JPH11232435A - Picture processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To extract only a part where luminance and the like locally protrude by changing the attribute value of an objective picture, based on a deviation and variations, which are operated from the attribute value of a picture element group surrounding the objective picture selected from an input picture and a dispersion degree. SOLUTION: A picture input means 1 has a CCD camera 1a and an input picture memory 1b and stores an input picture, for example. A picture element group selection means 2 selects a prescribed picture element group surrounding the objective picture element in the inputted picture. An average value operation means 3 operates the attribute value of the selected picture element group and a dispersion value operation means 4 operates the dispersion value of the attribute value of the picture element group. A deviation operation means 5 operates a deviation between the average value operated by the operation means 3 and the attribute value of the objective picture element. An attribute value change means 6 changes the attribute value of the objective picture element, based on the operated deviation and the dispersion value operated by the operation means 4. Thus, the output picture from which the picture element where a spot and the like locally protrude is extracted is obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image processing apparatus, and more particularly to an image processing apparatus for generating an output image by extracting a bright point from a picked-up image including a bright point generated by laser light irradiation. .

[0002]

2. Description of the Related Art For example, as shown in FIG. 5, a cargo handling vehicle disclosed in Japanese Patent Application No. 7-221724 is irradiated from a vehicle body 52 onto a pallet 51 when passing through a U-shaped pallet 51. The scattered light of the laser light 53 is imaged by the CCD camera 54, and the pallet 5
1 is measured. The scattering point of the laser light 53 on the pallet 51 is a luminescent point whose luminance is more prominent than other parts on the image. Measurement control is performed based on the bright spot. In addition to the above example, a point light source such as an LED or a mark colored in a predetermined color is attached to an object and an image is taken with a CCD camera or the like, and the point light source or a bright point by a sphere is recognized. There is also a device that measures the position and orientation of the target object by performing the operation. However, since the captured image includes not only the scattered light of the laser light but also the background light from another light source, it is necessary to perform image processing for emphasizing the bright spot and extract the image from the captured image. The simplest method of extracting a bright point is a method of binarizing an image with a predetermined luminance threshold. That is, in the input image, the brightness value of a portion darker than the threshold value is 0, and the brightness value of a bright portion is 1
In this method, a value image is created, and a portion having a luminance value of 1 in the binary image is recognized as a bright point. However, when the input image is directly binarized, the brightness of the background portion other than the bright spot fluctuates, and if there is a brighter portion than the bright spot, the bright spot may be missed or erroneously recognized. is there. It is also possible to reduce the influence of background light by attaching a filter that transmits only the wavelength of the laser beam used for position and orientation to a CCD camera or the like that performs image capturing. In addition, there are problems that it is difficult to sufficiently remove the background light and that the effect is not obtained so much for an object having poor monochromaticity such as an LED or a colored mark. On the other hand, as a method of extracting a sharp change in luminance value such as a bright spot, it has been conventionally known to perform a pre-processing using a spatial filter on an input image before binarization. For example, a portion where the luminance value changes is extracted by performing a differentiation process in one or both of the horizontal and vertical directions of the input image. FIG. 6 is a diagram illustrating an example of a simple first-order spatial differential filter. The mask of this spatial differential filter is 3 × 3, and the central pixel is the target pixel. FIG. 6A shows the configuration of the mask in the horizontal direction, and FIG.
Shows the configuration of the mask in the vertical direction. In this spatial filter, the influence of noise is somewhat suppressed by adding three adjacent differential values (difference values).
The primary spatial differential filter is effective, for example, when the luminance value changes in a step-like manner. A portion where the luminance value changes can be extracted based on whether or not the luminance value is larger.

[0003]

However, even if the technique of extracting a luminance value changing portion using a spatial filter is simply applied to the technique of measuring the position and orientation based on the bright spots and the technique of recognizing the bright spots, the above-described differential processing does not. In addition, not only the bright spot but also the edge portion where the luminance value changes linearly cannot be extracted, so that only the bright spot cannot be extracted, and accurate measurement of the position and orientation cannot be performed. The present invention has improved an image processing apparatus in order to solve such problems in the conventional technology,
It is an object of the present invention to provide an image processing apparatus that performs processing for extracting only a portion where an attribute value such as luminance locally protrudes.

