JP2003058892A - Device and method for processing animation - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device and a method for analyzing and processing a moving image in real time. SOLUTION: The animation processing device for analyzing a moving object from each of image data continuously picked up as animation by a digital image pickup means and transmitted, is composed of the digital image pickup means for continuously picking up images, a main body of processor for processing the animation transmitted from digital video and a connection means for connecting the digital image pickup means and the main body of processor. The main body of processor is composed of an image data memory part for storing image data, background video memory part for storing background data, background separation operating part for separating a background image, feature point extraction operating part for extracting the feature points of a moving object, matching operating part, feature point position fixing part, moving image moving locus operating/judging part for operating the locus of the moving object, and image display part for displaying the locus of the moving object.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving image processing apparatus and a moving image processing method for processing a moving image which is a digital image continuously captured. More specifically, a moving image processing apparatus that analyzes and processes a moving object from image data continuously captured and transmitted by digital image capturing means such as a digital camera and a digital video camera,
And a video processing method.

[0002]

2. Description of the Related Art In recent years, the field of information processing based on digital images from digital image pickup means such as digital video cameras and digital cameras has been developing, and such digital image processing has become indispensable to our lives. ing. In particular, in addition to the widespread use of digital image capturing means, the speed of image processing systems such as computers has increased, and the storage capacity of digital data such as hard disks has increased in capacity. There is increasing interest in processing.

Usually, the processing of such a moving image is performed based on the methodology of motion analysis. That is, in general, the input to the analysis system for the motion of continuously moving images is a series of instantaneous images (for example, several tens of frames of continuous images per second), and the image data to be processed is enormous. In particular, it becomes enormous in proportion to the increase in input (increase in input time). In addition, there is a method in which a color mark that can be easily detected is attached to a person and the position is detected by using the mark as a mark.

[0004]

In fields such as automatic navigation and security systems, where real-time analysis is always required, fast processing time is required. There is a high demand for processing devices and processing methods. Therefore, an object of the present invention is to provide a moving image processing apparatus and a processing method capable of supporting a processing speed required in a field where real-time analysis is always required, such as an automatic navigation system or a security system. .

As a result of earnest studies in view of the above problems, the present inventors have found color feature points of an image of a moving object in a video by using RGB data in a continuous image (moving image) from a digital camera video. The present invention has been found out that the moving object can be tracked from the image by detecting the moving distance and the direction of the moving object for each frame.

[0006]

That is, the moving picture processing apparatus of the present invention is a moving picture processing apparatus for analyzing a moving object from each image data continuously picked up as a moving picture by the digital image pickup means and transmitted. A digital image pickup means for continuously picking up an object, a processing device body for processing a moving image transmitted from the digital video, and a connection means for connecting the digital image pickup means and the processing device body, The processing device main body is stored in a video data memory unit that stores image data from the image pickup unit, a background video memory unit that stores background data from the image pickup unit, the video data memory unit, and the background video memory unit. An image separation calculation unit that separates a background image from image data, and a background image is separated from the image data by the image separation calculation unit. A feature point extraction calculation unit that extracts a feature point of a moving target from an image, a matching calculation unit that performs matching processing of the feature points extracted by the feature point extraction calculation unit, and a matching calculation unit that is extracted by the feature point extraction calculation unit And a moving image moving locus calculation / judgment unit that calculates the locus of the moving target from the characteristic point of the moving target, which is fixed by the characteristic point position fixing unit. A moving image moving locus display unit that displays a moving object locus from a moving object characteristic point calculated by the moving image moving locus calculation / judgment unit (claim 1). .

According to the moving picture processing apparatus of the present invention, it is possible to trace only the characteristic points by extracting the moving characteristic points from the moving picture. The parts can be selectively processed.

In the moving image processing apparatus of the present invention, the image separation calculation section calculates the image signals (R, G, B) from the digital image pickup means by the following equation (1).

