JP2001177836A - Method and device for detecting mobile object in moving picture - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect a mobile object with high accuracy by eliminating the effect of disturbance from moving picture signal coded data. SOLUTION: A moving picture signal processing section 1a decodes an image of an IP frame and outputs a motion vector of a P frame. An inter-frame difference calculation section 1b calculates an inter-frame difference of an IP frame decoded image. A motion vector distribution generating section 1c obtains a motion vector distribution of a macro block with a large inter-frame difference. A motion vector selection section 1d selects a motion vector whose distribution density is high. A mobile body detection section 1e detects the macro block with the large inter-frame difference and a selected motion vector as a mobile object. Selecting a motion vector with a large distribution density can exclude the effect by a distribution.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for detecting a moving object in a moving image, and more particularly to a method and an apparatus for detecting a moving object from moving image signal encoded data with high accuracy.

[0002]

2. Description of the Related Art As a conventional apparatus for detecting a moving object in a moving image, there is an apparatus disclosed in Japanese Patent Laid-Open No. 8-106534. The operation of this device will be described with reference to FIG.

[0003] The image input unit 35b takes in images taken by the camera 35a at regular time intervals. Change area detector 35c
Calculates a difference between an input image and an image input after a predetermined time, creates a difference image, determines a binarization threshold from the density distribution of the difference image, binarizes the difference image, and changes Extract the region. The area blocking unit 35d generates a block surrounding the extracted change area. Feature extractor 35
e calculates the feature amount (area, horizontal / vertical width, moving direction / distance, density distribution) of this block. Moving object identification unit
35f determines from these characteristics whether or not the object in the change area is a detection target. The result output unit 35g displays the result on the display device 35h based on the determination result.

[0004]

However, the above-mentioned conventional apparatus has a problem that an object which is not a detection target is detected. This is because conventional devices basically detect moving objects by detecting changes in luminance over time.Therefore, factors that are not detected, such as shadows and external light insertion, changes in sunshine, and camera gain control, are used. This is because even the change (disturbance) of the video signal due to the above is caused.
Therefore, even if the moving object is identified using the feature amount of the detection area, there is a problem that an object that is not a detection target is detected.

[0005] Even when a moving object is detected based on a change in luminance over time, only a part of the object may be detected, or a plurality of objects may be detected as one unit. Therefore, there is a problem that a moving object cannot be correctly identified due to a large difference in area, horizontal / vertical width, or moving direction / distance.

SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned conventional problems and to accurately detect a moving object from encoded video signal data.

[0007]

In order to solve the above-mentioned problems, the present invention provides a moving object detection method that decodes a moving image signal of an intra-frame coded image and a forward prediction inter-frame coded image. A video signal decoding processing step of outputting a motion vector of a forward prediction inter-frame coded image, an inter-frame difference calculation processing step of calculating an inter-frame difference from decoded image data, and a changed pixel in which an inter-frame difference is calculated. A motion vector distribution creation processing step of creating a distribution of motion vectors of a macroblock including: a motion vector distribution processing step of selecting a motion vector whose distribution density of the motion vector satisfies a predetermined condition to obtain a selected motion vector; A macro block including a change pixel for which an inter-frame difference is calculated and having a selected motion vector is set as a moving object area. And configured to include a moving object detection processing step of detecting by. With this configuration, by detecting the moving object using the motion vector in addition to the inter-frame difference, it is possible to eliminate the influence of disturbance and accurately detect the moving object from the moving image signal encoded data.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below in detail with reference to FIGS. In each of the drawings, components having the same function are denoted by the same reference numerals.

(First Embodiment) In a first embodiment of the present invention, a video signal of an I frame and a P frame is decoded, a motion vector of a P frame is output, and an inter-frame The difference is calculated, the motion vector distribution of the macroblock having a large interframe difference is obtained, a motion vector having a large motion vector distribution density is selected, and the macroblock having a large interframe difference and having the selected motion vector is moved to the moving object area. This is a moving object detection device that detects a moving object.

FIG. 1 is a functional block diagram of a moving object detecting device according to a first embodiment of the present invention. In FIG. 1, a video signal decoding processor 1a decodes video signal encoded data to generate an intra-frame encoded image (I frame).
And forward-predicted inter-frame coded image (P frame)
Is a means for outputting the decoded image and the motion vector of the P frame. The I frame and the P frame are collectively called an IP frame. The inter-frame difference calculation unit 1b is a unit that calculates an inter-frame difference of an IP frame decoded image. The motion vector distribution creating unit 1c is a unit that creates distribution data of a motion vector of a macroblock including a pixel having a changed inter-frame difference. The motion vector selection unit 1d is a unit that selects a motion vector based on the distribution density of the motion vector. The moving object detection unit 1e is a unit that detects a macroblock having the selected motion vector as a moving object region. FIG. 2 is a flowchart illustrating a processing procedure of the moving object detection device according to the first embodiment.

FIG. 3 is a diagram showing an example of encoded video signal data composed of an I frame, a P frame, and a B frame. In FIG. 3, an I frame 3a is an intra-frame coded image. The P frame 3b is a forward prediction inter-frame coded image. The B frame 3c is a bidirectionally predicted inter-frame coded image. FIG. 4 is a diagram illustrating an example of encoded video signal data composed of IP frames.

FIG. 5 is a functional block diagram of a moving image signal decoding processing unit according to the first embodiment. In FIG. 5, a header decoding unit 5a is a unit that decodes header information of encoded video signal data. Variable length decoding unit 5b
Is means for decoding variable-length coded data such as motion vector information and DCT coefficients. The inverse quantization unit 5c is means for inversely quantizing the decoded DCT coefficient. The IDCT unit 5d is means for performing an inverse cosine transform of the DCT coefficient.
The addition unit 5f is a unit that adds the decoded data to the predicted image data. The frame memory 5e is a memory that stores the decoded image as reference image data.

FIG. 6 is an explanatory diagram of the difference between frames.
In FIG. 6, a frame 6a is a reference I frame. The frame 6b is a P frame to be detected. The image 6c is a result of displaying changed pixels in the inter-frame difference image as black. The image 6d is an image of a person in the reference frame. The image 6e is an image of a person in the detection processing target frame. The image 6f is a pixel area corresponding to the position of the person in the reference frame. The image 6g is a pixel area corresponding to the position of the person in the detection processing target frame. The image 6h is a video of disturbance in the detection processing target frame. Image 6i is
These pixels are erroneously detected due to disturbance.

FIG. 7 is an explanatory diagram of the inter-frame difference result and the motion vector information of the P frame. In FIG. 7, a vector 7a is motion vector information of a P frame.
The macro block 7b is a macro block at a position where no person exists. The macro block 7c is an MB at a position where a person exists. The macro block 7d is a macro block at a position where a disturbance exists.

FIG. 8 is a diagram showing an example of a motion vector distribution. In FIG. 8, a distribution point 8a is an isolated distribution point. The distribution point 8b is a concentration distribution point.

The operation of the moving object detecting apparatus according to the first embodiment of the present invention configured as described above will be described.
First, moving image signal encoded data will be described. ITU-T H.261, MP
EG1, MPEG2, etc. are defined as international standards. ITU-T H.261 is used for communication media such as videophones, MPEG1 is used for storage media such as video CDs, and MP.
EG2 is used for storage media such as DVDs and communication media such as DTVs.

The frame structure of the encoded video signal data will be described with reference to FIGS. As shown in FIG. 3, 15 frames are normally defined as 1 GOP in MPEG1 and MPEG2. In the configuration of the 15 frames, the first frame in the transfer order is an I frame, and thereafter, two B frames and one P frame are periodically inserted.

The relationship between a motion vector and an image frame will be described with reference to FIG. For example, the encoded data at P5 in FIG. 3 includes difference data between forward prediction vector data using I2 as a reference image and a prediction image created based on this vector. Further, the encoded data of B3 in FIG. 3 includes difference data between a bidirectional prediction vector using I2 and P5 as reference images and a prediction image created based on the vectors.

As shown in FIG. 4, a GOP composed of only an I frame and a P frame is recognized in MPEG2. Although H.261 does not organize GOPs, the frame configuration is based on I frames and P frames, and has the same configuration as that of FIG. In the present embodiment, a case will be described as an example in which moving image signal encoded data having the frame configuration shown in FIG. 3 is used as input data. However, the present invention is limited to moving image signal encoded data having the above frame configuration. Not something.

Second, the operation of the moving picture signal decoding section 1a will be described. The moving image signal decoding processor 1a decodes the input moving image signal encoded data. The input coded video signal data has three types of frames, i.e., an I frame, a P frame, and a B frame. Output vector information.

Here, the operation of the video signal decoding processor 1a will be described in more detail with reference to FIG. The header decryption processing section 5a decrypts the header information of the encoded video signal data. An example of header information is MPE
Sequence Header, GOP Header, Picture Hea in G
der etc. The frame type information of the image data is Pic
This is information in the frame header. In this header decoding process, an I frame, a P frame, and a B frame are identified, and in the case of a B frame, the subsequent processes are not performed.

The variable length decoding unit 5b decodes variable length coded data such as motion vector information and DCT coefficients. The variable-length decoded DCT coefficients are inversely quantized by the inverse quantization unit 5c and inverse cosine transformed by the IDCT unit 5d.

