JPH06174818A - Image tracking device - Google Patents

Abstract

PURPOSE:To discriminate a target from a dummy target by excluding an area, where the maximum brightness of an object image is larger than a reference value and a difference between the maximum brightness and an average brightness is smaller than a reference value, as the dummy target. CONSTITUTION:The gravity center position in every connection area containing significant picture elements of a concentration image, binarized 2, is measured, an average brightness and the maximum brightness in every connection area are measured 7, 8 and a difference between the average brightness and the maximum brightness is measured 9, so that the maximum brightness may be compared with a reference value and then output to a bright difference judging means 11. The judging means 11 judges an area to be an effective area when the brightness difference is smaller than a reference value and the maximum brightness is smaller than a reference value and to be a dummy target when the brightness difference is smaller than the reference value but the maximum brightness is larger than the reference value. A significant target selecting means 12 selects an area where an error between the gravity center position and the previous tracking position is minimum out of the areas, judge to be effective areas, to output the gravity center position of the area as a current tracking position.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to improving the anti-interference property of an image tracking device.

[0002]

2. Description of the Related Art FIG. 10 is a block diagram of a conventional image tracking device,
FIG. 11 is a diagram showing a grayscale image provided by the image pickup apparatus.
In 0, 1 is an image pickup device, 2 is a binarizing unit, 3 is a barycentric position measuring unit for each area, 4 is a barycentric error measuring unit, 5 is a selecting unit based on a minimum value, and 6 is a tracking position output.

In the conventional image tracking device, the image pickup device 1 obtains a grayscale image as shown in FIG. The binarizing unit 2 binarizes the grayscale image, and the region-by-region centroid position measuring unit 3 determines the centroid position for each connected region composed of significant pixels of the binarized image obtained by the binarizing unit 2. To measure. The center-of-gravity error measuring means 4 measures the error between the center-of-gravity position of the area detected this time and the tracking position of the previous tracking position output 6. The selecting means 5 based on the minimum value selects the barycentric position of the area where the error output from the barycentric error measuring means 4 becomes the minimum. The center-of-gravity position of the area thus selected is output as the tracking position output 6 of this time.

FIG. 12 is a diagram showing data measured by a conventional image tracking device. The target and the pseudo target of the area detected as a significant pixel, the target barycentric position (XT, YT), and the pseudo target barycentric position ( XD, YD), and the center-of-gravity errors Tdif and Ddif of the target and pseudo target, respectively.

[0005]

Since the conventional image processing apparatus is configured as described above, the target and the pseudo target are close to each other, and the center of gravity of both of them and the tracking position of the previous tracking position output 6 are set. There is a problem in that the pseudo target may be mistakenly regarded as the target when the error of 1 is the same or the error of the pseudo target region is smaller than the error of the target region.

The present invention has been made in order to solve such a problem, and an object thereof is to obtain an image tracking device capable of discriminating both a pseudo target area and a target area even if they are close to each other. .

[0007]

An image tracking device according to the present invention, as means for discriminating between a target and a pseudo target, a maximum brightness measuring means for each area for measuring the maximum brightness of each area of a binary image, and a binary image. Area average brightness measuring means for measuring the average brightness of each area, a brightness difference measuring means for measuring two brightness differences of the output of the area maximum brightness measuring means and the area average brightness measuring means, and the area maximum brightness A maximum brightness determining unit that receives an output of the measuring unit and determines a region where the maximum brightness is greater than or equal to a reference value and a region that is less than the reference value; Brightness difference determining means and brightness difference determining means for determining an area for which the maximum brightness is determined to be a reference value or more by the maximum brightness determining means as a pseudo target, invalidating the pseudo target area, and valid for areas other than the pseudo target Is determined to be valid Error of the center of gravity position and the previous tracking position of the region within the region is provided with a significant target selection means for selecting a target area to be minimized.

