JPH0389141A - Binary-coded threshold determination for floc measuring apparatus - Google Patents

Abstract

PURPOSE:To enable recognition of all required flocs by a handy treatment by defining a value (integer) weighted by a value of an logarithmic average brightness as threshold of a floc image within a range of a specified formula. CONSTITUTION:When a floc image is binary coded with a floc measuring device with a lighting of an underwater camera detector section made in a parallel opposition type, a binary coded threshold is determined from a density histogram of the floc image. In other words, a logarithmic mean mu and a logarithmic standard deviation delta are calculated from a brightness (density) histogram of the floc image. A threshold S allowed for a characteristic coefficient (a) of a camera is within a range of a formula and hence, a value weighted thereon by a value of the logarithmic average brightness is defined as threshold of the floc image. Thus, even when an area and the number of flocs are measured continuously even for the flocs different in state of formation, required flocs can be all recognized by a handy treatment.

Description

[Detailed description of the invention] A. Industrial application field
The present invention relates to a method for determining a digitization threshold, and particularly relates to a method for determining a digitization threshold for a floc measurement device in which the underwater cameras of the detection unit are illuminated in parallel and opposite manner.

B0 Summary of the Invention The present invention provides a method for determining a binarization threshold value when a floc image is binarized using a floc measuring device in which the illumination of underwater cameras in the detection unit is arranged in a parallel opposing manner. By determining the binarization threshold value using the formula, it is possible to determine the binarization threshold value that takes into account differences in cameras, and it is also possible to continuously measure the area and number of flocs even for flocs with different formation states. , provides a technology that recognizes all required flocs with simple processing.

C9 Conventional technology The floc measuring device, as shown in the general configuration in FIG. 5, includes an underwater camera detection section 52 that detects flocs floating in a floc formation pond 5I, and an image processing device 53 that recognizes the flocs. and a host computer 54 that stores data for the image processing device. The detection unit 52 and the image processing device 53 are in a close relationship, and the illumination method is particularly important for the underwater camera detection unit 52. For example, if the brightness balance of the image to be measured is not uniform, 2
Value processing cannot be performed successfully. Therefore, complex preprocessing such as density processing and filter processing must be performed.

As for the binarization processing method, a wide variety of methods have been devised in the past, and in particular, the binarization method that aims to recognize filter images etc. scattered on the screen field with most of the screen as the background is as follows. There is something.

FIG. 6 is an explanatory diagram showing a brightness histogram of the entire image. When the monitor screen 61 as shown in FIG. 6(a) is dotted with flocks 62 having a luminance of black or higher, the screen luminance histogram becomes as shown in FIG. 6(b). The same figure (b
), the vertical axis indicates the pixel number ratio (%), and the horizontal axis indicates the luminance level obtained by A/D converting the video signal, divided into 256 gradations. In this luminance histogram, it can be seen that the shaded area occupies 80 to 90% of the whole.

From this distribution, we can define the image as “0”, which corresponds to black or white on the screen.
To determine the threshold for converting into a binary value of "or "1" - within the membrane, the cluster method.

There are various methods such as the P-tile method, the differential histogram method, and a combination method of the cluster method and the differential histogram method.

The cluster method is a method in which when two peaks appear in the brightness histogram, the valley between them is used as a threshold.If the distribution of the two peaks is a normal distribution, the threshold can be determined using this method.
In flock images, mountains rarely appear symmetrically;
This method is not effective for images in which the background occupies an overwhelming portion, such as flock images.

P-tile method (above, this method is applied when the ratio of the binarized image to the whole image is known in advance, but it is not effective for images where the ratio constantly changes, such as flock images) .

As shown in FIG. 7, the differential histogram method (including the Laplacian histogram method using second-order differentiation) or the combination method of the cluster method and the differential histogram method, as shown in FIG. The brightness change 72 corresponding to the center of the image has a large absolute value of the second derivative;
The brightness change 73 corresponding to the periphery takes advantage of the fact that the absolute value of the first-order differential is large, and although the brightness change in the peripheral part of the image is small, it can be regarded as the binarization threshold ° for the brightness change in the background. It is considered to be suitable for images with overwhelmingly large background parts, such as flock images. However, this method has the disadvantage that weak noise is emphasized and even things other than flocks are recognized.

As described above, a wide variety of binarization processing methods have been devised, but they all have advantages and disadvantages, and the two most suitable for flock recognition are
A value conversion method was long awaited.

Problems to be solved by the D0 invention In response to such demands, two
A method of calculating a threshold value as a pre-processing of the digitization process and binarizing the flock image based on that value is proposed by the applicant in Utility Model Application No. 80165/1983, in which a camera using parallel opposed illuminators is proposed. Viewing range (e.g. 50 vertically + 5 horizontally yxx
This is made possible by a lighting method that does not cause uneven brightness.