If the image is an I frame, the data after the IDCT processing becomes the decoded image information as it is. When the image is a P frame and a B frame, the addition processing unit 5f
Then, the data after the IDCT processing is added to the reference image data in the frame memory 5e and the prediction image data created based on the motion vector information, and decoded image information is obtained.

Third, the operation of the inter-frame difference calculator 1b will be described. The inter-frame difference calculator 1b calculates the difference between the decoded image information of two frames, binarizes the difference, and creates inter-frame difference information. The decoded image I0 (X, Y) obtained at a certain time and the decoded image I obtained at the immediately following time
A difference of 1 (X, Y) is calculated, and a pixel whose absolute value exceeds a predetermined threshold value is set as a changed pixel. That is, assuming that the difference image is D (X, Y) and the predetermined threshold is Dth, D (X, Y) is set to 1 for pixels satisfying | I0 (X, Y) -I1 (X, Y) |> Dth. And

The calculation of the inter-frame difference does not need to be performed on a pixel-by-pixel basis. The average value of pixels is calculated for each rectangular area such as 8 × 8 pixels or 16 × 16 pixels, and the inter-frame difference is calculated. Is also good. Alternatively, one pixel may be sampled for each rectangular area such as 8 × 8 pixels or 16 × 16 pixels, and an inter-frame difference may be calculated for the sample pixels.

Thereafter, the detection of the moving object is performed using the motion vector. However, since the motion vector cannot be obtained in the case of the I frame, the detection processing is not performed. In the case of an I frame, the difference information between frames may be directly output as a detection result.

Fourth, the operation of the motion vector distribution creating section 1c will be described. The motion vector distribution creating unit 1c extracts a motion vector of a macroblock including a changed pixel of the inter-frame difference information, and creates motion vector distribution information.

Here, the inter-frame difference and the motion vector will be described in more detail with reference to FIGS. In FIG. 6, the result of displaying the reference frame 6a as the I2 frame in FIG. 3, the detection processing frame 6b as the P5 frame in FIG. 3, and displaying the changed pixels in the inter-frame difference image as black is 6c. 6f is a pixel area 6d at the position of the person in the reference frame 6a,
6g is a pixel area 6e at the position of the person in the detection processing frame 6b. Further, an erroneously detected pixel area 6i is generated due to disturbance 6h generated in the detection processing frame 6b.

FIG. 7 is a diagram showing an example of an inter-frame difference result and P-frame motion vector information. In FIG. 7, the size of one grid is 16 pixels × 16 pixels, and H.2
61. A macroblock (MB) in MPEG is a unit of a pixel set using a motion vector. 7a is the motion vector information of the P frame, 7b is one of the MBs where the person is not present, 7c is one of the MBs where the person is present, and 7d is one of the MBs where the disturbance is present.

In the motion vector distribution creating section 1c, the M including the changed pixel areas 6f, 6g and 6i of the inter-frame difference information is calculated.
The motion vectors of B7b, 7c, and 7d are sequentially extracted. By referring to the difference image at the position corresponding to each MB, it is sequentially determined whether or not a changed pixel is included, and a motion vector V (Vx, Vy) of the MB having the changed pixel is extracted.

Then, using the vertical component Vx and the horizontal component Vy of these motion vectors, they are plotted on two-dimensional coordinates, and the motion vector distribution B (X, Y) is calculated. At this time, the distribution B (X, Y) counts the number of extracted vectors. That is, the vector V (V
When x, Vy) is extracted, 1 is added to B (Vx, Vy).

Fifth, the operation of the motion vector selector 1d will be described. The motion vector selection unit 1d selects a motion vector according to the distribution density of the motion vector distribution. The motion vectors obtained in the region of the object moving in the image are substantially the same, and when a vector distribution is created from these, the motion vectors are concentrated on the distribution. On the other hand, a motion vector obtained in a region having disturbance such as light reflection is a random vector, and the motion vectors do not concentrate on the distribution. Therefore, moving objects are detected by selecting vectors concentrated on the distribution and extracting macro blocks having these vectors.

As an example of a method of determining whether vectors are concentrated on a distribution, there is a method of determining whether a vector distribution point is an isolated point on the distribution. FIG. 8 is a diagram illustrating an example of a vector distribution. With respect to all the points where B (X, Y) is 1 (the plotted points in FIG. 8), it is determined whether or not the points are isolated points with reference to eight neighboring points adjacent thereto. If B (X0, Y0) of all eight neighboring points is 0 as in the distribution point 8a, this vector is not selected and B (X0, Y0) is set to 0.
And On the other hand, as shown in 8b, at least one of the eight neighboring points
If B (X1, Y1) of the point is 1, it is considered that the vector is concentrated, and this vector is selected, and the vector distribution B is left as it is. As described above, by removing the vector obtained in the disturbance area, only the object area can be detected with high accuracy.

Finally, the operation of the moving object detector 1e will be described. The moving object detection unit 1e detects, as a moving object region, an MB having a change pixel of an inter-frame difference and having a selected motion vector. For each MB, it is determined whether or not the difference image D (X, Y) includes a changed pixel whose value is 1, and further, the motion vector V (Vx,
Vy) on the selected vector distribution, B (Vx, V
j) is determined to be 1 and when both are satisfied,
This MB is detected as an animal region.

As described above, in the first embodiment of the present invention, the moving object detecting device decodes the I-frame and P-frame moving image signals, outputs the P-frame motion vectors, and outputs the decoded image data. Calculate the inter-frame difference from, create the motion vector distribution of the macroblock including the changed pixel with a large inter-frame difference, select a motion vector with a large motion vector distribution density, and select the large inter-frame difference. Since a macroblock having a motion vector is detected as a moving object area, the moving object can be detected with high accuracy by eliminating the influence of disturbance from the encoded video signal data.

(Second Embodiment) In a second embodiment of the present invention, when a motion vector is a field vector, the magnitude of the motion vector is normalized by the distance between fields, and then the distribution of the motion vector is determined. Is a moving object detection device that creates a moving object.

FIG. 9 is a flowchart showing a processing procedure of the moving object detection device according to the second embodiment of the present invention.
FIG. 10 is a diagram illustrating a frame vector and a field vector. In FIG. 10, an image 10a is a reference frame image. Image 10b is a processed frame image. Image 10c is a top field image. Image 10
d is a bottom field image. Image 10e is a processed top field image. Vector 10f is a frame vector. Vector 10g is a top field vector. Vector 10h is a bottom field vector. Image 10i is a processed bottom field image. Vector 10j is a top field vector. Vector 10h is a bottom field vector.

FIG. 11 is a functional block diagram of a moving object detection device according to the second embodiment. 1 are given the same reference numerals. In FIG. 11, a moving image signal decoding unit 1a is a means for decoding moving image signal encoded data and outputting a decoded image of an IP frame and a motion vector of a P frame. Inter-frame difference calculator 1b
Is means for calculating the inter-frame difference of the decoded IP frame image. The motion vector normalizing unit 11a is a means for normalizing the magnitude of the field vector of the MB including the pixel having the changed frame difference. The motion vector distribution creating unit 1c is a unit that creates motion vector distribution data. The motion vector selection unit 1d is a unit that selects a motion vector based on the distribution density of the vector distribution. The moving object detecting unit 1e calculates M having the selected motion vector.
This is a means for detecting B as a moving object area.

The operation of the moving object detecting apparatus according to the second embodiment configured as described above will be described. In the video signal decoding process and the inter-frame difference calculation process, the inter-frame difference is calculated in the same manner as in the first embodiment. In the motion vector distribution creation processing, the motion vector distribution is created after correcting the size of the field vector according to the inter-field distance.

In MPEG2, a P frame motion vector includes a frame vector and a field vector. The frame vector has one motion vector for one macroblock (MB), and the field vector is one MB (16 × 16 pixels).
Is divided into 16 × 8 pixels which are field images, each having a motion vector.

FIG. 10 is a diagram for explaining a frame vector and a field vector. In FIG. 10, the vector from the reference frame image 10a to the processing frame image 10b is shown.
10f is a frame vector. On the other hand, the processed top field image 10e includes the top field image 10c and the bottom field image 10c as reference field images.
Either 10d may be selected. At this time, the time interval between the processed field image and the reference field image differs between the top field vector 10g referring to the top field image 10c and the bottom field vector 10h referring to the bottom field image 10d.
Since the magnitude of the vector changes depending on which one is selected, the magnitude of the vector is normalized so that the magnitude does not change depending on the selected field.

In FIG. 10, the time interval between the reference image of the frame vector 10f and the processed image is 3 frames, that is, 6 fields. In the case of a field vector, the size of the vector is normalized so as to match the time interval of 6 fields. That is, since the bottom field vector 10h in FIG. 10 has a time interval of 5 fields, the magnitude of the vector is multiplied by 6/5, and the top field vector 10j is added to the processed bottom field image 10i.
Multiplies the size of the vector by 6/7 since the time interval is 7 fields.

In the motion vector distribution creation processing, it is determined whether or not each MB includes a changed pixel with reference to the difference image D (X, Y). In the case of an MB including a change pixel, it is determined whether or not the motion vector of the MB is a field vector.
Perform normalization. In the vector normalization, the horizontal / vertical components of the field vector are each multiplied by a predetermined coefficient, rounded off and converted to an integer. Then, a vector distribution is created using the frame vector and the normalized field vector. Note that the field vector is
There are two types, a top field vector and a bottom field vector. A vector distribution may be created using one of the field vectors, or a vector distribution may be created using both field vectors.

