JPH0687260B2 - Still image processing method - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a still image processing method for performing a panning process by shifting a part or all of a one-screen still image in a predetermined direction.

<Prior Art> As a method of optically performing the above-mentioned flow processing at the time of photographing, there are flow photographing, zooming during exposure and the like. The follow shot captures the movement of a moving subject (car, person, bicycle, etc.) at a slow shutter speed to stop the moving subject and move the background to obtain a sense of speed. Is.

On the other hand, zooming during exposure involves changing the focal length of the zoom lens during exposure to emphasize the subject at the center of the screen and let the image around the screen flow in the radial direction.

The shooting method as described above requires a very advanced technique.

Therefore, it is conceivable that the effect of the panning process can be obtained by performing the electric image processing on the image or the like in which the background is not flowing at the time of photographing.

However, for example, in the case of panning, there are the following problems.

In FIG. 2, A is an original still image, B is a car as a subject to be stopped, C is a background area to be flow processed, and D is a vector arrow indicating the direction and amount of flow processing.

When the flow process is executed, each image in the background area of the original image is shifted in the direction indicated by the arrow D, but the area E
Since there is no original image in C, it cannot be shifted into region C.

Further, since the original image does not exist in the area F after the shift due to the shadow of the car, the shift cannot be performed.

For this reason, there is a problem in that it is impossible to obtain an effective flow-through treatment effect.

<Objective> An object of the present invention is to provide a still image processing method capable of solving the above-mentioned problems and executing effective flow processing.

Here, the panning process means that the above-described optical panning or zooming during exposure is performed by image processing.

The present invention stores the original still image in the first memory, shifts the pixel data of the image in the first memory to be shifted in a predetermined direction, and stores the shifted pixel data again in the first memory. Then, the shifted pixel data is integrated into the second memory, and the flow processing is executed by repeating the shift processing, the re-storing processing, and the integration processing, and the data of the pixel in the specified still area of the original still image is executed. Is characterized by maintaining the value of the first memory.

<Embodiment> FIG. 1 shows a control block diagram of an embodiment of the present invention.

In the figure, reference numeral 11 is a region B in FIG. 2, that is, a stationary region designating unit for designating a stationary region, which is constituted by a mouse or the like. 13 is CPU
To execute the whole control. Reference numeral 15 is a RAM for temporarily storing data required for processing by the CPU 13. An image controller 17 is an interface between the computer bus CB and the image bus IB. The ROM 19 stores the control program of the CPU 13. A parameter input unit 20 inputs parameters of the direction and distance of the flow processing. 21 is an A / D converter that converts the input still image signal IN into a digital value, 23 is an image processor that executes image processing, 25 is an R, G,
B Image memory that stores each screen separately, 27, 29, 31 are work memories for image processing, 33 is an integrated screen memory that stores calculation results, and 35 is a processed image converted into an analog signal and output. It is a D / A converter.

FIG. 3 shows a processing program for obtaining the follow shot effect when the image processing apparatus having the above-described configuration is used.

This will be described below with reference to FIG.

First, in step S1, the operator designates the stationary region (FIG. 2B) by the designating unit 11, and designates the flowing direction and distance by the parameter input unit 20.

Next, the loop specified for defining the area is often broken, so the edge part of the loop is extracted and the still area becomes a closed loop by utilizing the processing of diffusion and thinning of the extracted edge. Process (S2). Then, a work map 1 showing a bit map showing a stationary area surrounded by a closed loop
It is stored in (Fig. 1, Fig. 27) (S3).

Next, the contents of the work memory 1 are copied to the work memory 2 (S4), and the work memory 2 is further shifted by the distance 1 in the flowing direction (S5).

And the work map of work memory 1 and work memory 2
The logical sum of the above bit maps is taken and the bit map is put into the work memory 2 again (S6).

Further, a bit is set in the work memory 2 corresponding to a line at the end of the screen (FIG. 2G) in the direction opposite to the flowing direction (S7).