[0004]

According to one aspect of the present invention, there is provided an image input means for inputting an image composed of a plurality of pixels having a predetermined attribute value. Group selecting means for selecting a predetermined pixel group surrounding a certain target pixel in the image input by the above, and representative value calculating means for calculating a representative value of the attribute value of the pixel group selected by the pixel group selecting means A variation calculating means for calculating the variation of the attribute value of the pixel group;
Deviation calculating means for calculating a deviation between the representative value calculated by the representative value calculating means and the attribute value of the target pixel;
On the basis of the deviation calculated by the deviation calculating means and the variation calculated by the variation calculating means,
The image processing apparatus includes an attribute value changing unit that changes the attribute value of the target pixel. Also,
According to a second aspect of the present invention, in the image processing apparatus according to the first aspect, the attribute value is changed by the attribute value changing means for each pixel included in a predetermined area of the image, and the attribute value is locally changed. The gist of the present invention is to create an output image in which pixels having a prominent value are extracted. According to a third aspect of the present invention, in the image processing apparatus according to the first or second aspect, the pixel group selected by the pixel group selecting means is formed in an annular shape separated by a predetermined distance from the target pixel. The gist is that it is a pixel group of. Further, the invention according to claim 4 is the invention according to any one of claims 1 to 3 above.
In the image processing device described in the paragraph, the gist is that the attribute value is a luminance value of each pixel.

According to a fifth aspect of the present invention, in the image processing apparatus according to the fourth aspect, the representative value is an average value of luminance values of pixels included in the pixel group. And According to a sixth aspect of the present invention, in the image processing apparatus according to the fourth or fifth aspect, the degree of variation is a variance value of a luminance value of each pixel included in the pixel group. I do. According to a seventh aspect of the present invention, in the image processing apparatus according to any one of the first to sixth aspects, the pixel group selecting means selects at least two different pixel groups for a certain target pixel. ,
The gist is that the calculation of the representative value by the representative value calculating means and the calculation of the variation degree by the variation degree calculating means are performed for different pixel groups. The invention described in claim 8 is the same as that in claims 1 to 7 above.
In the image processing apparatus according to any one of the above, the attribute value change means is proportional to the deviation calculated by the deviation calculation means, and adds a predetermined value to the variation calculated by the variation calculation means. The gist is that the attribute value of the target pixel is changed in inverse proportion to the calculated value.
According to a ninth aspect of the present invention, in the image processing apparatus according to any one of the first to eighth aspects, the image input by the image input means is provided so that the transport vehicle identifies an obstacle. The gist is that the scattered light of the laser light applied to the obstacle is imaged. Claim 1
In the image processing device according to any one of the above items 9 to 9, by comparing the target pixel with the attribute value of a pixel group surrounding the target pixel, a pixel having a locally prominent attribute value such as a bright spot is emphasized. An output image can be created, and a bright point can be easily detected by performing a binarization process or the like at a later stage. Therefore, when the image processing device is used for a carrier, favorable position and orientation measurement can be performed.

[0006]

Embodiments of the present invention will be described below with reference to the accompanying drawings to provide an understanding of the present invention.
The following embodiment is a specific example of the present invention and does not limit the technical scope of the present invention. FIG. 1 is a functional block diagram illustrating a schematic configuration of an image processing apparatus according to an embodiment of the present invention. As shown in FIG. 1, an image processing apparatus according to an embodiment of the present invention stores, for example, a CCD camera 1a to which an image composed of a plurality of pixels having attribute values such as luminance values is input and an input image. Image input means 1 having an input image memory 1b of
A pixel group selecting unit 2 for selecting a predetermined pixel group surrounding a certain target pixel in the image input by the image input unit 1
A mean value calculating means (corresponding to a representative value calculating means) 3 for calculating an average value of the attribute values of the pixel group selected by the pixel group selecting means 2, and a variance value of the attribute value of the pixel group. A variance value calculating means (corresponding to a variation degree calculating means) 4 for calculating; a deviation calculating means 5 for calculating a deviation between the average value calculated by the average value calculating means 3 and the attribute value of the target pixel; An attribute value changing means for changing the attribute value of the target pixel based on the deviation calculated by the deviation calculating means and the variance value calculated by the variance value calculating means; The attribute value is changed by the means 6 for each pixel included in a predetermined area of the image and output to the output image memory 7, and the output image from which the pixel whose attribute value locally protrudes, such as a bright spot, is extracted. Create