[0009]

[Equation 9]

The image RGB _diff (x, y) of the moving object is separated from the background image RGB _back (x, y) from the input image RGB (x, y) according to From the image of the moving object separated by the separation calculation unit, the distribution of the image data of the pixel matrix forming the image of the moving object or the center of the matrix showing a large value in the color space is set as the characteristic point, and the following equation ( 2)
~ (6):

[0011]

[Equation 10]

The feature point T of the object to be moved is extracted according to any one of the expressions (Claim 2). With this configuration, it is possible to easily separate the background image from the input image and calculate the moving feature point.

Further, in the moving image processing apparatus of the present invention,
When extracting a feature point, the feature point extraction calculation unit deducts from the variance value a value that decreases with increasing distance from the feature point around the already extracted feature point. , Equation (7)

[0014]

[Equation 11]

( _Wherein RGB _max , (x _cpi , y
_cpi ) is the calculated value and the position of the feature point according to any one of the equations (2) to (6), a is a positive constant, F is an increasing function, and RGB ′ _max is the processing result by the equation (7). )
The feature points are dispersed according to (Claim 4). With such a configuration, it becomes possible to appropriately disperse the characteristic points, and it becomes possible to disperse the characteristic points throughout the moving object.

Further, in the moving image processing apparatus according to the present invention, the matching calculation section may have a pixel information distribution similarity or a pixel information histogram similarity in an area closest to an area centered on the extracted feature point. Find the destination of the feature point from Equation (8):

[0017]

[Equation 12]

(Where k is a horizontal position value, l
Is a position value in the vertical direction) and changes in feature points are tracked (claim 5). With such a configuration, it is possible to perform image matching. Then, the feature points of the image thus matched are compared with the feature points of the next image,
It becomes possible to track a moving object.

The moving picture processing method of the present invention is a moving picture processing method for analyzing a moving object from each image data which is continuously picked up as a moving picture by the digital image pickup means, and which is (S1) When the first image is read by the means, (S2) the background image data of the first image is stored in the RGB background buffer, and (S3) the process of steps S1 and S2 is repeated, and then there is a moving object. The moving object is read, (S4) the input data is loaded into the RGB buffer (S5) the background image data is taken from the first image data, (S6) the characteristic point of the moving object is determined from step S5, and (S6) The positions of the characteristic points are calculated from the characteristic points obtained in step S6, and (S7) the nth image is obtained by the digital image pickup means. Capturing, (S8) repeating steps S3 to S7, and (S10) comparing the feature point of the n-th image with the feature point of the (n + 1) -th image (claim 6). ).

According to the moving picture processing method of the present invention, by extracting the characteristic points of the moving object from the moving picture, it becomes possible to track the movement by comparing only the characteristic points with the previous image, so that it is possible to shorten the movement. It is possible to selectively process a necessary portion from a large-capacity image in time.

In the moving image processing method of the present invention, the image signals (R, G, B) are expressed by the following equation (1).

[0022]

[Equation 13]

It is preferable to separate the moving image RGB _diff (x, y) from the input image RGB (x, y) in accordance with the background image RGB _back (x, y),
(Claim 7) Further, from the image of the object that moves the characteristic points obtained in step S5, the center of the matrix showing a large dispersion of the image data of the pixel matrix forming the characteristic points or the dispersion in the color space is determined. As the characteristic points, the following equations (2) to (6):

[0024]

[Equation 14]

It is preferable to extract the feature point T of the object to be moved according to any one of the expressions (claim 8), and when extracting the feature point, the feature point is extracted around the already extracted feature point. As a center, a value that decreases with an increase in the distance from the feature point is subtracted from the variance value, and the extracted feature point is expressed by equation (7).

[0026]

[Equation 15]

( _Wherein RGB _max , (x _cpi , y
_cpi ) is the calculated value and the position of the feature point according to any one of the equations (2) to (6), a is a positive constant, F is an increasing function, and RGB ′ _max is the processing result by the equation (7). ) Is used to disperse the feature points (claim 9), and the region closest to the region centered on the extracted feature point is the similarity of the pixel information distribution or the histogram of the pixel information. The destination of the feature point is searched for from the degree of similarity, and the image obtained in step S6 is calculated using equation (8).