In another embodiment, the brightness standard deviation measuring means for each area for measuring the brightness standard deviation of each area of the binary image,
A pseudo target and other areas are determined from the output of the area value measuring means for each area for measuring the area of each area, the output of the brightness standard deviation measuring means for each area, and the output of the area measuring means for each area, and the pseudo target area is invalidated or pseudo. Pseudo-target determination means that determines an area other than the target to be effective, and of the areas determined to be effective by the pseudo-target determination means, the area in which the error between the center of gravity of the area and the previous tracking position is the smallest is selected as the target. The target selecting means is provided.

In still another embodiment, the maximum brightness position measuring means for each area for measuring the position where the maximum brightness is given to each area of the binary image, and the maximum brightness position of each area outputted by the maximum brightness position measuring means for each area are set. Distance measuring means for measuring the distance to the position of the center of gravity, area-by-area area value measuring means for measuring the area of each area, and a distance normalized by the area is obtained from the output of the distance measuring means and the output of the area-by-area area value measuring means. If the value is less than or equal to the reference value, the pseudo target area is determined to be invalid by determining that it is a pseudo target, and the area other than the pseudo target is determined to be valid. The significant target selecting means is provided for selecting, as a target, a region in which the error between the barycentric position of the region and the previous tracking position is the smallest.

[0010]

It is known that an aircraft and a ship sail as a pseudo target while projecting flaming bombs. On the image picked up by the image pickup device, the flaming bullet has high brightness, and the brightness distribution spreads uniformly inside the area. On the other hand, the target aircraft and ship have different brightness in each part inside the target area and show a complicated brightness distribution. When the target and a plurality of pseudo targets are imaged, it can be said that the difference between the maximum brightness and the average brightness inside the target region is larger than the difference between the maximum brightness and the average brightness inside the pseudo target region. From this, the difference between the maximum brightness and the average brightness of each imaged object is calculated, and the area where the maximum brightness is larger than a certain reference value and smaller than the reference value where the difference between the maximum brightness and the average brightness is smaller is selected as a pseudo target. It can be seen that the target and the pseudo-target can be discriminated by considering them and excluding them.

When the target area has a certain area value, the brightness standard deviation is large because the brightness distribution is complicated inside the target area. On the other hand, in the pseudo target area, the brightness distribution spreads uniformly, so that the brightness standard deviation is smaller than that in the target area. With this, the area value and the brightness standard deviation of each imaged object are obtained, and a region whose area value is larger than a certain reference value and smaller than the brightness standard deviation is a standard value is selected and excluded as a pseudo target. Therefore, it can be seen that the target and the pseudo target can be distinguished.

Further, when comparing the position giving the maximum brightness and the barycentric position inside the area, in the pseudo target in which the brightness distribution is uniformly spread, the position giving the maximum brightness and the barycentric position are close to each other. On the other hand, in an aircraft or a ship, a part such as an engine provides a maximum brightness and is often not close to the center of gravity in the area. Further, the degree of these proximity depends on the area of the imaged pseudo target and the target region. Here, the distance between the maximum brightness position and the barycentric position of each imaged object is obtained, and the value is divided by the area value of each object to obtain a normalized distance that does not depend on the area. It is understood that the target and the pseudo target can be distinguished by comparing the normalized distance with a certain reference value, selecting an area smaller than the reference value, and regarding that area as a pseudo target and excluding it.

According to the present invention, the difference between the maximum brightness and the average brightness within the area of each imaged object is obtained, and the area having the maximum brightness is large and the difference between the maximum brightness and the average brightness is small is selected as a pseudo target. It is possible to distinguish the target from the pseudo-target by excluding it by considering it as. The target and the pseudo target can be distinguished by obtaining the area value and the brightness standard deviation of each imaged object, selecting a region having a large area value and a small brightness standard deviation, and excluding that region as a pseudo target. . In addition, the distance between the position giving the maximum brightness of each imaged object and the position of the center of gravity and the area value of each object are obtained, and the area in which the distance is normalized by the area value is selected as a pseudo target. By excluding it, the target and the pseudo target can be identified.

[0014]

【Example】

Example 1. FIG. 1 is a block diagram showing a first embodiment of an image tracking device according to the present invention. In FIG. 1, 7 is an average brightness measuring means for each area, 8 is a maximum brightness measuring means for each area, 9 is a brightness difference measuring means, 10 is a maximum brightness determining means, 11 is a brightness difference determining means, and 12 is a significant target selecting means. is there.

