JPH04239982A - Gradation picture processor - Google Patents

Abstract

PURPOSE:To facilitate a measurement condition setting operation and to reasonably obtain the density discrimination reference of a background area even if the brightness of the background picture area of a gradation picture is not uniform. CONSTITUTION:A picture signal from a camera 2 is inputted to a picture processor 3 and a picture input part 5 A/D-converts the picture signal and stores it in a picture memory 6. CPU 12 segments the gradation picture inside the specified measurement area for obtaining characteristic quantity by a binarization threshold, calculates a density average and a dispersion value about the picture area corresponding to the background of the gradation picture and obtains the density discrimination reference of the background area. CPU 12 obtains the characteristic quantity of the gradation picture while it discriminates whether the respective constitution picture elements of the gradation picture inside the measurement area are the picture area of the background or the picture area of an object based on the density discrimination reference.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a gradation image processing apparatus for processing a gradation image obtained by imaging an object within a predetermined measurement area to obtain feature quantities such as area, center of gravity, principal axis angle, etc.

0002

2. Description of the Related Art Conventionally, when determining feature amounts from a grayscale image, the following method has been used to eliminate errors caused by the influence of the background area of the grayscale image. First, a density measurement area is set within the image area corresponding to the background for a grayscale image, and the density average is determined for the image within that density measurement area.Then, a certain offset value is added to the density average to determine the density of the background area. Calculate the reference level. The offset value here is a positive value when the image area of the object is brighter than the background image area, and a negative value when it is darker.

Next, for each pixel constituting the grayscale image, it is determined whether the pixel constitutes the background image area or the image area of the object based on the density discrimination reference level, and if the pixel constitutes the image area of the object. If a pixel constitutes an image area of , it is treated as a target of the feature value calculation operation, and if it is a pixel that constitutes a background image area, it is excluded from the feature value calculation operation.

0004

[Problems to be Solved by the Invention] However, in the case of this type of method, it is necessary to set a density measurement area separately from a measurement area for feature quantities, which makes the measurement condition setting operation complicated. Furthermore, when the brightness of the background image area is not uniform, the offset value will differ depending on where the density measurement area is set, making it difficult to determine an appropriate discrimination reference level and reducing measurement accuracy. There is a problem.

[0005] The present invention was made in view of the above-mentioned problem, and the measurement condition setting operation is simple, and even if the brightness of the background image area is not uniform, the density determination standard of the background area can be used. An object of the present invention is to provide a gradation image processing device that can appropriately determine the .

[0006]

[Means for Solving the Problems] The grayscale image processing device of the present invention processes a grayscale image obtained by imaging an object within a predetermined measurement region to obtain feature quantities. Means for calculating a binarization threshold for binarizing a grayscale image, and calculating a density average and a variance value for an image area corresponding to the background of the grayscale image divided by the binarization threshold. means for determining a density discrimination criterion for a background region; and means for determining a background image region and an object image region based on the density discrimination criterion for a gradient image in the measurement region to obtain a feature amount of the gradient image. Equipped with

[0007]

[Operation] Divide the grayscale image within a predetermined measurement area to obtain the feature amount using a binarization threshold, calculate the density average and variance value for the image area corresponding to the background of the grayscale image, and calculate the density of the background area. Since a discrimination criterion is determined, there is no need to set a density measurement area separately from a measurement area for feature quantities, and the measurement condition setting operation is simple. Furthermore, even if the brightness of the background image area is not uniform, it is possible to determine an appropriate density determination standard for the background area, thereby improving measurement accuracy.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows an example of the configuration of a gray-scale image processing apparatus 1 according to an embodiment of the present invention, which includes a camera 2 and an image processing apparatus 3. As shown in FIG. The camera 2 is positioned facing downward and images an object 4 located within its observation field from above.

FIG. 2 shows an example of the circuit configuration of the grayscale image processing apparatus 1. As shown in FIG. The image input section 5 inputs an analog grayscale image signal from the camera 2, performs A/D conversion, and outputs a digital grayscale image signal to the image memory 6. The image memory 6 stores the input grayscale image in units of one pixel.

Character memory 10 stores font data for characters displayed on video monitor 8. The image output section 7 converts the image signals output from the image memory 6 and the character memory 10 into analog signals, and sends the image signals for one screen to the video monitor 8. The video monitor 8 displays input images, calculation results, etc. using image signals from the image output section 7.

The image memory 6 and the character memory 10 are connected to an address/data bus 11, and the address/data bus 11 is
The data bus 11 includes a CPU 12, a ROM 13, and a RAM 1.
4. The I/O port 15 is connected to constitute a microcomputer. In this microcomputer,
The CPU 12 decodes and executes programs stored in the ROM 13, reads and writes data to and from the RAM 14, and performs various image processing such as measuring feature amounts of grayscale images.

The timing control section 9 works in conjunction with the CPU 12 to output timing signals for controlling the input and output of data in the image input section 5, image memory 6, image output section 7, and character memory 10.

FIG. 3 shows a grayscale image 16 stored in the image memory 6. In the figure, the shaded area represents an image area 16A of the object, and the other area represents an image area 16B of the background. . In addition, 17 has an X coordinate value of X1 to X2.
This is the measurement area of the feature amount defined by the Y coordinate value in the range from Y1 to Y2.