By the above operation, the still area, the area obtained by shifting the still area by the distance 1 and the bit map indicating the screen end portion in the direction opposite to the flowing direction are formed in the work memory 2.

Next, the entire R screen of the original image is copied to the work memory 3, and L indicating the moved amount is set to 0 (S8). Then, the screen of the work memory 3 is shifted by a distance of 1 in the flowing direction and stored again in the work memory 3 (S9).

The contents of the work memory 3 will be sequentially added to the integrated screen memory 33, but if they are simply added, they will flow to the still area. Therefore, only the still area needs to be returned to the original image position. Further, the portion moved from the back side of the stationary area (Fig. 2F) is not the image of the background (Fig. 2C). or,
No image exists at the original image position to be flowed (FIG. 2E) at the screen edge portion in the direction opposite to the flow direction.

Therefore, the work memory 2 created in steps S1 to S7 is used. That is, in step S10, the original image (R image: stored in the image memory 25) of the part where the bit is standing in the work memory 2 is taken out, replaced with that part of the work memory 3, and again the work memory 3 is read. To store. By doing so, the static area portion in the work memory 3 is returned to the original image position. Also, the part moved from the still area is replaced with the background information of the original position,
The background information at the original position is also substituted for the portion moved from the screen end opposite to the flowing direction.

The image in the work memory 3 obtained by the above operation is added to the integrated image memory 33 (S11). Then, since one movement is completed, L is incremented and the operations of S9 to S13 are repeated until the distance at which L flows is reached.

The integrated screen memory 33 needs to have a bit number that does not cause an overflow even if the distance for flowing the maximum number of gradations of the image is added to one dot.

When the screen shift is completed by the flow distance, the data of each dot in the integrated screen memory is divided by the flow distance L to obtain the average density for each dot (S15). As described above, the flow process for the R screen is completed.

Next, for the G and B screens, the contents of the work memory 3 and the integration screen memory 33 are reset and the operations of S8 to S14 are repeated to complete the final panned image.

The images after panning processing and the original images obtained as described above are attached as Reference Photo 2 and Reference Photo 1, respectively.

In this embodiment, the flowing direction is set parallel to one edge of the image, but it may be set diagonally. In this case, the screen edge portion opposite to the flowing direction is at two positions, but the same processing can be performed.

Further, in the above embodiments, the flow processing has been described as an example, but the present invention can of course be applied to other flow processing such as zooming during exposure. In the case of zooming during exposure, the flow direction is radial, so the flow processing may be performed in each direction. Further, although the example in which only a part of the screen is flown has been described, the present invention can be applied to the case where the entire screen is flown.

<Effect> As described above, according to the present invention, not only optical flow processing, zooming during exposure, etc. can be realized by image processing, but also the still region can be surely stopped, so that a new still image without discomfort can be created. it can.

[Brief description of drawings]

FIG. 1 is a control block diagram of the image processing apparatus of this embodiment, FIG. 2 is an explanatory diagram of follow shot processing, and FIG. 3 is a control flow chart of the follow shot processing.

Front page continuation (72) Inventor Susumu Matsumura Kanagawa Prefecture Kawasaki City Takatsu-ku 770 Shimonoge Office, Tamagawa Works Canon Inc. (72) Inventor Hiroshi Omura, Tamano Works, Canon Inc., 770, Shimonoge, Takatsu-ku, Kawasaki City, Kanagawa Prefecture

Claims (1)

[Claims] 1. A still image for obtaining a new still image in which a part or all of the original still image appears to flow in a predetermined direction by performing image processing on pixels constituting the original still image of one screen. A method of processing, comprising storing the original still image in a first memory, shifting pixel data of an image in the first memory in the predetermined direction, and shifting the shifted pixel data to the predetermined direction. The image data in the first memory is stored again in the first memory, the shifted pixel data is integrated in the second memory, and the shift processing, the re-storing processing, and the integration processing are performed on the image in the first memory. Repeated a plurality of times, maintaining data of pixels of the original still image in the first memory for a specified still area of the original still image.