The image tracking device according to the present invention is an image pickup device 1.
To obtain a grayscale image. The binarization unit 2 binarizes the grayscale image, and the region-by-region barycentric position measuring unit 3 measures the barycentric position for each connected region composed of significant pixels of the obtained binary image. The area-based average brightness measuring means 7 measures the average brightness of each connected area. In the maximum brightness measuring means 8 for each area,
Maximum brightness is measured for each connected region. The brightness difference measuring means 9 obtains the difference between the maximum brightness output from the maximum brightness measuring means 8 for each region and the average brightness output from the average brightness measuring means 7 for each region by "Equation 1". Bmax in "Equation 1"
(I) is the maximum brightness of the area i output by the maximum brightness measuring means 8 for each area, and Bave (i) is the average brightness measuring means 7 for each area.
Is the average luminance of the region i output by Bdif (i) is the luminance difference of the region i output by the luminance difference measuring means 9.

[0016]

[Equation 1]

FIG. 4A is a diagram showing a luminance histogram inside the target area when the image of the target is picked up by the image pickup device 1. T is the threshold value of the binarization process, BTave is the average brightness inside the target area, BTmax is the maximum brightness inside the target area, and BTdif is the difference between the maximum brightness and the average brightness inside the target area. FIG. 4B is a diagram showing a luminance histogram inside the pseudo target area when the pseudo target is imaged by the imaging device 1. T is the threshold value of the binarization process, BDave is the average luminance inside the pseudo target area, BDmax is the maximum luminance inside the pseudo target area, and BDdif is the difference between the maximum luminance and the average luminance inside the pseudo target area. As shown in FIG. 4, the brightness histogram inside the target area has a large difference between the maximum brightness and the average brightness, whereas the brightness histogram inside the pseudo target area has a small difference between the maximum brightness and the average brightness and the maximum brightness. The value increases.

FIG. 5 is a flow chart of the maximum brightness determining means 10. In FIG. 5, BMref is a reference value for maximum brightness determination, and L (i) is a determination value of the maximum brightness determination means 10 in the area i. In step 13 of FIG. 4, the output value Bmax (i) of the maximum luminance measuring means 8 for each area and the reference value BMre.
Compare with f. If the maximum luminance output value Bmax (i) is equal to or greater than the reference value BMref, it is determined that the maximum luminance of the region i is equal to or greater than the reference value, step 14 is executed, and the determination value L (i) is set to 1. To do. Maximum brightness output value BMma
If x (i) is less than BMref, it is determined that the maximum brightness of the region i is less than the reference value, step 15 is executed, and the determination value L (i) is set to 0.

FIG. 6 is a flow chart of the brightness difference determining means 11. In FIG. 6, BDref is a reference value for brightness difference determination,
F (i) is a judgment value of the brightness difference judgment means 11 in the area i. In step 16, the output value Bdif (i) of the brightness difference measuring means 9 is compared with the reference value BDref. When the brightness difference output value Bdif (i) is larger than the reference value BDref, the region i is determined to be the valid region, step 17 is executed, and the determination value F (i) is set to 1. When the brightness difference output value Bdif (i) is less than or equal to the reference value BDref, step 18 is executed. In step 18, when the judgment value L (i) of the maximum brightness judgment means 10 is received and the judgment value L (i) is 0, the brightness difference of the region i is smaller than the reference value, but the maximum brightness is also smaller than the reference value. Since it is small, the region i is determined to be a valid region, step 17 is executed, and the determination value F (i) is set to 1
Set to. If the determination value L (i) is 1, the brightness difference in the area i is smaller than the reference value and the maximum brightness is larger than the reference value. Therefore, the area i is regarded as a pseudo target, and the area i is determined as a pseudo target. Is executed and the judgment value F (i) is set to 0.

Significant target selecting means 12 selects, from the areas determined to be effective areas by the brightness difference determining means 11,
The area where the error output from the center-of-gravity error measuring means 4 is minimized is selected, and the center-of-gravity position of the selected area is output as the current tracking position.

