JP2829960B2 - Background information extraction method - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for detecting a moving object such as a vehicle or an aircraft, and more particularly to a background information extracting method for extracting background information other than a moving object.

[Conventional technology]

When a moving object is detected by processing a video signal obtained from a television camera or the like, recognizing a portion that is not a moving object, so-called background information, is extremely heavy in order to obtain good detection accuracy. For this reason, Japanese Patent Application No. 55-188451 has been proposed. In this method, the difference between neighboring pixel values in the scanning line direction in an image sampled two-dimensionally is calculated, and the image is changed to “1” and “0” according to the magnitude of the difference.
Value. Then, the number of "1" s in the same scanning line direction in the measurement area is set as the characteristic value of the scanning line, and the presence of the traveling vehicle is detected using the characteristic value. In this case, as the extraction of the background value, the maximum value of the expected background value is set as the threshold value A in advance. When a characteristic value on a scanning line smaller than the threshold value A is obtained, it is recognized as a background value.

A method for automatically setting the feature value of the scanning line described above is described in IEICE Transactions Vol. J68-D No. 3, March 1985.
It is proposed in "Traffic flow measurement system using image processing method". In this method, the maximum value x and the minimum value y are obtained from the values generated as the scanning line characteristic values of the same scanning line within a certain period, except when it is determined that the vehicle is present. Thereafter, the background value is recognized as a value sandwiched between x + α, y−α (α and β are constants), and are sequentially measured.

[Problems to be solved by the invention]

However, in the above-described “vehicle presence position detection method”, since the threshold value A of the background information is set in advance, if the background information changes every moment due to weather conditions or the like, accurate position detection is impossible. Further, according to the scanning of the threshold value A, there is a possibility that a characteristic value of a scanning line generated when a vehicle is present is erroneously processed as background information.

"Traffic flow measurement system using image processing method"
Means that the background value is sequentially measured as a value sandwiched between x + α and y + β, so that the background value in the n-th cycle is x + n
α, y−nβ. That is, when the measurement was continued sequentially, the upper limit (x + nα) of the background value gradually increased, and the lower limit (y-nβ) approached 0, so that normal measurement was impossible.

As described above, there is no precedent for a method of effectively extracting a feature value corresponding to background information by a method of processing a scanning line and automatically performing the extraction.

[Means for solving the problem]

The background information extraction method of the present invention calculates the difference between neighboring pixels in the scanning line direction in an image sampled two-dimensionally, and calculates the difference between “1” and “0” based on a predetermined threshold. This is a method of processing as a scanning line feature value. When processing a signal that periodically generates a fixed number of images, the number of scanning line feature values “1” on the same scanning line within a predetermined time is determined. Then, the maximum value of the number of “1” of the scanning line feature value is extracted as background information in the image.

In addition, an area determination unit is provided, a measurement area excluding objects other than the background is determined by the area determination unit, and the number of scanning line feature values “1” on the same scanning line within a predetermined time in the measurement area is calculated. Then, the maximum value of the number of “1” of the scanning line feature value is extracted as background information in the image.

Further, an area determining unit and a background validity determining unit are provided, and the area determining unit determines a measurement area excluding objects other than the background,
Unless the presence of an object other than the background is determined in the measurement area by the predetermined condition of the background validity determination unit, the number of “1” of the scanning line feature value on the same scanning line within a predetermined time is calculated. Then, the maximum value of the number of “1” of the scanning line feature value is extracted as background information in the image.

[Action]

The difference between neighboring pixel values in the scanning line direction in the image is calculated, and the image information is changed to “1” and “0” according to the magnitude of the value.
At the same time, the value “1” on the scanning line that occurs most frequently within a predetermined time is extracted as background information in the image.

Further, the background information that has already been extracted is compared with the number of “1” on the scanning line, and the background information is extracted except when there is an object other than the background.

Further, a range between scanning lines on the image is set, and after it is determined whether an object other than the background exists within this range, background information is extracted.

〔Example〕

Next, the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a first embodiment according to the present invention.
In the figure, 10 is an input unit for inputting a video signal of a television camera, 11 is a shift register, 12 is a difference calculation unit that calculates an absolute value by subtracting a signal from the input unit 10 and the shift register 11,
13 is a threshold value comparing unit that compares the calculation result of the difference calculating unit 12 with a predetermined threshold value, and 14 is a feature value updating unit that updates a feature value on a scanning line based on the result of the threshold value comparing unit 13 ,
15 is an area storage unit that stores the measurement area, 19 is a storage command unit that operates the feature value storage unit 20 by the feature value update unit 14,
20 is a feature value storage unit that stores the frequency of occurrence of scanning line feature values in each scanning line, 21 is a pixel number counting unit that measures the pixel number of the screen, 22 is a cycle counting unit that measures the cycle of the screen, 23 Is a cycle storage unit that stores the number of screen cycles, 24
Is a background information extraction unit that extracts background information from the feature value storage unit 20, and 25 is a background information storage unit that stores the extraction result of the background information extraction unit.