[0028]

[Equation 16] It is preferable to track the change of the feature points according to the equation (k is a horizontal position value and l is a vertical position value). With this configuration, each step can be easily executed.

[0029]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of a moving image processing apparatus and a moving image processing method according to the present invention will be described below in detail with reference to the accompanying drawings. The present invention is not limited to these embodiments.

FIG. 1 is a schematic view showing the whole moving picture processing apparatus according to the present invention, and FIG. 2 is a schematic view showing the structure of the processing apparatus main body of the moving picture processing apparatus according to the present invention. Figure 3
Shows an example of an image input / processed in the present invention, FIG. 3 (a) shows an image before processing, FIG. 3 (b) shows an example of a background image separated by image separation processing, and FIG. 3C shows an example of a moving target image separated from the background image. FIG. 4 is a diagram showing an example of extracting color feature points in the present invention, and FIGS.
It shows in (e). FIG. 4A shows the equation (2) as shown in FIG.
Is the equation (3), FIG. 4 (c) is the equation (4), and FIG.
4A and 4E are diagrams showing color feature points when the formula (5) is used and the formula (6) is used. 5 (a) to 5
4E is a diagram showing an example in which the feature points in FIGS. 4A to 4E are dispersed. 6 and 7
FIG. 5 is a diagram showing an example in which a positional relationship with a next matching block is compared for each frame indicating a feature point.

First, the overall structure of the moving image processing apparatus of the present invention will be described with reference to FIGS. 1 and 2. As shown in FIG. 1, the moving image processing apparatus of the present invention includes a digital image pickup means 1 capable of continuously picking up images of a digital camera, a digital video camera or the like as a moving image, and a moving image transmitted from the digital image. The processing device main body 2 (computer system) for processing and the digital image pickup means 1 and the processing device main body 2 are mainly connected to each other, and the connection means 3 is provided for transmitting and receiving digital data.

The processing apparatus main body 2 in the present invention is usually a predetermined operating system (for example, WIN).
It is driven by a DOWS (registered trademark) type OS, a UNIX (registered trademark) type OS, a Mac (registered trademark) OS, etc., and a display connected to a central operator (not shown), a memory, a hard disk, a video card, and the video card. It is mainly composed of a display which is a means.

As shown in FIG. 2, the image processing apparatus main body 2 includes a camera image capturing section A for transmitting and receiving data to and from the digital image pickup means, an image separation calculation section P1, a feature point extraction calculation section P2, A matching calculation unit P3, a feature point position determination unit P4, calculation units P1 to P5 such as a moving image movement locus calculation / judgment unit P5, a camera image data acquisition control unit C,
The image data memory unit M1, the background image memory unit M2, the feature point position determination unit M3, and the moving image movement locus display unit D are included. The image data memory unit M2, the background image memory unit M2, and the feature point position determination unit M3 may be a storage area secured in the hard disk or a temporary memory such as RAM, and each operation unit P1.
~ P5 are arithmetic programs stored on the hard disk.

Next, a moving image processing operation by the moving image processing apparatus of the present invention having such a configuration will be described. In this description, it is assumed that the digital image pickup means 1 is fixed.

First, as shown in FIG. 2, the initial image (background image) which does not include a moving object and which is imaged by the image pickup means 1 is obtained from the image data from the image pickup means 1 via the camera image capturing section A. It is stored in the background image memory unit M2. Then, the image data RGB _max shown in FIG.
2 is a video data memory unit M1 for storing the image data from the image pickup means 1 through the connection means 3 as shown in FIG.
Memorized in. Then, the image data RGB _max stored in the video data memory unit M1 is subjected to a background image separation process from the entire image by the first calculation unit P1, and FIG.
The background image RGB _back shown in FIG. 3B and the moving target RGB _diff shown in FIG. 3C are separated. At this time, the image separation process is performed by the following equation (1).

[0036]

[Equation 17]

The threshold value at this time is appropriately determined by the average value of the previous frame. At this time, it is assumed that the imaging means is fixed, but it is also possible to move it appropriately by using a platform or the like. In this way, the background image RGB
Image RGB of moving object separated from _{back
The diff} requires a very complicated calculation and time to find and track the motion of a moving object in consideration of all the pixels.