Example 2. FIG. 2 is a block diagram showing a second embodiment of the image tracking device according to the present invention. In FIG. 2, 20 is an area value measuring means for each area, 21 is a brightness standard deviation measuring means for each area, and 22 is a pseudo target determining means.

The image tracking device according to the present invention comprises an image pickup device 1.
To obtain a grayscale image. The binarizing unit 2 binarizes the grayscale image, and the region-by-region centroid position measuring unit 3 measures the centroid vector for each connected region formed by the significant pixels of the obtained binary image. The area-based average brightness measuring means 7 measures the average brightness of each connected area. The area value measuring means 20 for each area measures the area value for each connected area. The brightness standard deviation measuring means for each area 21 receives the output of the average brightness measuring means for each area 7 and the output of the area value for each area measuring means 20, and calculates "Equation 2".
The luminance standard deviation of each connected area is measured by. "Number 2"
, B (p, q) is the brightness at the point (p, q) inside the region, S (i) is the output value of the area value measuring means 20 for each region,
Bsd (i) is an output value of the brightness standard deviation measuring means 21.

[0023]

[Equation 2]

The pseudo target judging means 22 invalidates the area which is the pseudo target by judging the output of the brightness standard deviation measuring means 21 for each area and the output of the area value measuring means 20 for each area.
Other areas are determined to be valid.

FIG. 7 is a flow chart of the pseudo target judging means 22. In FIG. 7, BSref is a reference value for brightness standard deviation determination, Sref is a reference value for area value determination, and FD (i) is a determination value of the pseudo target determination means 22 in the region i. In step 23, the output value B of the brightness standard deviation measuring means 21
The sd (i) is compared with the reference value BSref. When the brightness standard deviation output value Bsd (i) is larger than the reference value BSref, the region i is determined to be a valid region, step 24 is executed, and the determination value FD (i) is set to 1. When the brightness standard deviation output value Bsd (i) is less than or equal to the reference value BSref, step 25 is executed. In step 25, the output value S (i) of the area value measuring means for each area is compared with the reference value Sref. If the area S (i) is less than the reference value Sref, the area may be a small target, so step 24 is executed and the determination value FD (i) is set to 1. If the area value S (i) is greater than or equal to the reference value Sref, the area value is large, but the standard deviation is small. Therefore, the area i is regarded as a pseudo target and the area i is determined as a pseudo target. And sets the determination value FD (i) to 0.

The significant target selecting means 12 selects and selects the area in which the error outputted by the center-of-gravity error measuring means 4 is the smallest, from the areas judged to be the effective areas by the pseudo target judging means 22. The center-of-gravity position of the selected region is output as the tracking position this time.

Example 3. FIG. 3 is a block diagram showing a third embodiment of the image tracking device according to the present invention. In FIG. 3, 27 is a maximum brightness position measuring unit for each area, 28 is a distance measuring unit, and 29 is a normalized distance determining unit.

The image tracking device according to the present invention comprises an image pickup device 1.
To obtain a grayscale image. The binarizing unit 2 binarizes the grayscale image, and the region-by-region centroid position measuring unit 3 measures the centroid vector for each connected region formed by the significant pixels of the obtained binary image. The maximum brightness position measuring means 27 for each area measures the position of the point giving the maximum brightness for each connected area. The area value measuring means 20 for each area measures the area value for each connected area. The distance measuring means 28 receives the output of the maximum brightness position measuring means 27 for each area and the output of the center of gravity position measuring means 3 for each area, and measures the distance between the maximum brightness position and the center of gravity position for each connected area by "Equation 3". . In “Equation 3”, (BMx
(I), BMy (i) is the maximum brightness position of the area i output by the maximum brightness position measuring means 27 for each area, (Gx (i),
Gy (i) is the barycentric position of the region i output by the barycentric position measuring unit 3 for each region, and A (i) is the distance of the region i output by the distance measuring unit 28.