As described above, by normalizing the field vector, the vector distribution can be created with high accuracy, and thus the moving object detection accuracy is improved. After that, in the motion vector selection processing and the moving object detection processing, the same operation as in the first embodiment is performed to detect the moving object.

As described above, according to the second embodiment of the present invention, when the moving object detection device is a motion vector that is a field vector, the motion vector is normalized by the inter-field distance, Since the vector distribution is configured, it is possible to correct the variation in the time interval between the detection field and the reference field and detect the moving object with high accuracy.

(Third Embodiment) In a third embodiment of the present invention, a moving object is detected by calculating the magnitude and angle of a motion vector and creating a motion vector distribution on a magnitude / angle coordinate system. Device.

FIG. 12 is a flowchart showing a processing procedure of the moving object detecting device according to the third embodiment of the present invention. FIG. 13 is a functional block diagram of the moving object detection device according to the third embodiment. 1 are given the same reference numerals. In FIG. 13, the video signal decoding processing unit 1a decodes the video signal encoded data,
This is a means for outputting a decoded image of an IP frame and a motion vector of a P frame. The inter-frame difference calculator 1b calculates I
This is a means for calculating the inter-frame difference of the P frame decoded image. The motion vector polar coordinate calculation unit 13a is a unit that calculates the magnitude and angle of the vector of the MB including the changed pixel of the inter-frame difference. Motion vector distribution creation unit 1c
Is a means for creating a distribution of motion vectors. The motion vector selection unit 1d is means for selecting a motion vector based on the distribution density of the vector distribution. Moving object detector 1e
Is means for detecting an MB having the selected motion vector as a moving object area.

The operation of the moving object detecting apparatus according to the third embodiment of the present invention configured as described above will be described.
In the video signal decoding process and the inter-frame difference calculation process, the inter-frame difference is calculated in the same manner as in the first embodiment. In the motion vector distribution creation processing, the magnitude and the angle of the motion vector are calculated, and the motion vector distribution is created on a two-dimensional coordinate system in which the magnitude is set on the vertical axis and the angle is set on the horizontal axis.

Since the motion vector output in the moving picture signal decoding process is represented by horizontal / vertical components, it is converted into a magnitude / angle. When the motion vector is V (V
x, Vy), the magnitude r and the angle θ can be calculated by r = ｒ (Vx ² + Vy ² ) θ = tan ⁻¹ (Vy / Vx).

In the motion vector distribution creation processing, it is determined whether or not each MB includes a changed pixel with reference to the difference image D (X, Y). In the case of an MB including a change pixel, the size r
And the angle θ are calculated. Then, it is plotted at a point (r, θ) on the size / angle coordinate system, and the vector distribution B (r,
θ).

When the vector distribution is created using the horizontal / vertical components of the motion vector, the degree of concentration of the vector obtained from the object differs depending on the coordinate position. The horizontal component Vx and the vertical component Vy can be expressed as Vx = r × cos θ Vy = r × sin θ, respectively. For example, for an object with a large amount of movement, that is, an object with a large r, V
The difference between x and Vy increases. As described above, when the horizontal / vertical components are used, the distribution becomes large on the distribution coordinates,
By creating the distribution on the size / angle coordinates, the distribution density does not vary depending on the coordinate position, and the distribution densities can be easily compared.

Thereafter, in the motion vector selection processing and the moving object detection processing, a moving object is detected in the same manner as in the first embodiment.

As described above, in the third embodiment of the present invention, the moving object detecting apparatus calculates the magnitude and angle of a motion vector and creates a motion vector distribution on a magnitude / angle coordinate system. With the configuration, the motion vector distribution density can be easily compared regardless of the position on the distribution coordinates.

(Fourth Embodiment) In a fourth embodiment of the present invention, a motion vector distribution is divided into a plurality of rectangular areas, and the number of motion vectors distributed in each rectangular area is calculated. This is a moving object detection device that selects a motion vector belonging to a rectangular area in which the number of motion vectors exceeds a predetermined threshold.

FIG. 14 is a flowchart showing a processing procedure of the moving object detection device according to the fourth embodiment of the present invention. FIG. 15 is a diagram illustrating an example in which the motion vector distribution is divided into a plurality of rectangular regions. In FIG. 15, the rectangular area 15
a is an area obtained by dividing the (r, θ) space. Rectangular area
15b is a region where the motion vector distribution is large. The rectangular area 15c is an area where the motion vector distribution is small.

FIG. 16 is a functional block diagram of a moving object detecting device according to the fourth embodiment of the present invention. 1 are given the same reference numerals. In FIG. 16, a moving image signal decoding processing unit 1a is means for decoding moving image signal encoded data and outputting a decoded image of an IP frame and a motion vector of a P frame. The inter-frame difference calculation unit 1b is a unit that calculates an inter-frame difference of an IP frame decoded image. The motion vector distribution creation unit 1c
This is a means for creating a distribution of motion vectors of an MB including a pixel having a change between frames. Motion vector number calculation unit
16a divides the motion vector distribution into a plurality of rectangular areas,
This is a means for calculating the number of motion vectors in a rectangular area. The intra-region motion vector selection unit 16b is a means for selecting a motion vector in a rectangular region having a number of vectors exceeding a predetermined threshold. The moving object detection unit 1e is a unit that detects an MB having the selected motion vector as a moving object area.

The operation of the moving object detecting apparatus according to the fourth embodiment of the present invention configured as described above will be described.
In the moving image signal decoding process, the inter-frame difference calculation process, and the motion vector distribution creation process, the same operation as in the first embodiment is performed to create a motion vector distribution. In the motion vector selection processing, the motion vector distribution is divided into a plurality of rectangular areas, the number of motion vectors distributed in each rectangular area is calculated, and the number of motion vectors in a rectangular area in which the motion vector exceeds a predetermined threshold is calculated. Select the motion vector to which it belongs.

FIG. 15 is a diagram showing an example in which the motion vector distribution is divided into a plurality of rectangular areas. Although this motion vector distribution is shown as a distribution on the magnitude / angle coordinate system, it may be a vector distribution created from vertical / horizontal components. In FIG. 15, the motion vector distribution B (r, θ) is divided into rectangular areas 15a, and the number of vectors distributed in each rectangular area is calculated. All the vectors in the rectangular area 15b whose vector number exceeds the predetermined threshold are selected, and all the vectors in the rectangular area 15c whose vector number does not exceed the predetermined threshold are deleted.

As a result, it is possible to easily select a portion where the motion vector distribution is concentrated, and it is possible to reduce the amount of calculation when selecting a vector. In the moving object detection processing based on the selected motion vector, a moving object is detected in the same manner as in the first embodiment.

As described above, in the fourth embodiment of the present invention, the moving object detection device divides the motion vector distribution into a plurality of rectangular areas and calculates the number of motion vectors distributed in each rectangular area. Since the calculation is performed and a motion vector belonging to a rectangular area where the number of motion vectors exceeds a predetermined threshold value is selected, the calculation of the distribution density can be easily performed, and the amount of calculation when selecting the motion vector can be reduced. .

(Fifth Embodiment) A fifth embodiment of the present invention is a moving object detecting apparatus which selects only a motion vector having a predetermined size or angle after selecting a motion vector. .

FIG. 17 is a flowchart showing a processing procedure of the moving object detection device according to the fifth embodiment of the present invention. FIG. 18 is a diagram showing the distribution of the selected motion vector. In FIG. 18, an area 18a has a vector size r.
And the area in which the range of the movement direction θ is limited. Point 18b is
This is a point representing the selected motion vector.

FIG. 19 is a functional block diagram of a moving object detection device according to the fifth embodiment. 1 are given the same reference numerals. In FIG. 19, a moving image signal decoding processing unit 1a is a means for decoding moving image signal encoded data and outputting a decoded image of an IP frame and a motion vector of a P frame. Inter-frame difference calculator 1b
Is means for calculating the inter-frame difference of the decoded IP frame image. The motion vector distribution creating unit 1c is a unit that creates a distribution of a motion vector of an MB including a changed pixel of an inter-frame difference. The motion vector selection unit 1d is means for selecting a motion vector based on the distribution density of the vector distribution. The predetermined motion vector selection unit 19a is means for further selecting a motion vector having a predetermined size and angle. The moving object detection unit 1e is a unit that detects an MB having the selected motion vector as a moving object area.

The operation of the moving object detecting apparatus according to the fifth embodiment of the present invention having the above-described configuration will be described.
In the video signal decoding process, the inter-frame difference calculation process, the motion vector distribution creation process, and the motion vector selection process, the first
The selected motion vector information is created in the same manner as in the embodiment. Only motion vectors having a predetermined magnitude or angle are further selected.

FIG. 18 is a diagram showing an example of a selected motion vector. The range of the vector size r of the object to be detected and the moving direction θ are limited, and only the selected motion vector 18b belonging to the range 18a is further selected. This allows
It is possible to extract only an object having a specific movement direction and movement amount (for example, an object entering from an entrance portion, a vehicle running at high speed, and the like). In the moving object detection processing based on the selected motion vector, a moving object is detected in the same manner as in the first embodiment.