Next, the operation will be described. In general, i rows j
The video signals ai, j that can be assumed to be in a column array are quantized,
It is input to the input unit 10. The input unit 10 sequentially receives values corresponding to each pixel, and inputs the values to the shift register 11 and the difference calculation unit 12, respectively. The difference calculation unit 12 receives the video signal from the shift register 11 and the input unit 10, and calculates the absolute value of the difference. The delay time from input to output in the shift register 11 means the interval between pixels to be compared on the same scanning line. If the delay time is k, the output of the shift register is a
_{i, j−k} . Therefore, the difference calculation unit 12 calculates | a _{i, j−k} −a
_{i, j} | is output to the threshold value comparison unit 13. The threshold comparing unit 13 includes a predetermined threshold TH and a difference calculating unit.
12 are compared, and if | a _{i, j−k} −a _{i, j} | ≧ TH, “1” is set, and if | a _{i, j−k} −a _{i, j} | ≦ TH, “0” is set as a pixel. _Is sent to the feature value updating unit 14 as the feature value P _{i, j−k} . The feature value updating unit 14 reads the left and right values (j ₁ , j ₂ ) of the region of the i-th row from the region storage unit 15 that stores the measurement region when counting the sum of “1” in the measurement region. _{If 1} j−kj ₂ , the above-mentioned feature value P _{i, j−k} is added, and V _i + P _{i, j−k}
Updating scan line feature value V _i as. Feature value updating unit 14 outputs the scanning line feature values V _i and completes the process in the storage instruction unit 19. As described above, the storage command unit 19 reads the left and right values (j ₁ , j ₂ ) of the i-th row area and stores the frequency of occurrence of the scanning line characteristic value Vi of each scanning line if jk = j _2. Feature value storage unit
The value of the corresponding area of 20 is incremented by one. The feature value storage unit 20 is a two-dimensional memory, and stores the frequency of occurrence of the scanning line feature value Vi in the i-th scanning line as an element of the i-th row and the Vi column. When the processing of the storage command unit 19 is completed, the control is transferred to the pixel number counting unit 21. Note that in the above description, j
The -k = j ₂ unless storage instruction unit 19 to the currency control the pixel number counting unit 21 moves to. The pixel number counting unit 21 measures the pixel number of the screen, and when the value of the j-th column reaches the pixel number corresponding to the right end of the screen, the scan line number of the i-th row is incremented by one, and the value of the j-th column is the left end of the screen. Match to value 1. When the values of the i-th row and the j-th column correspond to the pixel number corresponding to the lower right corner of the screen, the control is transferred to the cycle counting section 22. If the pixel number at the lower right of the screen has not been reached, the control is returned to the input unit 10, and the above-described processing is repeated. Period counting section
Numeral 22 measures the well-known state of the screen, and increases one count when control is transferred. Further, the value of the cycle number storage unit 23 in which the cycle for extracting the background is stored in advance by the number of screens is compared with the value measured by the cycle count unit 22. And
If the values match, the content of the cycle counting unit 22 is reset to 0, and control is transferred to the background information extraction unit 24. If they do not match, control is returned to the input unit 10. The background information extraction unit 24 extracts the maximum value among the scanning line feature values of each row of the feature value storage unit 20 as background information of each scanning line in order to extract background information, and outputs it to the background information storage unit 25. Then, the value of the i-th row and the j-th column is reset to 0, and the control is returned to the input unit 10. The background information storage unit 25 can store the background information in a one-dimensional array.

FIG. 2 is a block diagram showing a second embodiment according to the present invention. In the figure, the same parts as those in FIG. 1 are denoted by the same reference numerals. 16 is an area determination unit.

In this embodiment, the operation from the input unit 10 to the feature value updating unit 14 is the same as in the first embodiment. When the area determination unit 16 receives the scanning line feature value Vi from the feature value update unit 14, the left and right values (j ₁ , j ₂ ) of the i-th row area are received from the area storage unit 15.
Reading, transfers control to store instruction unit 19 if j-k = j _2. If this condition is not satisfied, control is transferred to the pixel number counting section 21. The storage command unit 19 refers to the content + _hoi of the background information storage unit 25 that stores the background information, and uses the preset values α and β to
h _oi -αVi + h _oi + transfers control if β to the pixel number counting section 21. When this condition is not satisfied, the value of the corresponding area of the feature value storage unit 20 storing the occurrence frequency of the scanning line feature value obtained by the feature value update unit 14 is incremented by one. Subsequent operations are the same as in the first embodiment. Therefore, in the second embodiment, the measurement area is limited by the area determination unit 16 and compared with the data of the background information storage unit 25 already stored by the storage command unit, so that an object other than the background exists. And the number of the most frequently occurring scanning line feature values can be extracted as background information in the image.