Therefore, in the present invention, the image RGB _diff of the moving object is transmitted to the feature point extraction calculation section P2 in the _subsequent stage. The feature point extraction calculation unit P2 extracts and specifies some color feature points T instead of all the pixels to be moved.

In the present invention, in order to accurately match the color feature point of the previous frame and the position corresponding to the next frame, the process of extracting this color feature point includes The image is scanned with a matrix centered around the points and the pixels R. The center of the matrix that maximizes the value of the combination of G and B values will be the feature point. From such a concept, it is preferable to use any one of the following formulas (2) to (6) when extracting the color feature points.

[0040]

[Equation 18]

In this way, from the moving image, equations (2)-
An example of the color feature points extracted using any of (6) is shown in FIGS. 4 (a) to 4 (e). FIG. 4 (a) shows the equation (2).
4 (b) is the equation (3), and FIG. 4 (c) is the equation (4).
4 (d) shows the color feature points when the equation (5) is used, and FIG. 4 (e) shows the color feature points when the equation (6) is used. In the figure, a, b, ... Show the extracted feature points. With such a configuration, it is possible to handle not the entire moving object but the moving object using the extracted color feature points.

Further, the color feature points thus extracted are the same as the positions of the feature points in the subsequent stage, and the feature point position determination unit P4 is used.
It is also possible to confirm by
It may be partly gathered and may not be suitable for representing the whole moving object. Therefore, it is preferable to disperse the partially collected feature points in the feature point extraction calculation unit P2. The distribution of such feature points is
The feature point boundary technique of the present invention can be used to achieve this by using an algorithm that distributes feature points over a moving object.

As such an algorithm, for example, equation (7)

[0044]

[Formula 19]

( _Wherein RGB _max , (x _cpi , y
_cpi ) is the calculated value and the position of the feature point according to any one of the equations (2) to (6), a is a positive constant, F is an increasing function, and RGB ′ _max is the processing result by the equation (7). ) Can be represented by.

5 (a) to 5 (e) respectively show FIG. 4 (a).
4 shows the result obtained by dispersing the characteristic points shown in FIG. 4 (e) by the equation (7). In the figure, a, b, ... Show feature points dispersed corresponding to the feature points a, b, ... Shown in FIG. As shown in FIGS. 5A to 5E, it can be seen that the characteristic points a, b, ... Located at the center of the moving object are distributed in a wide range.

Then, the feature points dispersed as described above in some cases are sequentially processed by the matching calculation unit P3, for example, 8 ×.
It is expanded to a square matching block of 8 pixels, the position of the feature point is determined by the feature point position determination unit P4, the locus of the moving target from the feature point of the moving target is calculated by the moving image moving locus calculation / determination unit P5, and The moving image moving locus display unit D displays the moving object locus from the moving object feature point.

That is, for example, the data determined by the arithmetic unit P4 is sequentially stored in the memory M3, and the preceding and following matching blocks are sequentially compared with the corresponding blocks by the moving image movement locus calculation / judgment unit P5 to trace the movement of the moving object. To do. In this comparison, the positions of the matching blocks of the previous frame in the horizontal direction and the vertical direction are confirmed. The change value of the comparison at this time is, for example, Equation (8).

[0049]

[Equation 20]

(Where k is a horizontal position value, and l
Is the position value in the vertical direction). In this way, the result of tracking the movement of the moving object is displayed on the moving image movement locus display unit D. That is,
For example, as shown in FIGS. 6 and 7, the positional relationship with the next matching block for each frame, that is, the changing portion of the values of k and l indicates the movement distance and direction. Therefore, it becomes possible to easily analyze the characteristics and speed of the movement of the moving object.

Next, a moving image processing method according to the second embodiment of the present invention will be described with reference to FIGS. 8 is a flowchart showing the whole process of the moving image processing method of the present invention, and FIG. 9 is a flowchart showing the process of separating the background image and the moving target image in the moving image processing method of the present invention. Is a flowchart showing a process of calculating feature points in the moving image processing method of the present invention, and FIG. 11 is a flowchart showing an example of a process of performing image matching in the moving image processing method of the present invention.