FIG. 4 shows a grayscale image 1 within the measurement area 17.
6 shows a density histogram HT for 6, where the horizontal axis is the density level and the vertical axis is the number of pixels. Each constituent pixel of the grayscale image 16 has a density level of a predetermined gradation, and the density histogram HT indicates how many pixels exist for each density level of each gradation.

In the illustrated example density histogram HT,
One peak 18 is placed in the image area of the object, and the other peak 19
correspond to the background image area, respectively. In the case of this embodiment, in order to obtain the feature amount of the grayscale image 16 for the measurement area 17, first, the density histogram HT is obtained,
Next, after calculating a binarization threshold TH for binarizing the gradation image in the measurement area 17 using Otsu's threshold method, the gradation image is divided by this binarization threshold TH. The density average a and the variance value σ are determined for the image area corresponding to the background, and the density discrimination standard for the background area is determined. Then, the background image area is determined for the grayscale image in the measurement area 17 based on the density discrimination standard. The feature amount of the grayscale image is determined while determining the image area of the object and the image area of the object.

FIG. 5 shows a grayscale image 1 produced by the CPU 12.
6 shows the procedure for calculating the feature amount. In the illustrated example, the gray center of gravity of the grayscale image 16 is calculated as the feature amount, but the present invention is not limited to this, and it is also possible to calculate an area or a principal axis angle as the feature amount. In addition, in the illustrated example, measurement area 1
Although Otsu's threshold method is used to binarize the grayscale image in 7, other methods such as the mode method, differential histogram method, and discriminant analysis method can also be used.

In step 1 in FIG. 5 (indicated by "ST1" in the figure), a grayscale image of the object 4 captured by the camera 2 is taken into the image input section 5 and converted into a digital grayscale image signal. The image is stored in the image memory 6.

Next, the CPU 12 selects a predetermined measurement area 17 including the object image area 16A from the grayscale image 16.
is set, a density histogram HT is generated for the image within this measurement area 17, and is stored in the RAM 14 (step 2). In the next step 3, the CPU 12 uses the density histogram HT to determine a binarization threshold value T for binarizing the grayscale image within the measurement area 17 using Otsu's threshold method.
Calculate H.

In the subsequent step 4, the CPU 12 calculates an image area corresponding to the background of the grayscale image divided by the binarization threshold TH, that is, in this embodiment, for each pixel having a density smaller than the binarization threshold TH. Calculate the concentration average a and the variance value σ.

Now, if the density level of each component pixel of the grayscale image 16 in the measurement area 17 is i (where i=0 to 255), and the number of pixels having each density level is Hi, then the density average a and the variance are The value σ can be found from the following equations (1) and (2). In the equation (2), b is the density determination reference level of the background area, which is given by the following equation (2).

[0021]

[Math 1]

[0022]

[Math 2]

[0023]

[Math 3]

Next, in step 5, the CPU 12 determines whether each image is a background image area 16B or an object image area 16A based on the density discrimination reference level b for the grayscale images 16 in the measurement area 17. The density gravity center G (XG, YG) of the density image is determined.

[0025] The following equations (1) and (2) are expressed as
This is a formula for calculating the coordinate value XG and the Y coordinate value YG, in which f(X, Y) indicates the density data of the pixel at the position of the coordinate (X, Y). g(X,Y) is concentration data f(X,Y)
and the concentration discrimination reference level b, and this function g(X, Y) is "0" when f(X, Y)≧b, and when f(X, Y)<b When , it becomes "1".

[0026]

[Math 4]

[0027]

[Math 5]

[0028]

Effects of the Invention As described above, the present invention divides a grayscale image within a predetermined measurement area to obtain feature quantities using a binarization threshold, and calculates density average and variance values for the image area corresponding to the background of the grayscale image. and calculates a density discrimination standard for the background region, and determines the feature amount of the grayscale image by determining the background image region and the object image region based on the density discrimination criterion for the grayscale image in the measurement region. Because of this, there is no need to set a density measurement area separately from a measurement area of feature quantities, and the measurement condition setting operation becomes easy. In addition, even when the brightness of the background image area is not uniform, an appropriate discrimination reference level can be determined, and measurement accuracy is improved, achieving the remarkable effects of achieving the purpose of the invention.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a grayscale image processing device according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a circuit configuration of an image processing device.

FIG. 3 is an explanatory diagram showing a grayscale image stored in an image memory.

FIG. 4 is an explanatory diagram showing a density histogram for a grayscale image within a measurement area.

FIG. 5 is a flowchart showing a procedure for calculating feature amounts of a grayscale image.

[Explanation of symbols]

1. Grayscale image processing device 2. Camera 3 Image processing device 5 Image input section 6 Image memory 12 CPU

Claims (1)

[Claims] Claims: 1. A grayscale image processing device that processes a grayscale image obtained by imaging a target object within a predetermined measurement region to obtain a feature amount, wherein a grayscale image processing device for binarizing the grayscale image within the measurement region is provided. means for calculating a valorization threshold; and means for calculating a density average and a variance value for an image region corresponding to the background of the grayscale image divided by the binarization threshold to obtain a density discrimination criterion for the background region. , means for determining a background image area and an image area of an object based on the density determination standard for the grayscale image in the measurement area, and determining a feature amount of the grayscale image.