Hereinafter, the details of the bright spot extraction processing by the image processing apparatus will be described. FIG. 2 and FIG. 3A are diagrams illustrating examples of an input image processed by the image processing apparatus. FIG. 3B is a diagram showing an example of an output image processed by the image processing apparatus. In the above image processing apparatus, an input image taken by the CCD camera 1a is stored in an input image memory 1b. The input image is obtained by, for example, capturing an image of scattered light of a laser beam applied to the obstacle by the carrier to identify the obstacle.
In this input image, a plurality of pixels are arranged in a rectangular shape as shown in FIGS. 2 and 3A. The numbers shown in the grid in FIG. 3B represent the attribute values of each image. This attribute value is, for example, a luminance value. Further, the pixel represented by T in FIG. 3A corresponds to the center of the luminescent spot. This pixel T has the same attribute value 129 as the pixel D before image processing is performed. The input image is stored in the memory space of the input image memory 1b for each pixel. In FIG. 2, for example, the input image memory 1b
Has a row number of H, a column number of W, and an i row j in the input image memory 1b.
The pixel in the column is P (i, j) [1 ≦ i ≦ H, 1 ≦ j ≦ W]
It is represented by In the image processing apparatus, the input image memory 1
b, for example, E ≦ i ≦ HE, E
The following processing is performed for each pixel in a predetermined area where ≦ j ≦ WE (E is a constant). First, a target pixel P (i, j) is selected, and the pixel group selecting unit 2
An annular pixel group separated by a predetermined interval from the target pixel P (i, j) is selected. In FIG. 2, the pixel group selected by the pixel group selecting means 2 is a plurality of pixels indicated by oblique lines. The predetermined interval is set based on, for example, the approximate radius of the luminescent spot to be detected. In the example of FIG. 2, this interval is set to 4 in units of pixels. Also,
The constant E is provided to limit the range of image processing so that a pixel group does not protrude from a specified area when processing pixels at the periphery of an image, and is set to a value larger than the above-described interval (see FIG. In the example of 2, it is 5 in units of pixels). still,
Peripheral pixels (0 ≦ i ≦ E, HE ≦ i ≦ H, 0 ≦ j <
E, WE <pixels with j <W), the attribute value of which is set to 0 in both the input image and the output image. The average value calculating means 3 calculates the average value Ic of the attribute values of the pixel group selected by the pixel group selecting means 2. I ₁ the attribute value of each pixel of the pixel group, I _2, ..., When I _n (n is the number of pixels of the pixel group), the average value Ic is

(Equation 1) It is. On the other hand, the variance calculating unit 4 calculates the variance Sc of the attribute value of the pixel group selected by the pixel group selecting unit 2. The variance Sc is

(Equation 2) It is. The average value Ic of the pixel group calculated by the average value calculating means 3 is supplied to the deviation calculating means 5, where the attribute value I (i, j) of the target pixel P (i, j) is provided.
j) and the deviation Is = I (i, j) −I between the average value of the pixel group
c is calculated. Then, in the attribute value changing means 6, the attribute value I ′ (i, i, j) in the output image of the target pixel P (i, j)
j) is calculated according to the following equation. However, when the calculated attribute value I ′ (i, j) becomes negative, the attribute value of the pixel in the output image is changed to 0. I ′ (i, j) = k ₁ (I (i, j) −Ic) / (Sc
+ K ₂ ) Here, k ₁ and k ₂ are predetermined constants, where k ₁ corresponds to a gain and k ₂ corresponds to an offset.

These processes are repeated for each pixel included in the predetermined area, and an output image having the attribute value I ′ (i, j) is stored in the output image memory 7. The output image is an image in which a luminescent spot or the like whose attribute value protrudes locally in the image is extracted. The bright spot in the image becomes a peak having a larger luminance value than the background portion in the vicinity. Therefore, if the target pixel P (i, j) is the center T of the bright spot, the deviation I
Since s has a large positive value, the attribute value I ′ (i, j) in the output image also has a large value. On the other hand, the deviation Is is small in the pixels other than the bright spots whose luminance values are close to uniform, and the attribute value I ′ (i, j) in the output image is also small. In a portion where the luminance value changes linearly, such as an edge portion of an image, the deviation Is becomes relatively large.
Since the variance value Sc also increases, the attribute value I ′ (i, j) in the output image does not become a large value. For this reason, in the output image, only a portion corresponding to a bright point has a large attribute value, and a bright point can be easily detected by performing a binarization process or the like in a subsequent stage. For example, FIG.
This is an output image stored in the output image memory 7 obtained by repeating the above processing for each pixel of the input image shown in FIG. The numbers shown in this output image are the attribute values of each pixel, as in the case of FIG. In FIG. 3B, which is the output image, it is understood that only the pixel at and near the center T of the bright spot has a large attribute value and the peripheral attribute value is small. As described above, according to the image processing apparatus according to the embodiment of the present invention, it is possible to create an output image in which only the pixels near the luminescent spot are extracted, and perform the luminescent processing by performing a binarization process or the like in the subsequent stage. Points can be easily extracted. Therefore, when this is used for a cargo handling carriage, accurate position and orientation measurement can be performed.