As shown in FIG. 8, in step S0,
Waiting for an input command. Then, in step S1, the first image is read from the digital image pickup means. The first image at this time is a background image in which there is normally no moving object. Then, in step S2, background image data is stored in the RGB background buffer for the image thus read.

Then, when there is a moving object, that is, when the digital imaging means images the moving object, the moving object is read (step S3), and the input data is read as R.
It is taken into the GB buffer (step S4). And
In step S5, the background image thus obtained and the moving target image are separated. The background image separation process is performed, for example, as shown in FIG. 9 described later.

Then, it is judged whether or not the characteristic point is determined by this separation process (step S6), and if the characteristic point is not determined, the characteristic point is determined (step S).
7). When the feature point is determined in step S6, or after the feature point is determined in step S7, a matching process (step S8) is performed and comparison with the feature point of the next image is performed (step S9).

The image separation process in step S5 is performed, for example, as shown in FIG. That is, RG
The vertical position y of B is set to 0 (step S1
1), the horizontal position x is set to 0 (step S1)
2). And moving the target of _{RGB di} ff (x, y) of the entire image from the _{RGB back} of the background image RGB (x,
y) Obtained as a subtracted value (step S13). Then, it is determined whether or not this RGB _diff (x, y) is larger than a predetermined threshold value (step S14).

[0056] _RGB diff (x, y) if it is less than the threshold value _RGB diff (x, y) was set to 0 (step S14), and then increases the value of x (step S
15). If RGB _diff (x, y) is greater than or equal to the threshold value, the value of x is increased as it is in step S16.

Next, it is judged whether or not the value of x increased in step S17 is smaller than 320, and if it is smaller, the process returns to step S13 (step S18). When the value of x is larger than 320, the value of y is increased (step S19: 41). Next, in step S20, it is determined whether or not the value of y is smaller than 240. If the value of y is smaller than 240, the process returns to step S12.
If the value of y is greater than 240, the image separation process is completed (return).

The determination of the characteristic points in S6 in FIG. 8 is performed according to the flow shown in FIG. 10, for example.
That is, RGB _max is obtained according to equations (2) to (6) (step S21). Then, i = 0 is set (step S22), the specific point CP [i] is selected, and its position is held (step S23). Then, the characteristic points are dispersed according to, for example, the above-mentioned equation (7) (step S24), i is increased (S25), and it is determined whether or not i is less than the required amount of characteristic points. If i is less than the required amount of feature points, the process returns to S23 and continues. I
If is greater than or equal to the amount of required feature points, then the feature points are determined. The matching process of step S8 in FIG. 8 is performed according to the flow shown in FIG. 11, for example.

First, the horizontal position k is set to 0 (step S31), and the vertical position l is set to 0 (step S32). Then, the position M (k, l) of the feature point is calculated using the calculation formula shown in the formula (8) (step S33). Next, it is determined whether M (k, l) is less than Mmin (step S34). M (k, l)
Is less than Mmin, the Mmin value is set to M (k, l).
Is set (step S35), the l value is increased (step S36), and it is determined whether or not the l value is lmax (step S37).

On the other hand, if M (k, l) is greater than or equal to Mmin, it is determined in step S37 whether the l value exceeds lmax. If 1 is less than lmax, then k
The value is increased (step S38), and k is kmax
It is determined whether or not it is less than (step S39). If k is less than kmax, the process returns to step S32 to continue the process. On the other hand, when k exceeds kmax, the matching is completed.

With such a method, it is possible to trace only the characteristic points by extracting the characteristic points of the moving object from the moving image, so that the necessary portion can be selected from the large-capacity image in a short time. Can be processed as desired.

