shift,
This can be corrected by providing a motion vector correction circuit.
This makes it possible to improve the detection accuracy of the image motion vector of the image pickup device, hold the image pickup device directly by hand,
Even if the imaging device is mounted on a vibrating moving body such as a vehicle, it is possible to reliably correct the shake of the screen display during imaging.

[Brief description of drawings]

FIG. 1 is a block diagram of an image motion detecting device according to an embodiment of the present invention, FIG. 2 is a block diagram of a motion vector correcting circuit shown in FIG. 1, and FIG. 3 is a block diagram of a conventional image motion detecting device. FIG. 4 is a state diagram showing fields in interlaced scanning. 1 ... Representative point storage circuit, 2 ... Correlation calculation circuit, 3 ... Motion vector detection circuit, 4 ... Motion vector correction circuit, 4a
...... Adder, 4b …… Subtractor, 4c …… Switch circuit.

Claims (2)

(57) [Claims] 1. A video signal divided into a plurality of areas holds a video signal at a representative point of each area of the current field as a modern front point signal, and outputs a previous representative point signal at a representative point of each area of the previous field. A representative point storage circuit, a correlation calculation circuit that performs a correlation calculation process between the previous representative point signal from the representative point storage circuit and the corresponding video signal of the current field, and an image from the calculation result of the correlation calculation circuit. A motion vector detection circuit for detecting the motion vector of the motion vector, and an image motion detection device for use in an interlaced video signal processing device, comprising: A motion vector correction circuit for performing vector correction by adding or subtracting the vertical scanning position shift amount between the current field and the previous field Image motion detecting device for. 2. The image motion detection device according to claim 1, wherein the scanning position shift amount of the motion vector correction circuit is half the line interval in one field.