As described above, in the fifth embodiment of the present invention, the moving object detection device is configured to select only a motion vector having a predetermined size or angle after selecting a motion vector. Therefore, it is possible to easily extract only the object moving in a predetermined direction or only the object moving by a predetermined moving amount from the moving objects in the image.

(Sixth Embodiment) In a sixth embodiment of the present invention, a first motion vector distribution is created from a frame vector and a top field vector, and a second motion vector distribution is created from a frame vector and a bottom field vector. , A motion vector is selected from each of the two motion vector distributions, and a first moving object detection result and a second moving object detection result are calculated from the two selected motion vectors to obtain two moving objects. This is a moving object detection device that integrates detection results and outputs a final detection result.

FIG. 20 is a flowchart showing a processing procedure of the moving object detection device according to the sixth embodiment of the present invention. FIG. 21 is a functional block diagram of the moving object detection device according to the sixth embodiment. 1 are given the same reference numerals. In FIG. 21, the video signal decoding processor 1a decodes the video signal encoded data to
This is a means for outputting a decoded image of a P frame and a motion vector of the P frame. First motion vector distribution creating unit 21a
The first motion vector selector 21c and the first moving object detector 21e are means for calculating a first moving object detection result from the frame vector and the top field vector.
The second motion vector distribution creating section 21b, the second motion vector selecting section 21d, and the second moving object detecting section 21f are means for calculating a second moving object detection result from the frame vector and the bottom field vector. Detection result integration means 21
g is a unit for calculating a final detection result.

The operation of the moving object detecting apparatus according to the sixth embodiment of the present invention configured as described above will be described.
In the video signal decoding process and the inter-frame difference calculation process, the inter-frame difference information is calculated in the same manner as in the first embodiment. Using a frame vector and a top field vector, in a first motion vector distribution creation process, a first motion vector selection process, and a first moving object detection process,
A first moving object detection result is calculated. In the first motion vector distribution creation processing, the first motion vector selection processing, and the first moving object detection processing, the same processing as in the first embodiment is performed.

Further, by using the frame vector and the bottom field vector, a second motion vector distribution creating process, a second motion vector selecting process, and a second moving object detecting process are performed. Is calculated. Second
In the motion vector distribution creation process, the second motion vector selection process, and the second moving object detection process,
The same processing as in the first embodiment is performed.

At this time, if the top field vector and the bottom field vector are the same vector, the first
And the second moving object detection result are the same. However, even with the same MB, the top field vector and the bottom field vector are not always the same. In particular, M in which luminance change has occurred due to the influence of disturbance
In B, the top field vector and the bottom field vector often differ greatly. On the other hand, in an actual moving object, the top field vector and the bottom field vector are almost the same vector.

Therefore, in the detection result integration processing, the first moving object detection result is compared with the second moving object detection result, and the MB detected by only one of the results is the result of disturbance detection. Judge that there is. Also, it is determined that the MBs detected in both are correct moving objects. That is, a logical product of the first moving object detection result and the second moving object detection result is calculated and output as a final result.

As described above, according to the sixth embodiment of the present invention, the moving object detecting apparatus generates the first motion vector distribution from the frame vector and the top field vector, and generates the first motion vector distribution from the frame vector and the bottom field vector. , A motion vector is selected from the two motion vector distributions, and a first moving object detection result and a second moving object detection result are calculated from the two selected motion vectors. Since one moving object detection result is integrated to output the final detection result, the moving object can be detected with high accuracy using both the top field and bottom field motion vectors.

(Seventh Embodiment) In a seventh embodiment of the present invention, the number of each vector is calculated for all the motion vectors in an image, and the number of vectors that exceeds a predetermined threshold is calculated. Is a moving object detection device that determines whether or not a moving object exists and does not detect a moving object when a large number of vectors exist.

FIG. 22 is a flowchart showing a processing procedure of the moving object detecting apparatus according to the seventh embodiment of the present invention. FIG. 23 is a functional block diagram of the moving object detection device according to the seventh embodiment. 1 are given the same reference numerals. In FIG. 23, the moving image signal decoding processing unit 1a decodes the moving image signal
This is a means for outputting a decoded image of a P frame and a motion vector of the P frame. The inter-frame difference calculation unit 1 b
This is a means for calculating the inter-frame difference of the frame decoded image. The majority vector existence determination unit 23a is a unit that calculates a majority vector. The stop signal sending section 23b is means for sending a process stop signal when a large number of vectors exist.
The motion vector distribution creating unit 1c is a unit that creates a distribution of motion vectors of an MB including changed pixels of an inter-frame difference when a large number of vectors do not exist. The motion vector selection unit 1d is means for selecting a motion vector based on the distribution density of the vector distribution. The moving object detection unit 1e
This is a means for detecting an MB having the selected motion vector as a moving object area.

The operation of the moving object detecting apparatus according to the seventh embodiment of the present invention configured as described above will be described.
In the video signal decoding process and the inter-frame difference calculation process, the inter-frame difference information is calculated in the same manner as in the first embodiment. In the majority vector calculation process, the number of each vector is calculated for all the motion vectors in the image, and the majority vector having a number equal to or larger than a predetermined threshold is calculated.

Since the horizontal and vertical components (Vx, Vy) are given to the motion vector in the P frame motion vector information, the vectors of all MBs in the image are checked, and the component is (Vx, Vy). Number of vectors Num (Vx, V
y) is calculated. After calculating Num (Vx, Vy) for all the vectors in the image, those whose values are larger than a predetermined threshold value Nth are extracted.

The threshold value Nth is set to a considerably large value. At this time, the existence of a vector satisfying Num (Vx, Vy)> Nth means that many MBs in an image have the same vector. In such a case, the movement of the camera itself has occurred. It is considered to be a motion vector.

Usually, the movement of the camera can be known by giving the operation information of the camera. However, when the operation device of the camera and the moving object detection device are located at a distance from each other, or when the moving image signal When a moving object is detected from the conversion data, camera operation information cannot be obtained. However, the movement of the camera can be known by calculating the number of vectors by the above method.

When the camera is moving, the moving object cannot be correctly detected, and the subsequent processing is stopped. On the other hand, when there is no large number of vectors, the detection of the moving object is performed in the motion vector distribution creation processing, the motion vector selection processing, and the moving object detection processing in the same manner as in the first embodiment.

As described above, in the seventh embodiment of the present invention, the moving object detecting apparatus calculates the number of each vector for all the motion vectors in an image, and It is determined whether or not there exists a large number of vectors, and when a large number of vectors exist, the detection of the moving object is not performed. Therefore, when the camera is moving, the detection of the moving object can be prevented. .

(Eighth Embodiment) In an eighth embodiment of the present invention, the number of each vector is calculated for all the motion vectors in an image, and the number of vectors that exceeds a predetermined threshold is calculated. Is a moving object detection device that determines whether or not a motion vector exists, and subtracts a vector from all motion vectors when a large number of vectors exist.

FIG. 24 is a flowchart showing a processing procedure of the moving object detection device according to the eighth embodiment of the present invention. FIG. 25 is a block diagram of functions of the moving object detection device according to the eighth embodiment. 1 and 23 are denoted by the same reference numerals. In FIG. 25, a moving image signal decoding unit 1a is a means for decoding moving image signal encoded data and outputting a decoded image of an IP frame and a motion vector of a P frame. Inter-frame difference calculator 1
b is means for calculating the inter-frame difference of the decoded IP frame image. The majority vector existence determination unit 23a is a unit that calculates a majority vector. Overall vector subtraction unit 25a
Is a means for subtracting the majority vector from all the vectors when the majority vector exists. The motion vector distribution creating unit 1c is a unit that creates a motion vector distribution of the MB including the changed pixel of the inter-frame difference based on the corrected motion vector. Motion vector selector 1d
Is means for selecting a motion vector according to the distribution density of the vector distribution. The moving object detection unit 1e is a unit that detects an MB having the selected motion vector as a moving object area.

The operation of the moving object detecting apparatus according to the eighth embodiment of the present invention configured as described above will be described.
In the video signal decoding process and the inter-frame difference calculation process, the inter-frame difference information is calculated in the same manner as in the first embodiment. In the majority vector calculation process, as in the seventh embodiment, for all the motion vectors in the image,
The number of each vector is calculated, and a large number of vectors (Vxg, V
yg) is determined.

When a large number of vectors (Vxg, Vyg) are present, the entire screen is moved by (Vxg, Vyg), so that the vector of each MB is equal to the original motion vector (Vx, Vy) and the entire vector. (Vxg, Vyg). That is, by subtracting a large number of vectors from the motion vector of each MB, an original vector can be obtained, and using this, a moving object can be detected even when the camera is moving.

Therefore, the vector subtraction process subtracts a large number of vectors from all the motion vectors, and the motion vector distribution creation process creates a motion vector distribution using these motion vectors. After that, the motion vector selection processing and the moving object detection processing perform the same operations as in the first embodiment,
A moving object is detected.

As described above, in the eighth embodiment of the present invention, the moving object detecting apparatus calculates the number of each vector for all motion vectors in an image, It is determined whether or not there is a large number of vectors, and when the large number of vectors exists, the vectors are subtracted from all the motion vectors. Therefore, a moving object can be detected with high accuracy even in scenes of pan and camera shake.

(Ninth Embodiment) A ninth embodiment of the present invention is an automatic alarm generation device that detects a moving object and generates an alarm signal based on the detection result information.