[0029]

[Equation 3]

The normalized distance determining means 29 receives the output of the distance measuring means 28 and the output of the area-by-area area value measuring means 20, obtains the distance normalized by the area, and determines the value to be a pseudo target. It is determined that the area is invalid and the other areas are valid.

FIG. 8 is a flow chart of the normalized distance judging means 29. In FIG. 8, P (i) is a normalized distance value in the area i, Pref is a reference value for determining the normalized distance, and FA (i) is a normalized distance determining means 2 in the area i.
It is a judgment value of 9. In step 30, the distance value P (i) normalized by the area value is obtained from "Equation 4".

[0032]

[Equation 4]

FIG. 9A is a diagram showing the relationship between the position of the center of gravity and the position at which the maximum luminance value is given when the target is imaged. In FIG. 9A, TB is a position where the maximum brightness is given, and TG is a center of gravity position. FIG. 9B is a diagram showing the relationship between the position of the center of gravity and the position at which the maximum brightness value is given when the pseudo target is imaged. In FIG. 9B, DB is a position where the maximum brightness is given, and DG is a barycentric position. As shown in the figure, when the target is imaged, the position T that gives the maximum brightness
B is separated from the center of gravity position TG, and the distance is large. on the other hand,
When the pseudo target is imaged, the area spreads out in a circle centering on the center of gravity point, and the position DB and the center of gravity position DG that give the maximum brightness are obtained.
Are close.

The distance between the position of maximum brightness and the position of the center of gravity depends on the area as the area of the area increases. When the area spreads in a circle like a pseudo target, the square root of the area value is proportional to the radius of the circle. By dividing the distance between the position giving the maximum brightness and the position of the center of gravity by the square root of the area value, the normalized distance can be compared without depending on the area.

In step 31, the value of the normalized distance P (i) calculated in step 30 is compared with the reference value Pref. When the normalized distance P (i) is larger than the reference value Pref, it is determined to be the effective area, step 32 is executed, and the determination value FA (i) is set to 1. When the normalized distance P (i) is less than or equal to the reference value Pref,
It is determined to be a pseudo target, step 33 is executed, and the determination value FA (i) is set to 0.

The significant target selecting means 12 selects and selects the area in which the error outputted by the center-of-gravity error measuring means 4 is the smallest, out of the areas judged by the normalized distance judging means 29 to be the effective areas. The center of gravity of the area is output as the tracking position for this time.

[0037]

As described above, according to the present invention, the maximum brightness of each region of the binary image is measured, and the brightness difference between the maximum brightness and the average brightness is measured, and the brightness difference is equal to or less than the reference value. And the difference between the center of gravity of the area and the previous tracking position within the area determined to be valid by the brightness difference determination means is provided by providing a brightness difference determination means that determines that the area whose maximum brightness is greater than or equal to the reference value is a pseudo target. By including the significant target selecting unit that selects the region in which the target is the minimum, the target can be tracked even when the target and the pseudo target are imaged in close proximity.

According to another embodiment, the brightness standard deviation and the area value of each area of the binary image are measured, and an area having a small brightness standard deviation despite a large area value is regarded as a pseudo target and is invalid. And a significant target selecting means for selecting, as a target, an area in which the error between the center of gravity of the area and the previous tracking position is the smallest among the areas determined to be effective by the pseudo target determining means. With the provision, the target can be tracked even when the target and the pseudo target are imaged close to each other.

According to still another embodiment, the distance between the position giving the maximum brightness and the position of the center of gravity of each region of the binary image is measured, the area value of each region is measured, and the distance is calculated as the area value. The normalized distance obtained by dividing by, the normalized distance determining means that regards the area with a small normalized distance as a pseudo target and determines that it is invalid is provided, and within the area determined to be valid by the normalized distance determining means. Tracking the target even when the target and the pseudo target are imaged close to each other by providing the significant target selection means for selecting the region where the error between the center of gravity of the region and the previous tracking position is the minimum as the target. You can

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a first embodiment of an image tracking device according to the present invention.

FIG. 2 is a configuration diagram showing a second embodiment of the image tracking device according to the present invention.

FIG. 3 is a configuration diagram showing a third embodiment of an image tracking device according to the present invention.