FIG. 3 is a block diagram showing a third embodiment according to the present invention. In the figure, the same parts as those in FIGS. 1 and 2 are denoted by the same reference numerals. Reference numeral 17 denotes a background validity determination unit, and reference numeral 18 denotes a flag storage unit.

In this embodiment, the operation from the input unit 10 to the area determination unit 16 is the same as in the second embodiment. When the background validity determining unit 17 receives the control from the area determining unit 16, the background validity determining unit 17 refers to the content + h _oi of the background information storage unit 25 and uses the preset values α and β to obtain + h _oi −αVi + h _oi
If it is + β, the value of the i-th row of the corresponding scanning line in the flag storage unit 18 storing the scanning line characteristic value of each scanning line in the same image is written as the scanning line characteristic value Vi. If this condition is not satisfied, the flag ** is written and the storage instruction unit 19 is written.
Transfer control to As shown in the arrangement diagram of the scanning line characteristic values shown in FIG. 4, the storage command unit 19 obtains the characteristics from the scanning line i-n-m to the scanning line i when the scanning line characteristic value Vi of the scanning line i is obtained. Whether the value is effective as a background or not is determined based on the value of the flag storage unit 18. That is, if the ** flag is written in the flag storage unit 18, it is determined that there is something different from the background from the scanning line i-n-m to the scanning line i, and the scanning line i-n
Feature value storage unit not recognized as background information scanning line feature value V _in the
Do not manipulate the value of 20. On the other hand, if the ** flag is not written in the flag storage unit 18, it is determined that the background from the scanning line inm to the scanning line i is the background, and the scanning line characteristic value Vin of the scanning line _in is determined. As background information, i-
the value of n row V _in the column is incremented by one. Here, m and n indicate the number of scanning lines. When the above operation is completed, the storage unit 19 transfers control to the pixel number counting unit 21. Subsequent operations are the same as in the first embodiment. Therefore, in the third embodiment, the range between scanning lines is set by the storage command unit 19, and after recognizing whether or not an object other than the background exists in this range, the most frequently occurring scanning line characteristics are set. The number of values can be extracted as background information in the image.

(Effect of the present invention)

As described above, according to the background information extracting method of the present invention, a difference between neighboring pixel values in the scanning line direction in an image sampled two-dimensionally is obtained, and the image information is set to “1” based on the magnitude of the value. ”And“ 0 ”, and the number of“ 1 ”s on the scanning line that occurred most frequently within a predetermined time is extracted as background information in the image. Even if it changes, background information that satisfies such conditions can be extracted. As a result, conventional vehicles,
When detecting a moving object such as an aircraft, the measurement accuracy can be improved.

Further, since background information can be automatically extracted, the processing can be speeded up.

Furthermore, even if an object other than the background exists in the same image information, the background information can be extracted by excluding only the object, so that the moving object and the background information are not confused. It has a great effect.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment according to the present invention,
FIG. 2 is a block diagram showing a second embodiment, and FIG.
FIG. 4 is an arrangement diagram of scanning line characteristic values. 10 input section, 11 shift register, 12 difference calculating section, 13 threshold comparing section, 14 feature value updating section, 15
… Area storage unit, 16… Area determination unit, 17… Background validity determination unit, 18… Flag storage unit, 19… Storage instruction unit, 20…
Feature value storage unit 21, pixel number counting unit 22, period counting unit 23, period number storage unit 24, background information extraction unit

Claims (3)

(57) [Claims] A difference between neighboring pixels in a scanning line direction in an image sampled two-dimensionally is calculated, and the difference is calculated based on a predetermined threshold value as a scanning line feature value of “1” or “0”. When processing a signal that periodically generates a fixed number of images, the number of scanning line feature values “1” on the same scanning line within a predetermined time is calculated, and scanning is performed. A background information extraction method, wherein a maximum value of the number of line feature values "1" is extracted as background information in an image. 2. A scanning line characteristic value of "1" and "0" is obtained based on a predetermined threshold value by taking a difference between neighboring pixels in a scanning line direction in an image sampled two-dimensionally. In the case of processing a signal that periodically generates a fixed number of pixels, an area determining unit is provided, and the area determining unit defines a measurement area excluding objects other than the background, Calculating the number of scanning line feature values “1” on the same scanning line within a predetermined time, and extracting the maximum value of the number of scanning line feature values “1” as background information in the image. Characteristic background information extraction method. 3. A difference between neighboring pixels in a scanning line direction in an image sampled in a two-dimensional manner is obtained, and the difference is calculated based on a predetermined threshold value as a scanning line characteristic value of "1" or "0". When processing a signal that periodically generates a fixed number of images, a region determining unit and a background validity determining unit are provided, and the region determining unit excludes objects other than the background. A measurement area is determined, and a scanning line feature value on the same scanning line within a predetermined time period except when it is determined in the measurement area that an object other than the background is determined based on a predetermined condition of the background validity determination unit. A method of calculating the number of “1” in the image data and extracting the maximum value of the number of “1” in the scanning line feature value as background information in the image.