FIG. 26 is a functional block diagram of the automatic alarm generator according to the ninth embodiment of the present invention. In FIG. 26, a moving object detection unit 26a is means for detecting a moving object and outputting detection result information by the same method as in the first to eighth embodiments. The alarm signal generation unit 8b is a unit that generates and outputs an alarm signal when a moving object exists in the image.

The operation of the automatic alarm generator configured as described above will be described. The moving object detection unit 26a includes
Similarly to the method described in the eighth embodiment, a moving object is detected from encoded video signal data, and moving object detection result information is output. The alarm signal generator 8b generates and outputs an alarm signal when a moving object exists in the image. Note that the alarm signal generation unit 8b may set a monitoring area frame in an image and generate an alarm signal when a moving object exists in the monitoring area frame.

As described above, in the ninth embodiment, the automatic alarm generator detects a moving object and generates an alarm signal based on the detection result information. Can accurately detect a moving object at and automatically generate an alarm.

(Tenth Embodiment) A tenth embodiment of the present invention is directed to an automatic moving picture which encodes a moving picture signal from a camera and records moving picture signal encoded data when an alarm is generated. It is a signal encoded data recording device.

FIG. 27 is a functional block diagram of an automatic moving picture signal encoded data recording apparatus according to the tenth embodiment of the present invention. In FIG. 27, an imaging device 27a is a video camera for monitoring. Video signal encoding device 27b
Is means for encoding a moving image signal. The automatic alarm generating device 27c is means for detecting a moving object and generating an alarm. The moving image signal encoded data recording device 27
This is means for recording moving image signal encoded data when an alarm signal is generated.

The operation of the automatic moving picture signal coded data recording system configured as described above will be described. Imaging device 27
The moving image signal picked up by a
b. The automatic alarm generator 27c has a ninth
The same operation as in the embodiment is performed to generate an alarm signal. The moving image signal encoded data recording device 27d records the encoded moving image signal data when an alarm signal is generated.

As described above, in the tenth embodiment,
Since the automatic moving picture signal encoded data recording device is configured to encode the moving picture signal from the camera and record the moving picture signal encoded data when an alarm is generated, a moving object in the moving picture signal encoded data is recorded. With high accuracy,
Only when there is a movement, moving picture signal encoded data can be automatically recorded.

(Eleventh Embodiment) In an eleventh embodiment of the present invention, a moving image signal from a camera is encoded, a moving object is detected from the encoded moving image signal, an alarm is generated, and a moving image is generated. An automatic moving image recording system that decodes encoded image signal data and records a decoded image signal when an alarm is generated.

FIG. 28 is a functional block diagram of the automatic moving picture recording system according to the eleventh embodiment of the present invention. The same components as those in FIG. 27 are denoted by the same reference numerals.
27 is different from FIG. 27 in that a moving image signal decoding device 28a is provided and an image signal recording device 28b is provided instead of the moving image signal encoded data recording device 27d. In FIG. 28, a moving image signal decoding device 28a is means for decoding an encoded moving image signal. The image signal recording device 28b is means for recording the decoded moving image signal.

The operation of the automatic moving picture signal recording system configured as described above will be described. Automatic alarm generator 27c
Detects the moving object from the moving image signal encoded data and generates an alarm signal in the same operation as in the ninth embodiment. The moving image signal recording device 28b records the moving image signal decoded by the moving image signal decoding unit 28a when an alarm signal is generated.

As described above, in the eleventh embodiment of the present invention, the automatic video recording system encodes a video signal from a camera, detects a moving object from the coded video signal, and issues an alarm. The moving image signal encoded data is generated, and the decoded image signal is recorded when an alarm is generated. Therefore, when a moving object in the moving image signal encoded data is accurately detected and there is a movement. Only the moving image signal can be automatically recorded.

(Twelfth Embodiment) In a twelfth embodiment of the present invention, a moving object is detected by inputting information necessary for moving object detection from a moving picture signal decoding apparatus, and an alarm is generated. This is an automatic moving image recording system that records a decoded image signal when it is performed.

FIG. 29 is a functional block diagram of the automatic moving picture recording system according to the twelfth embodiment of the present invention. 27 and 28 are denoted by the same reference numerals. The difference from FIG. 28 is that the moving picture signal decoding apparatus
The processing contents of 29a and the processing contents of the automatic alarm generator 29b are shown.

Referring to FIG. 29, only the part of the operation of the automatic moving picture signal recording system configured as described above which is different from the eleventh embodiment will be described. The automatic alarm generating device 29b inputs information necessary for detecting a moving object from the moving image signal decoding unit 29a, and based on the information, detects the moving object from the moving image signal encoded data in the same operation as in the ninth embodiment. Detect and generate alarm signal.

By the way, the operation of the automatic alarm generating apparatus according to the ninth embodiment requires a video signal decoding process as shown in FIG. However, the video signal decoding process is also performed by the video signal decoding device 29a. Therefore, the automatic alarm generating device 29b inputs information necessary for detecting a moving object, that is, a decoded image of an IP frame and a motion vector of a P frame from the moving image signal decoding device 29a. As a result, the video signal decoding process in the automatic alarm generator 29b can be omitted.

The moving picture signal decoding apparatus 29a normally outputs the decoded picture of the IP frame and the motion vector of the P frame, and only when the automatic alarm generating apparatus 29b outputs the alarm signal, the entire IPB frame is output. May be decoded.

As described above, in the twelfth embodiment,
Since the automatic moving image recording system receives information necessary for moving object detection from a moving image signal decoding device, detects a moving object, and records a decoded image signal when an alarm is generated, It is possible to accurately detect a moving object in moving image signal encoded data with a small amount of processing, and to automatically record a moving image signal only when there is a movement.

(Thirteenth Embodiment) In a thirteenth embodiment of the present invention, a moving picture signal from a camera is encoded, and the moving picture signal encoded data is remotely transmitted via a communication path and transmitted. Receiving the encoded moving image signal data, decoding the encoded moving image signal data, displaying the decoded moving image signal,
This is an automatic alarm system that detects a moving object, generates an alarm, and sounds an alarm when the alarm is generated.

FIG. 30 is a functional block diagram of the automatic alarm system according to the thirteenth embodiment of the present invention. FIG.
7, the same functional portions as those in FIG. 28 are denoted by the same reference numerals.
28 is different from the transmitting device 30a and the receiving device 30 in FIG.
b, a display device 30c, and an alarm generator 30d, and no image signal recording device 28b. In FIG. 30, a transmitting device 30a is means for transmitting moving image signal encoded data by wire or wirelessly. The receiving device 30b
This is means for receiving moving image signal encoded data by wire or wirelessly. The display device 30c is a unit that displays a moving image signal. The alarm generator 30d is a unit that issues an alarm in response to an alarm signal.

Referring to FIG. 30, only the part of the operation of the automatic alarm system configured as described above that is different from the eleventh embodiment will be described. The moving image signal encoded data encoded by the moving image signal encoding device 27b is remotely transmitted from the transmission device 30a via a wired or wireless communication path. Moving image signal encoded data transmitted remotely is
The signal is received by the receiving device 30b and sent to the video signal decoding device 28a and the automatic alarm generating device 27c.

The automatic alarm generation device 27c detects a moving object from encoded video signal data and generates an alarm signal in the same operation as in the ninth embodiment. Alarm device 30
d sounds an alarm when an alarm signal is generated to notify the observer that there is a moving object. In addition,
The alarm device may be a means for automatically calling the police or a security company, or may be a device that notifies by light or vibration.

[0110] The display device 30c includes a moving image signal decoding device 28.
The video signal decoded in a is displayed. Note that, instead of the video signal decoding device 28a and the automatic alarm generation device 27c, a video signal decoding unit 29a and an automatic alarm generation unit 29b in an automatic video signal recording system may be used. Also, by providing the same transmission / reception unit for each automatic recording system, an automatic recording system for remote monitoring or the like can be easily configured.

As described above, in the thirteenth embodiment,
The automatic alarm system encodes the moving image signal from the camera, remotely transmits the moving image signal encoded data via a communication path, receives the transmitted moving image signal encoded data, and transmits the moving image signal encoded data. And display the decoded moving image signal, detect the moving object, generate an alarm, and sound an alarm when the alarm is generated, so that the moving object in the moving image signal encoded data can be accurately detected. Detect
When there is a movement in the monitored area, the monitoring person can be automatically notified by an alarm.

(Fourteenth Embodiment) In a fourteenth embodiment of the present invention, moving picture signals from a plurality of cameras are encoded and multiplexed, data is remotely transmitted, and received multiplexed data is transmitted to a plurality of cameras. Decode into moving image signal encoded data and decode,
An automatic video switching system that selects and displays a moving video signal based on a plurality of alarm signals generated by detecting a moving object from video data encoded data.

FIG. 31 is a functional block diagram of the automatic video switching system according to the fourteenth embodiment of the present invention.
In FIG. 31, a plurality of imaging devices 31a are a plurality of video cameras provided at a monitored location. The plurality of video signal encoding devices 31b are means for encoding a video signal.
The multiplexing device 31c is means for multiplexing a plurality of encoded video signal data. The transmission device 31d is means for transmitting the multiplexed data by wire or wirelessly. Receiver 31e
Is means for receiving multiplexed data by wire or wirelessly. The separating device 31f is a unit that separates the multiplexed data into a plurality of encoded moving image signal data. The plurality of moving image signal decoding devices 31g are means for decoding moving image signal encoded data. The plurality of automatic alarm generating devices 31h are means for detecting a moving object from moving image signal encoded data and generating an alarm. The moving image signal switching device 31i is means for selecting and outputting one moving image signal based on a plurality of alarm signals. The display device 30c is a unit that displays the selected moving image signal.