FIG. 4 is a configuration diagram showing a luminance histogram of a target and a pseudo target.

FIG. 5 is a flowchart showing an explanation of the operation of the maximum brightness determining means.

FIG. 6 is a flowchart showing the operation of the brightness difference determination means.

FIG. 7 is a flowchart showing the operation of the pseudo target determination means.

FIG. 8 is a flowchart showing the operation of the normalized distance determination means.

FIG. 9 is a diagram for explaining a distribution of luminance of a target and a pseudo target.

FIG. 10 is a configuration diagram showing a conventional image tracking device.

FIG. 11 is a diagram showing a grayscale image provided by the image pickup device.

FIG. 12 is a diagram for explaining the operation of a conventional image tracking device.

[Explanation of symbols]

 7 Average brightness measurement means for each area 8 Maximum brightness measurement means for each area 9 Brightness difference measurement means 10 Maximum brightness determination means 11 Brightness difference determination means 12 Significant target selection means 20 Area value measurement means for each area 21 Brightness standard deviation measurement means for each area 22 Pseudo target determining means 27 Maximum brightness position measuring means for each area 28 Distance measuring means 29 Normalized distance determining means

Claims (3)

[Claims] 1. A significant region is selected as a target from a region in a screen obtained by obtaining an image signal from an image pickup device and binarizing the image signal, and a barycentric position of the region is set as a tracking position. In the image tracking device that outputs and tracks,
The binarizing means for binarizing the image signal, the region-by-region centroid position measuring means for measuring the region-by-region centroid position obtained by the binarizing means, and the region-by-region centroid position measuring means output. A center-of-gravity error measuring unit that measures the error between the center-of-gravity position and the previous tracking position, an average brightness measuring unit for each region that measures the average brightness of each region, and a maximum brightness measuring unit that measures the maximum brightness of each region. A brightness difference measuring means for measuring a brightness difference between the output of the area average brightness measuring means and the area maximum brightness measuring means, and the maximum brightness by receiving the area maximum brightness measuring means output. Is a maximum brightness determining means for determining an area having a reference value or more and an area less than the reference value, and the brightness difference measuring means of the area having the maximum brightness determined by the maximum brightness determining means as a reference value or more. If the brightness difference output is less than the reference value And a significant target selecting unit that determines the output of the center-of-gravity error measuring unit and the output of the luminance difference determining unit and selects the most significant region as a target. An image tracking device characterized by the above. 2. A binarizing unit for binarizing an image signal, a region-by-region centroid position measuring unit for measuring the region-by-region centroid position obtained by the binarizing unit, and the region-by-region centroid position. A center-of-gravity error measuring unit that measures an error between the center-of-gravity position output by the measuring unit and the previous tracking position, a region-standard luminance standard deviation measuring unit that measures the luminance standard deviation of each region, and an area value of each region are measured. Area-by-area area value measuring means, pseudo-target determining means for determining an area that is a pseudo-target from the output of the area-by-area luminance standard deviation measuring means, and the output of the area-by-area area value measuring means, and the centroid error measurement 2. The image tracking apparatus according to claim 1, further comprising a significant target selecting unit that determines the most significant region as a target by judging from the output of the unit and the output of the pseudo target determining unit. 3. A binarizing unit for binarizing an image signal, a region-by-region centroid position measuring unit for measuring the region-by-region centroid position obtained by the binarizing unit, and the region-by-region centroid position. A center-of-gravity error measuring means for measuring an error between the center-of-gravity position output by the measuring means and a previous tracking position, and a maximum-brightness position measuring means for each area for measuring a position providing maximum brightness for each area,
An area value measuring means for each area for measuring an area value for each area, a distance measuring means for measuring the distance between the maximum brightness position output by the maximum brightness position measuring means for each area and the barycentric position for each area, Normalized distance determining means for determining a normalized distance based on the area value from the area value measuring means for each area and the output of the distance measuring means, and the center of gravity error 2. The image tracking device according to claim 1, further comprising a significant target selecting unit that determines from the output of the measuring unit and the output of the normalized distance determining unit to select the most significant region as a target.