[0062]

As described above, the present invention has the following excellent effects. According to claim 1. By extracting the feature points of the moving object from the moving image, it is possible to track only the feature points, so that it is possible to selectively process a necessary portion from a large-capacity image in a short time. Further, it becomes possible to extract the feature points with high accuracy. According to the second aspect, it is possible to easily separate the background image from the input image and calculate the feature point of the moving target.

According to the third aspect, the characteristic points can be preferably dispersed, and can be dispersed throughout the moving object. Therefore, it is possible to capture the entire movement of the moving object with few feature points. According to claim 4, it is possible to perform image matching. Then, the feature points of the image matched in this way are compared with the feature points of the next image, and it becomes possible to track the moving object.

According to the fifth aspect, by extracting the characteristic points of the moving object from the moving image, it becomes possible to track the movement by comparing only the characteristic points with the previous image, and therefore, it is possible to make a large movement in a short time. It is possible to selectively process a necessary portion from the image of the capacity. Also, the feature points are extracted accurately,
Moreover, as a result of processing a small number of feature points, it becomes possible to process at extremely high speed.

According to the sixth to ninth aspects, each step can be easily executed.

[Brief description of drawings]

FIG. 1 is a schematic diagram showing an entire moving image processing apparatus according to the present invention.

FIG. 2 is a schematic diagram showing a configuration of a processing device main body in the moving image processing device of the present invention.

3A and 3B show an example of an image input / processed in the present invention. FIG. 3A is an image before processing, and FIG.
An example of a separated background image is shown by the image separation process, and FIG. 3C shows an example of a moving target image separated from the background image.

FIG. 4 is a diagram showing an example of extracting color feature points in the present invention.

FIG. 5 is a diagram showing an example in which the characteristic points in FIG. 4 are dispersed.

FIG. 6 is a diagram illustrating an example in which a positional relationship with a next matching block is compared for each frame indicating a feature point.

FIG. 7 is a diagram illustrating an example in which a positional relationship with a next matching block is compared for each frame indicating a feature point.

FIG. 8 is a flowchart showing the entire process of the moving image processing method of the present invention.

FIG. 9 is a flowchart showing a process of separating a background image and a moving target image in the moving image processing method of the present invention.

FIG. 10 is a flowchart showing a process of calculating feature points in the moving image processing method of the present invention.

FIG. 11 is a flowchart showing an example of a process of performing image matching in the moving image processing method of the present invention.

[Explanation of symbols]

M1 to M5 memory section P1 to P5 calculation unit 1 Digital imaging means 2 Processing device body 3 connection means

Continued front page    (72) Inventor Takashi Yamaguchi             Saitama Prefecture Minami Saitama District Miyashiro Town Gakuendai 4-1 Japan             Inside the university F-term (reference) 5B050 DA04 EA04 EA09 GA01                 5B057 AA19 BA11 CA01 CA08 CA16                       CB01 CB08 CB20 CH01 CH11                       DA07 DA11 DA17 DB06 DB09                       DC25                 5L096 AA02 AA06 BA02 CA04 FA15                       FA33 FA35 FA66 GA51 HA04                       HA05 JA03 JA11 LA01 LA05

Claims (10)