The operation of the automatic video switching system according to the fourteenth embodiment configured as described above will be described. Moving image signals captured by the plurality of imaging devices 31a at the monitored location are encoded by the plurality of moving image signal encoding devices 31b, respectively. The plurality of encoded video signal data are
The data is multiplexed by the multiplexing device 31c, and is remotely transmitted from the transmitting device 31d via a wired or wireless communication path.

The multiplexed data transmitted remotely is received by the receiving device 31e and sent to the separating device 31f. The separating device 31f separates the multiplexed data into a plurality of encoded moving image signal data, and sends them to the image signal decoding device 31g and the automatic alarm generating device 31h, respectively. The plurality of automatic alarm generators 31h detect the moving object from the moving image signal encoded data and transmit an alarm signal to the moving image signal switching device 31i in the same operation as in the ninth embodiment. The plurality of moving image signal decoding devices 31g decode the transmitted moving image signal encoded data, and convert the moving image signals to a moving image signal switching device.
Send to 31i.

The moving image signal switching device 31i selects and outputs one moving image signal based on a plurality of alarm signals. As an example of a selection method, when there is one alarm, a moving image signal in which an alarm has occurred is selected, and when there are a plurality of alarms, a priority order is determined for an input channel.
There is a method of selecting a moving image signal having a higher priority. The moving image signal selected by the moving image signal switching device 31i is displayed on the display device 30c.

The functions of the automatic alarm generating device 31h and the moving image signal decoding device 31g correspond to the moving image signal decoding device 28a and the automatic alarm generating device 27c in the automatic moving image signal recording system described in the eleventh embodiment. Or the same as the automatic alarm generating unit 29b and the moving image signal decoding unit 29a in the automatic moving image signal recording system described in the twelfth embodiment. Further, the output destination of the moving image signal switching unit 31i is the display device 30c, but may be a moving image signal recording device.

As described above, in the fourteenth embodiment,
The automatic video switching system is used to remotely transmit multiplexed data obtained by encoding moving image signals from a plurality of cameras, and to separate and decode the received multiplexed data into a plurality of encoded moving image signal data to decode the moving image signal. A moving image signal having a motion is selected and displayed based on a plurality of alarm signals generated by detecting a moving object from the encoded data, so that the moving object in the encoded video signal data can be accurately detected. ,
It is possible to select an image of a moving place from images of a plurality of monitored places and automatically switch the image.

(Fifteenth Embodiment) A fifteenth embodiment of the present invention multiplexes a plurality of moving picture signal coded data obtained by coding moving picture signals from a plurality of video cameras and remotely transmits the multiplexed data. Separating the received multiplexed data into a plurality of encoded moving image signal data, detecting a moving object from the encoded moving image signal data, generating an alarm, and generating a moving image signal based on the plurality of alarm signals. This is an automatic video switching system that selects, decodes, and displays encoded data.

FIG. 32 is a functional block diagram of the automatic video switching system according to the fifteenth embodiment of the present invention.
The same components as those in FIG. 31 are denoted by the same reference numerals. FIG.
The difference from 1 is that a moving picture signal coded data switching apparatus 32a and a moving picture signal decoding apparatus 32b are provided instead of the plurality of moving picture signal decoding apparatuses 31g and the moving picture signal switching apparatuses 31i. In FIG. 32, the moving picture signal coded data switching device 32a is means for selecting and outputting one moving picture signal coded data based on a plurality of alarm signals. The moving picture signal decoding device 32b is means for decoding the selected moving picture signal encoded data.

Referring to FIG. 32, only the operation of the automatic video switching apparatus according to the fifteenth embodiment configured as described above, which is different from the fourteenth embodiment, will be described. The plurality of automatic alarm generators 31h detect the moving object from the moving image signal encoded data and transmit an alarm signal to the moving image signal encoded data switching device 32a in the same operation as in the ninth embodiment.

The moving picture signal coded data switching device 32a
One of the moving image signal encoded data is selected and output based on the plurality of alarm signals. As an example of the selection method,
If the number of alarms is one, the moving picture signal coded data in which the alarm has occurred is selected. There is a way to select data.
The selected encoded video signal data is decoded by the video signal decoding device 32b and displayed on the display device 30c.

The moving image signal coded data switching device 32
The output destination of “a” may be a moving image signal encoded data recording device. Further, the output destination of the video signal decoding device 32b may be a video signal recording device.

As described above, in the fifteenth embodiment,
The automatic video switching system multiplexes a plurality of moving image signal encoded data obtained by encoding moving image signals from a plurality of video cameras, remotely transmits the data, and converts the received multiplexed data into a plurality of moving image signal encoded data. Separated, a moving object is detected from the moving image signal encoded data, an alarm is generated, and moving image signal encoded data having motion is selected, decoded, and displayed based on a plurality of alarm signals. It is possible to select and automatically switch the moving image signal encoded data of a moving place from a plurality of monitored place images and decode and display the moving image signal with one moving image signal decoding unit.

(Sixteenth Embodiment) A sixteenth embodiment of the present invention encodes moving image signals from a plurality of video cameras, detects a moving object from the moving image signal encoded data, and generates an alarm. Selecting and remotely transmitting moving image signal encoded data based on a plurality of alarm signals,
This is an automatic video switching system that decodes and displays received moving image signal encoded data.

FIG. 33 is a functional block diagram of the automatic video switching system according to the sixteenth embodiment of the present invention.
In FIG. 33, a plurality of imaging devices 31a are a plurality of video cameras provided at a monitored location. The plurality of video signal encoding devices 31b are means for encoding a video signal.
The plurality of automatic alarm generating devices 31h are means for detecting a moving object from moving image signal encoded data and generating an alarm. The moving picture signal encoded data switching device 32a converts the moving picture signal encoded data to one based on a plurality of alarm signals.
Means for selecting and outputting one. The transmission device 31d is means for transmitting the moving image signal encoded data by wire or wirelessly. The receiving device 31e is a unit that receives moving image signal encoded data by wire or wirelessly. Video signal decoding device
32b is means for decoding moving picture signal encoded data. The display device 30c is a unit that displays a moving image signal.

The operation of the automatic video switching device according to the sixteenth embodiment configured as described above will be described. Moving image signals captured by the plurality of imaging devices 31a at the monitored location are encoded by the plurality of moving image signal encoding devices 31b, respectively. The plurality of automatic alarm generators 31h detect a moving object from the moving image signal encoded data and transmit an alarm signal to the moving image signal encoded data switching device 32a in the same operation as in the ninth embodiment.

The moving picture signal coded data switching device 32a
One of the moving image signal encoded data is selected and output based on the plurality of alarm signals. As an example of the selection method,
If the number of alarms is one, the moving image signal coded data in which the alarm has occurred is selected. If the number of alarms is plural, the priority order is determined for the input channel, and the moving image signal code having the higher priority is determined. There is a method of selecting the coded data.

The selected encoded moving picture signal data is remotely transmitted from the transmitting device 31d via a wired or wireless communication path. The coded video signal transmitted remotely is received by the receiving device 31e, and the video signal decoding device 32b
And then displayed on the display device 32c.

The output destination of the receiving device 31e may be a moving image signal encoded data recording device. Further, the output destination of the video signal decoding device 32b may be a video signal recording device.

As described above, in the fifteenth embodiment,
The automatic video switching system encodes moving image signals from a plurality of video cameras, detects a moving object from the encoded moving image signal data, generates an alarm, and generates a moving image signal based on a plurality of alarm signals. Since the coded data is selected and transmitted remotely, and the received coded video signal data is decoded and displayed, a moving object in the coded video signal data can be accurately detected, and monitoring of a plurality of locations can be performed. It is possible to select and automatically switch the moving image signal coded data of a moving place from the video of the place and transmit it.

(Seventeenth Embodiment) In a seventeenth embodiment of the present invention, a moving object is detected by coding moving image signals from a plurality of video cameras and inputting information necessary for detecting the moving object. Generate an alarm, select moving image signal encoded data having motion based on the plurality of alarm signals,
This is an automatic video switching system that remotely transmits and decodes and receives encoded moving image signal encoded data.

FIG. 34 is a functional block diagram of the automatic video switching system according to the seventeenth embodiment of the present invention.
In FIG. 34, a plurality of imaging devices 31a are a plurality of video cameras provided at a monitored location. The plurality of moving picture signal coding devices 34a are means for coding moving picture signals and outputting information necessary for moving object detection. The plurality of automatic alarm generators 34b are means for detecting a moving object and generating an alarm. The moving picture signal coded data switching device 32a is means for selecting and outputting one moving picture signal coded data based on a plurality of alarm signals. The transmission device 31d is means for transmitting the moving image signal encoded data by wire or wirelessly. The receiving device 31e is a unit that receives moving image signal encoded data by wire or wirelessly.
The moving picture signal decoding device 32b is means for decoding moving picture signal encoded data. The display device 30c is a unit that displays a moving image signal.