FIG. 8 is a schematic explanatory diagram of a parallel-opposed illuminator;
) shows a pair of projectors 81a in front view.

81b are arranged on both sides of the screen 82 as shown in FIG. . The screen 82 in the figure is black,
The projected flock should be projected with a brightness higher than black.

Binarization processing sets a threshold at a certain luminance scale, converts images with a luminance lower than that value to "0" (black), and converts images with luminance above that value to "l" (white). do.

FIG. 9 is a graph showing one screen of a flock image, in which the horizontal axis shows brightness (density) and the vertical axis shows the number of pixels.

In the figure, (a) shows the relationship between brightness and the number of pixels, and in addition to that, figure (b) shows the relationship between brightness and the logarithmic density function f (X). Note that each summer pixel is a unit of an image, and each has brightness (density) information. The shaded area in the figure indicates the background area other than the flock, and this area always changes depending on the screen (effects of the size of the flock, density of the flock, etc.). Therefore, in order to calculate the area and number of flocs, it is necessary to calculate the optimal value for the threshold value Si indicating the limit of the shaded area according to each screen.

The present invention was created in view of these problems, and
The threshold for binarization is determined by taking into account differences in cameras, and the required flocs can be easily calculated even when the area and number of flocs are continuously measured, even for flocs with different formation states. It is an object of the present invention to provide a method for determining a binarization threshold value for a floc measuring device that can recognize all types of flocs.

80 Means for Solving the Problems Means for solving the above problems in the present invention is a floc measuring device in which the illumination method of the underwater camera detection section is a parallel-opposed type, and the floc image is binarized using a predetermined threshold value. 2 (In the iI conversion threshold determination method, a binary chemical threshold determination method in which a value compensated by the value of logarithmic average luminance within a required range is used as the threshold of the flock image, and each concentration (luminance) The formula for calculating the logarithmic mean value μ and logarithm standard deviation σ from cr/LO
G(n)), and the formula for calculating the standard error Gs between the calculated threshold Si and the optimal threshold TSi is Gs= (S i -TS i) +'rs i, and the values of A and H are expressed as shown in Fig. 4. It is preferable that the formula for calculating the optimal threshold value TSi is TS i =S i + (1-(AXKw+B)).

11 Effects The present invention determines a binarization threshold value using a predetermined formula from the density histogram of the floc image when converting the floc image into a 21 scale using a floc measuring device in which the illumination of the underwater camera detection unit is of a parallel and opposing type. be.

As shown in Figure 9(a), the image of the flocks produced by the parallel facing illumination method is biased towards low brightness, and when the logarithmic average value and logarithmic standard deviation are calculated and the logarithmic density function is drawn, the image shown in the figure (
As shown in b), if the coefficient considering the characteristics of the camera that captures the flock image is a, then from the relationship between the logarithmic mean value μ, logarithm standard deviation σ, and threshold value S, μ+σ×a≦S≦μ+σx( It was found that the threshold value S falls within the range of a+0.2). If a value (integer) weighted by the value of the logarithmic average brightness within this range is used as the threshold for the flock image, the flock image can be binarized using the threshold.

G. Embodiments Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a process diagram showing an example of the binarization threshold determination method of the present invention. This embodiment is carried out using the floc measuring device shown in FIG. 5, and the steps are first to image the flocs using the underwater camera detection section 52 and input one screen of image data. Next, the image processing device 53 creates a density histogram based on the image data. The histogram is input to the host computer 54, and the following calculations are performed.

(1) Luminance integrated sum of one screen of flock images (K w
) calculation, (2) calculation of the calculation threshold (Si), (3) calculation of the optimal threshold ('l5i) using the optimal brightness correlation formula.

When each of the above calculations is completed, the host computer 54
A command to perform binarization processing is issued to the image processing device 53 using the optimum threshold value (TSi), and the binarization processing is performed.

FIG. 2 is a graph showing a density histogram of an image in which a group of flocks is illuminated by the parallel opposed illuminators shown in FIG. In the figure, the horizontal axis shows the luminance of the flock image, and the vertical axis shows the number of pixels. First, the logarithmic average value μ and logarithmic standard deviation σ are calculated from the number of images T: of each luminance X (density; l to 255) @, and the values are substituted into the known logarithmic density function formula % formula %).

On the other hand, the formula stated on the action side %formula%) By substituting values into and overlapping them, the result will be as shown in Fig. 3.

The threshold range of the flock image is calculated using this formula, and within that range, weighting is performed using the logarithmic average brightness using the following formula.

S(μ)(((μ +a ・+7 +0.2・(7/
LOG (128)) - (1) The threshold value S (μ) obtained as a result is set as the threshold value St of the target flock image.

By performing the binarization process using this threshold ftfisr, shading within the screen field of view caused by the parallel opposed illuminators is eliminated, and the following effects are produced.

(1) All flocs scattered on the screen can be recognized (however, since it is related to pixel resolution,
(for flocs of m or more).