[0009]

In the above embodiment, the average value and the variance are calculated for the same pixel group. However, at least two different pixel groups are selected for a certain target pixel by the pixel group selection means 2 and the average value calculation means is selected. 3 and the variance calculation by the variance calculator 4 may be performed on different pixel groups. The minimum number of pixels required to maintain a certain detection accuracy is
The average value and the variance value are different. Accordingly, the number of pixels to be selected can be reduced as a whole by performing calculations on different pixel groups, so that the subsequent attribute value calculation processing can be reduced. For example, a pixel group shown in FIG. 4 that shares some pixels may be selected for the pixel group shown in FIG. Such an image processing device is also an example of the image processing device in the present invention. In the above embodiment, the case where the attribute value is luminance has been described. However, the present invention is also applied to an image in which another physical quantity or information is represented as an attribute value, and a pixel whose value locally protrudes is determined. It is also possible to extract. Such an image processing device is also an example of the image processing device in the present invention. Further, in the above embodiment, the extraction processing is performed by calculating the average value and the variance value for the pixel group. However, the present invention is not limited to this, and instead of the average value, a representative value such as a median or the maximum value is used instead of the variance value. The degree of variation such as the difference between the value and the minimum value may be calculated. Such an image processing device is also an example of the image processing device in the present invention.

[0010]

As described above, any one of claims 1 to 9 can be used.
According to the image processing device described in the section, by comparing the attribute value of the pixel group surrounding the target pixel with the target pixel, an output image that emphasizes a pixel having a locally prominent attribute value such as a bright spot is created. It is possible to easily detect a bright spot by performing a binarization process or the like in a later stage. Therefore,
When the image processing device is used for a carrier, favorable position and orientation measurement can be performed.

[Brief description of the drawings]

FIG. 1 is a diagram showing a schematic configuration of an image processing apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating an example of an input image.

FIG. 3 is a diagram showing an example of an input image and an output image obtained by extracting bright points from the input image.

FIG. 4 is a diagram illustrating a pixel group.

FIG. 5 is a view for explaining a cargo handling carriage.

FIG. 6 is a view for explaining a conventional bright spot extraction method.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Image input means 2 ... Pixel selection means 3 ... Average value calculation means (representative value calculation means) 4 ... Dispersion value calculation means (variation degree calculation means) 5 ... Deviation calculation means 6 ... Attribute value change means

Claims (9)

[Claims] 1. An image input unit to which an image composed of a plurality of pixels having a predetermined attribute value is input, and a pixel for selecting a predetermined pixel group surrounding a certain target pixel in the image input by the image input unit Group selecting means, representative value calculating means for calculating a representative value of the attribute value of the pixel group selected by the pixel group selecting means, and dispersion degree calculating means for calculating the degree of dispersion of the attribute value of the pixel group; A deviation calculating means for calculating a deviation between the representative value calculated by the representative value calculating means and the attribute value of the target pixel; and a deviation calculated by the deviation calculating means and calculated by the variation degree calculating means. An image processing apparatus comprising: an attribute value changing unit that changes an attribute value of the target pixel based on the degree of variation. 2. The method according to claim 1, wherein the attribute value is changed by the attribute value changing means for each pixel included in a predetermined area of the image.
The image processing apparatus according to claim 1, wherein an output image in which pixels in which the attribute values protrude locally are extracted is created. 3. The image processing apparatus according to claim 1, wherein the pixel group selected by the pixel group selecting means is an annular pixel group separated by a predetermined distance from the target pixel. 4. The image processing apparatus according to claim 1, wherein the attribute value is a luminance value of each pixel. 5. The image processing apparatus according to claim 4, wherein the representative value is an average value of luminance values of each pixel included in the pixel group. 6. The image processing apparatus according to claim 4, wherein the degree of variation is a variance value of a luminance value of each pixel included in the pixel group. 7. The pixel group selecting means selects at least two different pixel groups for a certain target pixel, and the calculation of the representative value by the representative value calculating means and the calculation of the variation degree by the variation degree calculating means are performed. 7. The image processing apparatus according to claim 1, wherein the image processing is performed on different pixel groups. 8. The method according to claim 1, wherein the attribute value changing means is proportional to the deviation calculated by the deviation calculating means and inversely proportional to a value obtained by adding a predetermined value to the variation calculated by the variation calculating means. The image processing apparatus according to claim 1, wherein an attribute value of a pixel is changed. 9. The image input by the image input means is obtained by capturing an image of scattered light of a laser beam applied to an obstacle by the carrier to identify the obstacle.
9. The image processing device according to any one of items 8.