Of the operations of the automatic video switching system according to the seventeenth embodiment configured as described above,
Only the parts different from the third embodiment will be described with reference to FIG. The plurality of automatic alarm generators 34b each input information necessary for detecting a moving object from the moving image signal encoding unit 34a, and perform the same operation as in the ninth embodiment to extract the moving object from the moving image signal encoded data. Is detected, and an alarm signal is sent to the moving picture signal coded data switching device 32a.

The moving object detection requires a moving image signal decoding process as shown in FIG. Since the information obtained in the moving image signal decoding process is information existing in the moving image signal encoding device 34a, the automatic alarm generating device 34b inputs information necessary for moving object detection from the moving image signal encoding device 34a. By doing so, the video signal decoding processing can be omitted.

As described above, in the seventeenth embodiment,
The automatic video switching system encodes moving image signals from multiple video cameras, inputs information necessary for moving object detection, detects a moving object, generates an alarm, and detects movement based on multiple alarm signals. Since a certain moving picture signal coded data is selected, remotely transmitted, and the received moving picture signal coded data is decoded and displayed, a moving object in the moving picture signal coded data can be accurately detected with a small amount of processing. It is possible to detect and select moving image signal encoded data of a moving place from video images of a plurality of monitored places, automatically switch the data, and transmit the data.

[0137]

As is apparent from the above description, the present invention provides a moving object detection method which decodes a moving image signal of an intra-frame coded image and a forward prediction inter-frame coded image and performs forward prediction. A video signal decoding processing step of outputting a motion vector of an inter-frame coded image, an inter-frame difference calculation processing step of calculating an inter-frame difference from decoded image data, and a macroblock including a changed pixel whose inter-frame difference has been calculated A motion vector distribution creation processing step of creating a motion vector distribution having a motion vector distribution density, and a motion vector selection processing step of selecting a motion vector whose motion vector distribution density satisfies a predetermined condition to obtain a selected motion vector; Detecting a macroblock including the calculated change pixel and having a selected motion vector as a moving object area Since a structure that includes a moving object detection processing step, by eliminating the influence of the disturbance, there is an advantage that it is possible to accurately detect the moving object from the moving image signal encoding data.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of a moving object detection device according to a first embodiment of the present invention;

FIG. 2 is a processing flowchart of the moving object detection device according to the first embodiment of the present invention;

FIG. 3 is a frame configuration diagram of moving image signal encoded data composed of IPB frames;

FIG. 4 is a frame configuration diagram of moving image signal encoded data composed of IP frames;

FIG. 5 is a functional block diagram of a moving image signal decoding unit according to the first embodiment of the present invention;

FIG. 6 is an explanatory diagram of an inter-frame difference obtained by the moving object detection device according to the first embodiment of the present invention;

FIG. 7 is an explanatory diagram of a motion vector obtained by the moving object detection device according to the first embodiment of the present invention;

FIG. 8 is a diagram showing a motion vector distribution obtained by the moving object detection device according to the first embodiment of the present invention;

FIG. 9 is a processing flowchart of a moving object detection device according to a second embodiment of the present invention;

FIG. 10 is an explanatory diagram of a field vector obtained by the moving object detection device according to the second embodiment of the present invention;

FIG. 11 is a functional block diagram of a moving object detection device according to a second embodiment of the present invention;

FIG. 12 is a processing flowchart of a moving object detection device according to a third embodiment of the present invention;

FIG. 13 is a functional block diagram of a moving object detection device according to a third embodiment of the present invention;

FIG. 14 is a processing flowchart of a moving object detection device according to a fourth embodiment of the present invention;

FIG. 15 is a diagram showing a motion vector distribution obtained by the moving object detection device according to the fourth embodiment of the present invention;

FIG. 16 is a functional block diagram of a moving object detection device according to a fourth embodiment of the present invention;

FIG. 17 is a processing flowchart of a moving object detection device according to a fifth embodiment of the present invention;

FIG. 18 is an explanatory diagram for selecting a specific motion vector obtained by a moving object detection device according to a fifth embodiment of the present invention;

FIG. 19 is a functional block diagram of a moving object detection device according to a fifth embodiment of the present invention;

FIG. 20 is a processing flowchart of the moving object detection device according to the sixth embodiment of the present invention;

FIG. 21 is a functional block diagram of a moving object detection device according to a sixth embodiment of the present invention;

FIG. 22 is a processing flowchart of a moving object detection device according to a seventh embodiment of the present invention;

FIG. 23 is a functional block diagram of a moving object detection device according to a seventh embodiment of the present invention;

FIG. 24 is a processing flowchart of the moving object detection device according to the eighth embodiment of the present invention;

FIG. 25 is a functional block diagram of a moving object detection device according to an eighth embodiment of the present invention;

FIG. 26 is a functional block diagram of an automatic alarm generation device according to a ninth embodiment of the present invention;

FIG. 27 is a functional block diagram of an automatic moving picture signal encoded data recording system according to a tenth embodiment of the present invention;

FIG. 28 is a functional block diagram of an automatic moving image signal recording system according to an eleventh embodiment of the present invention;

FIG. 29 is a functional block diagram of an automatic moving image signal recording system according to a twelfth embodiment of the present invention;

FIG. 30 is a functional block diagram of an automatic alarm system according to a thirteenth embodiment of the present invention;

FIG. 31 is a functional block diagram of an automatic video switching system according to a fourteenth embodiment of the present invention;

FIG. 32 is a functional block diagram of an automatic video switching system according to a fifteenth embodiment of the present invention;

FIG. 33 is a functional block diagram of an automatic video switching system according to a sixteenth embodiment of the present invention;

FIG. 34 is a functional block diagram of an automatic video switching system according to a seventeenth embodiment of the present invention;

FIG. 35 is a functional block diagram of a conventional moving object detection device.

[Explanation of symbols]

1a Moving image signal decoding processing unit 1b Inter-frame difference calculating unit 1c Motion vector distribution creating unit 1d Motion vector selecting unit 1e Moving object detecting unit 3a Intra-frame coded image (I frame) 3b Forward predicted inter-frame coded image (P 3c Bidirectionally predicted inter-frame coded image (B frame) 5a Header decoding process 5b Variable length decoding process 5c Dequantization 5d IDCT 5e Frame memory 5f Addition process 6a Reference frame of frame for detecting moving object 6b Moving object 6c Inter-frame difference result 6d 6a 6a person position 6e 6b person position 6f, 6g, 6i 6c set of pixels regarded as a moving object at 6c Disturbance position at 6h 6b 7a Motion vector 7b 6f pixel Example of MB containing 7c Example of MB including 7d Example of MB including 6i pixels 8a Unselected motion vector distribution 8b Selected motion vector distribution 10a Reference frame image 10b Processing frame image 10c Reference top field image 10d Reference bottom field image 10e Processing field image 10f frame Vector 10g Top field vector 10h Bottom field vector 10i Processed field image 10j Top field vector 10i to processed field image 11a Motion vector normalizer 13a Motion vector polar coordinate calculator 15a Rectangular area 15b divided vector distribution 15b Number of vectors Rectangular area exceeding the value 15c Rectangular area where the number of vectors is less than or equal to the threshold value 16a Motion vector number calculating unit 16b Intra-region motion vector selecting unit 18a Range indicating predetermined moving direction 18b Motion vector in predetermined moving direction 19a Predetermined motion vector selector 21a First motion vector distribution generator 21b Second motion vector distribution generator 21c First motion vector selector 21d Second motion vector selector 21e First moving object detector 21f 2 Moving object detecting unit 21g Detection result integrating unit 23a Multiple vector existence determining unit 23b Stop signal sending unit 25a Total vector subtracting unit 26a Moving object detecting unit 26b Alarm signal generating unit 27a Imaging device 27b Moving image signal encoding device 27c Automatic alarm Generating device 27d Video signal encoded data recording device 28a Video signal decoding device 28b Video signal recording device 29a Video signal decoding device 29b Automatic alarm generating device 30a Transmitting device 30b Receiving device 30c Display device 30d Alarm device 31a Multiple imaging Apparatus 31b Multiple video signal encoding apparatuses 31c Multiplexing apparatus 31d Transmitting apparatus 31e Receiving apparatus 31f Separating apparatus 31g Plural video signal decoding devices 31h Plural automatic alarm generation devices 31i Video signal switching device 32a Video signal encoding data switching device 32b Video signal decoding device 34a Plural video signal encoding devices 34b Plural automatic alarm generation Device 35a Camera 35b Image input unit 35c Change area detection unit 35d Area blocking unit 35e Feature extraction unit 35f Moving object identification unit 35g Result output unit 35h Display device

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Masato Nishizawa 4-3-1 Tsunashima Higashi, Kohoku-ku, Yokohama-shi, Kanagawa F-term (reference) 5C059 MA00 MA05 MA23 NN01 NN21 NN28 PP05 PP06 RB01 SS11 UA05

Claims (25)