[Claims] 1. A moving image processing apparatus for analyzing a moving object from each image data transmitted and continuously captured as a moving image by the digital image capturing means, the digital image capturing means for continuously capturing the object. A processing device body for processing a moving image transmitted from the digital image pickup means, and a connection means for connecting the digital image pickup means and the processing device body, wherein the processing device body is image data from the image pickup means. A video data memory section for storing the background data, a background video memory section for storing the background data from the image pickup means, an image separation calculation section for separating a background image from the image data stored in the video data memory section and the background video memory section, A feature point for extracting a feature point of a moving target from an image in which a background image is separated from the image data by an image separation calculation unit Output calculation unit, a matching calculation unit that performs matching processing of the characteristic points extracted by the characteristic point extraction calculation unit, and a characteristic point that determines the position of the characteristic points extracted by the characteristic point extraction calculation unit and subjected to matching processing by the matching calculation unit A position determination unit, a moving image moving locus calculation / judgment unit for calculating a locus of a moving object from the characteristic point of the moving object decided by the characteristic point position fixing unit, and a moving object calculated by the moving image movement locus calculation / judgment unit. A moving image processing apparatus, comprising: a moving image movement locus display unit that displays a locus of an object that moves from a feature point. 2. The image separation calculation unit calculates the image signals (R, G, B) from the digital image pickup means by the following formula (1): The input image RGB (x, y) from the moving target image RGB _diff (x, y) according to the background image RGB.
_back (x, y), and the feature point extraction calculation unit distributes the image data of the pixel matrix forming the image of the movement target from the image of the movement target separated by the image separation calculation unit, or the color. Using the center of the matrix showing a large value of spatial variance as a feature point, the following equation (2)
(6): [Equation 2] The moving image processing apparatus according to claim 1, wherein the feature point T of the target to be moved is extracted according to any one of the expressions. 3. The moving image processing according to claim 2, wherein the matching calculation unit divides the image of the moving target that is separated when the feature points are extracted, and extracts the feature points for each division. apparatus. 4. The matching calculation section extracts feature points.
Center the feature points around the already extracted feature points.
As the distance from the feature point increases
The value is subtracted from the dispersion value to obtain the formula (7). [Equation 3] (In the formula, RGB_max(X_cpi, Y_cpi) Is each formula
Positions of calculated values and feature points according to any of (2) to (6)
Where a is a positive constant, F is an increasing function, and RGB '
_maxIs the processing result by equation (7))
The points are dispersed, according to claim 2 or 3.
Video processing device. 5. The matching calculation unit searches for a destination of a feature point in a region closest to a region centered on the extracted feature point from the similarity of pixel information distribution or the similarity of histogram of pixel information. Equation (8): The change of the feature point is tracked according to (wherein, k is a position value in the horizontal direction and l is a position value in the vertical direction) according to any one of claims 2 to 4. Video processing device. 6. A moving image processing method for analyzing a moving object from each image data transmitted and continuously imaged as a moving image by the digital image pickup means, wherein: (S1) the first image from the digital image pickup means. (S2) The background image data of the first image is stored in the RGB background buffer, and (S3) After the processes of steps S1 and S2 are repeated, if a moving object exists, the moving object is read (S4). ) Inputting the input data into the RGB buffer (S5) Separating the background image data from the first image data, (S6) Determining the target feature point to move from step S5, (S6) obtained in step S6 The position of the characteristic point is calculated from the characteristic point, and (S7) the nth image is captured from the digital imaging means, (S8) step S A method of processing a moving image, comprising repeating steps 3 to S7, and (S10) comparing the feature point of the nth image with the feature point of the (n + 1) th image. 7. In step S5, the image signals (R, G, B) are expressed by the following equation (1): According to the input image RGB (x, y), the moving target image RGB _diff (x, y) is _converted into the background image RGB.
The moving image processing method according to claim 6, wherein the moving image processing method is separated from _back (x, y). 8. From the image of the moving object that moves the feature points obtained in step S5, the variance of the image data of the pixel matrix forming the image of the moving object or the center of the matrix showing a large variance in the color space is set. As the characteristic points, the following equations (2) to (6): 8. The moving image processing method according to claim 7, wherein the feature point T of the target to be moved is extracted according to any one of the expressions. 9. When extracting a feature point, it has already been extracted.
Is the feature point centered around the feature point?
The value that decreases with increasing distance from
Subtract and extract the extracted feature points using equation (7). [Equation 7] (In the formula, RGB_max(X_cpi, Y_cpi) Is each formula
Positions of calculated values and feature points according to any of (2) to (6)
Where a is a positive constant, F is an increasing function, and RGB '
_maxIs the processing result by equation (7))
The process according to claim 8, wherein the process of dispersing
How to process the video. 10. The destination of the feature point is searched for from the similarity of the pixel information distribution or the similarity of the pixel information histogram in the region closest to the region centered on the extracted feature point,
The image obtained in step S6 is represented by equation (8) 10. The moving image processing method according to claim 8 or 9, wherein changes in feature points are tracked according to (wherein, k is a horizontal position value and l is a vertical position value). .