(2) A logarithmic average value and a logarithmic standard deviation can be calculated from the brightness (density) histogram value, and a binarization threshold value can be obtained using the above equation (1).

(3) Therefore, it is possible to obtain the optimum threshold value from each brightness (density) histogram even for flock images having different formation states.

(4) Further, according to (3) above, even when continuously measuring the area and number of flocs, continuous measurement is possible by automatically calculating the optimal threshold value using a computer or the like.

(5) There is no need for density conversion or filter processing during binarization, which simplifies the processing.

Incidentally, in the above processing, the equation includes a term a for taking into account the characteristics of the camera. This term a requires adjustment of its value when the camera is replaced, which can be troublesome. Therefore, it is possible to consider a method of fixing this term a=0.9 to eliminate the need for adjustment to the camera.

Let Si be the calculation threshold calculated by the above formula (1),
Let TSi be the optimal threshold actually measured in Fig. 2, and let Gs be its standard error, then Gs==(Si-TSi)+TSi ・・・・・・(
2), and furthermore, the sum of the products of the values of the horizontal axis and the corresponding vertical axis in FIG.
When w is taken, the error Gs is plotted on the vertical axis, and the sum Kw is plotted on the horizontal axis, the characteristics shown in FIG. 4 are obtained, and it is found that a very good correlation exists.

Therefore, by substituting the calculation threshold Si obtained in the above equation (1) into this linear equation (2), TS i =S i -i-(1-(AxKw+B))
...(a), and the optimal threshold value is calculated. Furthermore, with this method, there is no need to adjust for differences in cameras, and even if the camera is replaced, the optimum threshold value can be calculated without changing the coefficients.

I (, As described in detail, according to the present invention, the binarization threshold value is determined in consideration of the difference in cameras, and the area and number of flocs are continuously calculated even for flocs with different formation states. Even when measuring, floc measurement allows you to remove and recognize the required flocs with simple processing.

A method for determining a binarization threshold for an apparatus can be provided.

[Brief explanation of drawings]

Figure 1 is a process diagram of an embodiment of the present invention, Figures 2 and 3 are luminance histograms of the embodiment, Figure 4 is a graph showing the relationship between standard error and luminance sum, and Figure 5 is a flock measuring device. Fig. 6 is an explanatory diagram of the brightness histogram, Fig. 7 is an explanatory diagram of flock luminance changes, Fig. 8 is an explanatory diagram of parallel opposed illuminators, and Fig. 9 is a graph of the relationship between luminance and number of pixels. It is. 51... Floc formation pond, 52... Detection section, 53.
...Image processing device, 54...Host computer, 8
I. Floodlight, 82...Screen, 83...Optical fiber, 84...Light, 85...Camera. Fig. 1 Process diagram of one implementation of the present invention Fig. 2 V1 wave histogram Fig. 3 Military histogram brightness (density) Ishitatsu Genji 7 Graph Y-
A x 10 B Fig. 7 Flow 1 Tsuk overturning & changing the direction of the date and moon prisoner Quantification direction Fig. 8 Parallel force-directed type torture device 5's footwork (Q) Teruhama (dark 5) (b) hti positive Supplementary document on content procedures (penalty January 9, 1990)

Claims (3)

[Claims] (1) The illumination method of the underwater camera detection section is a parallel facing type,
In a method for determining a binarization threshold of a floc measurement device that performs binarization processing of a floc image using a predetermined threshold, the logarithmic mean value (μ) is determined from the luminance (density) histogram of the floc image.
, the logarithmic standard deviation (δ) is calculated, and the threshold value (s) considering the characteristic coefficient (a) of the photographing camera is within the range of μ+δ×a≦S≦μ+δ×(a+0.2), and the logarithmic average luminance is 1. A method for determining binarization of a flock measuring device, characterized in that a value weighted by a value is used as a threshold value of a flock image. (2) The lighting method of the underwater camera detection section is a parallel facing type,
In a method for determining a binarization threshold of a floc measurement device that performs binarization processing of a floc image using a predetermined threshold, the logarithmic mean value (μ) is determined from the luminance (density) histogram of the floc image.
, the logarithmic standard deviation (δ) is calculated, and when a is a coefficient that takes into account the characteristics of the camera that captures the flock image, the threshold value Si is Si(μ)∝{μ+a・δ+0.2・δ/lo
1. A method for determining a binarization threshold for a floc measuring device, characterized in that the binarization threshold for a floc image is determined by the formula: g(128)}. (3) In the method for determining a binarization threshold value for a flock measuring device according to claim (2), values A and B determined from the relationship between the luminance integrated sum (Kw) of one screen of flock images, the standard error, and the luminance sum. The optimal threshold TSi is calculated as TSi=Si÷{1-(
A x Kw + B)} A method for determining a binarization threshold for a floc measuring device, characterized in that the threshold is calculated by the following formula.