[Claims] A moving image signal decoding step of decoding a moving image signal of an intra-coded image and a forward prediction inter-coded image, and outputting a motion vector of the forward predicted inter-coded image; An inter-frame difference calculation processing step of calculating an inter-frame difference from the decoded image data; and a motion vector distribution creation processing step of creating a distribution of a motion vector of a macroblock including the changed pixel in which the inter-frame difference is calculated, A motion vector selection processing step of selecting a motion vector whose distribution density satisfies a predetermined condition to obtain a selected motion vector; and a macroblock including the changed pixel whose inter-frame difference has been calculated and having the selected motion vector. Moving object detection processing step of detecting a moving object region as a moving object region. Moving object detection method. 2. The movement according to claim 1, wherein when the motion vector is a field vector, the distribution of the motion vector is created after normalizing the magnitude of the motion vector by an inter-field distance. Object detection method. 3. The distribution of the motion vector according to claim 1, wherein the magnitude and the angle of the motion vector are calculated, and the distribution of the motion vector is created on a two-dimensional coordinate system using the magnitude and the angle as coordinates. Moving object detection method. 4. The method according to claim 1, wherein the motion vector distribution area is divided into a plurality of rectangular areas, the number of motion vectors distributed in each rectangular area is calculated, and the number of the motion vectors exceeds a predetermined threshold. And selecting a motion vector belonging to the set as the selected motion vector.
4. The moving object detection method according to any one of 3. 5. The method according to claim 1, wherein after selecting the motion vector, only a motion vector having a predetermined size or angle is further selected as the selected motion vector. Moving object detection method. 6. A motion vector distribution creating step of creating a first motion vector distribution from a frame vector and a top field vector, and creating a second motion vector distribution from the frame vector and a bottom field vector. A motion vector selection processing step of selecting a selected motion vector from each of the two motion vector distributions; and calculating a first moving object detection result and a second moving object detection result from the two selected motion vectors, respectively. The moving object detection method according to any one of claims 1 to 5, further comprising: a moving object detection processing step of integrating one moving object detection result and outputting a final detection result. 7. The number of each motion vector is calculated for all motion vectors in an image, and it is determined whether or not there is a large number of vectors having a number equal to or more than a predetermined threshold. The moving object detection method according to claim 1, further comprising a multiple vector calculation processing step that does not detect a moving object. 8. The number of each motion vector is calculated for all motion vectors in an image, and it is determined whether or not there is a large number of vectors having a number equal to or more than a predetermined threshold value. 7. The moving object detection method according to claim 1, further comprising a majority vector calculation processing step of subtracting the majority vector from all motion vectors. 9. A video signal for decoding encoded video signal data and outputting a decoded image of an intra-frame encoded image, a forward predicted inter-frame encoded image, and a motion vector of a forward predicted inter-frame encoded image. A decoding processing unit, an inter-frame difference calculating unit that calculates an inter-frame difference of the decoded image, a motion vector distribution creating unit that creates a distribution of a motion vector of a macroblock including a changed pixel of the inter-frame difference, A motion vector selection unit that selects a motion vector whose distribution density satisfies a predetermined condition to obtain a selected motion vector, and a macroblock including the changed pixel of the inter-frame difference and having the selected motion vector as a moving object area A moving object detection device, comprising: a moving object detection unit for detecting. 10. The moving object detecting apparatus according to claim 9, further comprising a motion vector normalizing unit that normalizes the magnitude of the motion vector according to an inter-field distance when the motion vector is a field vector. 11. The moving object detection device according to claim 9, further comprising a motion vector polar coordinate calculation unit that calculates a magnitude and an angle of the motion vector. 12. A motion vector number calculation unit that divides the motion vector distribution area into a plurality of rectangular areas and calculates the number of motion vectors distributed in each rectangular area, and the number of motion vectors in the rectangular area. 12. An intra-region motion vector selection unit which selects a motion vector belonging to a rectangular region exceeding a predetermined threshold value and sets the selected motion vector as the selected motion vector. Moving object detection device. 13. The moving object detection device according to claim 9, further comprising a predetermined motion vector selection unit that selects only a motion vector having a predetermined size or angle. 14. A first motion vector distribution creating unit for creating a first motion vector distribution from a frame vector and a top field vector, and a first motion vector distribution selecting unit for selecting a first selected motion vector from the first motion vector distribution. A first motion vector selecting unit, and a first motion vector
A first moving object detection unit for calculating a moving object detection result of the second motion vector distribution unit, a second motion vector distribution creating unit for creating a second motion vector distribution from the frame vector and the bottom field vector, and the second motion vector A second motion vector selecting unit that selects a second selected motion vector from the distribution, a second moving object detecting unit that calculates a second moving object detection result from the second selected motion vector, The moving object according to any one of claims 9 to 13, further comprising: a detection result integrating unit that integrates the moving object detection result of the second moving object and the second moving object detection result and outputs a final detection result. Object detection device. 15. A multiple vector existence determining unit that calculates the number of motion vectors for all motion vectors in an image and determines whether there is a multiple vector having a number equal to or greater than a predetermined threshold value, The moving object detection device according to any one of claims 9 to 14, further comprising a stop signal sending unit that sends a processing stop signal to the motion vector distribution creating unit when a vector exists. 16. A multi-vector existence determining unit that calculates the number of motion vectors for all motion vectors in an image and determines whether or not there is a multi-vector having a number equal to or more than a predetermined threshold value; The moving object detection device according to claim 9, further comprising a majority vector subtraction unit configured to subtract the majority vector from a motion vector. 17. A moving object detecting apparatus according to claim 9, further comprising: an alarm signal generating unit configured to generate an alarm signal based on a moving object detection result of the moving object detecting apparatus. And an automatic alarm generator. 18. An imaging device, a moving image signal encoding device that encodes a moving image signal from the imaging device, and
7. An automatic moving picture signal coded data recording system comprising: the automatic alarm generating apparatus according to claim 7; and a moving picture signal coded data recording apparatus that records moving picture signal coded data when an alarm is generated. 19. An imaging device, a video signal encoding device that encodes a video signal from the imaging device, and
7. An automatic moving picture, comprising: the automatic alarm generating apparatus according to claim 7, a moving picture signal decoding apparatus for decoding moving picture signal encoded data, and a moving picture signal recording apparatus for recording a decoded picture signal in response to an alarm. Image signal recording system. 20. An imaging apparatus, a moving image signal encoding apparatus that encodes a moving image signal from the imaging apparatus, and a moving image that decodes encoded moving image signal data and outputs information necessary for detecting a moving object. 18. The automatic alarm generating device according to claim 17, wherein a moving object is detected by inputting information from the image signal decoding device, the moving image signal decoding device, and a moving image signal recording device that records a decoded image signal in response to an alarm. And an automatic moving picture signal recording system. 21. An imaging device, a video signal encoding device that encodes a video signal from the imaging device, a transmission device that remotely transmits encoded video signal data via a communication path, and the communication device. A receiving device that receives the encoded video signal data transmitted via a channel, a video signal decoding device that decodes the encoded video signal data, a display device that displays the decoded video signal, Item 18. An automatic alarm system comprising: the automatic alarm generation device according to Item 17; and an alarm device that sounds an alarm sound in response to the alarm signal. 22. A plurality of imaging devices; a plurality of video signal encoding devices for encoding video signals from the imaging devices; a multiplexing device for multiplexing a plurality of encoded video signal data; A transmitting apparatus for remotely transmitting multiplexed data via a communication path, a receiving apparatus for receiving multiplexed data transmitted via the communication path, and separating the multiplexed data into a plurality of encoded moving image signal data A separating device, a plurality of video signal decoding devices for decoding the video signal encoded data, a plurality of automatic alarm generation devices according to claim 17,
An automatic video switching system, comprising: a video signal switching device that selects a moving video signal based on a plurality of alarm signals; and a display device that displays the selected video signal. 23. A plurality of imaging devices, a plurality of video signal encoding devices for encoding video signals from the imaging devices, a multiplexing device for multiplexing a plurality of encoded video signal data, A transmitting apparatus for remotely transmitting multiplexed data via a communication path, a receiving apparatus for receiving multiplexed data transmitted via the communication path, and separating the multiplexed data into a plurality of encoded moving image signal data A plurality of automatic alarm generating devices according to claim 17, a moving image signal coded data switching device for selecting moving image signal coded data having motion based on a plurality of alarm signals, An automatic video switching system comprising: a video signal decoding device for decoding coded video signal data; and a display device for displaying a video signal. 24. A plurality of imaging devices, a plurality of video signal encoding devices for encoding video signals from the imaging devices, a plurality of automatic alarm generation devices according to claim 17,
A moving image signal coded data switching device that selects moving image signal coded data having motion based on a plurality of alarm signals, and a transmitting device that remotely transmits the selected moving image signal coded data via a communication path, A receiving device that receives the moving image signal encoded data transmitted via the communication path, a moving image signal decoding device that decodes the moving image signal encoded data, and a display device that displays a moving image signal Automatic video switching system characterized by the following. 25. A plurality of imaging devices, a plurality of video signal encoding devices that encode video signals from the imaging devices and output information necessary for detecting a moving object, and the video signal encoding. 18. A plurality of automatic alarm generation devices according to claim 17, wherein information from the device is input to detect a moving object, and moving image signal encoding for selecting moving image signal encoded data having motion based on a plurality of alarm signals. A data switching device, a transmitting device that remotely transmits selected moving image signal encoded data via a communication path, and a receiving device that receives moving image signal encoded data transmitted via the communication path,
An automatic video switching system, comprising: a video signal decoding device for decoding the video signal coded data; and a display device for displaying a video signal.