JPH0636187B2 - Image processing device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus, and more particularly to an image processing apparatus for determining the length of a filamentous shape from a picked-up image of a filamentous object such as a filamentous microorganism.

[Background of the Invention]

At the sewage treatment plant, air is blown into the inflow water (aeration) in the air tank to remove the organic matter in the inflow water by allowing the microorganisms to ingest the organic matter. It has been released. Therefore, it is necessary to ingest organic matter and maintain microorganisms with good sedimentation. These microorganisms are roughly classified into aggregating microorganisms and filamentous microorganisms. Of these, if the filamentous microorganisms reproduce too much (called the bulking phenomenon), the sedimentation property deteriorates. It is important not to propagate filamentous microorganisms, because if the sedimentability deteriorates, the microorganisms will flow out from the sedimentation pond. Examples of filamentous microorganisms include Sphaerotilus. As described above, when filamentous microorganisms propagate in the sewage treatment process, the quality of treated water is deteriorated, and such microorganisms flow out.

In order to prevent these, it is important to measure and monitor the amount of filamentous microorganisms. At present, the length, that is, the amount of filamentous microorganisms is measured by a manual operation by an observer. For example, in the sewage treatment process, the length of the filamentous microorganisms on the photograph is measured by a kilvimeter or by microscopic observation. However, since the conventional measuring method is a manual operation by a monitor, even if a trained monitor performs the measurement, it takes several hours for one measurement. For this reason, continuous monitoring and early detection cannot be performed, and there is a drawback that the reproduction state of filamentous microorganisms cannot be well managed.

In order to solve this shortcoming, the applicant has previously filed Japanese Patent Application No. 58-1
No. 40422 (JP-A-60-31889) proposes a method for measuring the length of filamentous microorganisms. This measuring method detects an enlarged image of microorganisms with an industrial television camera (ITV),
It recognizes filamentous microorganisms by applying image processing technology.

Specifically, the image of the filamentous microorganism is enlarged, the pixels corresponding to the filamentous microorganism are extracted by focusing on the brightness level, and the number of the pixels is integrated to convert the length of the filamentous microorganism. . However, since only the brightness level of filamentous microorganisms is focused on in this method, protozoa or small aggregating microorganisms having the same luminance as filamentous microorganisms may be mistakenly recognized as filamentous microorganisms. Therefore, even if the pixels of the filamentous microorganisms are integrated, it will be converted into a length longer than the actual length of the filamentous microorganisms.

[Object of the Invention]

An object of the present invention is to accurately calculate, display or judge the length of a filamentous object by shaping an imaged image of the filamentous object, for example, a filamentous microorganism.

[Outline of Invention]

In order to achieve the above object, the present invention provides a binarization processing unit that binarizes an image converted into luminance information by an imaging device; the obtained binarized image is thinned to have a thickness of all 1s. Thinning processing means for extracting a line-shaped image of pixels; reduction / expansion processing means for reducing and expanding the obtained binarized image to extract a non-linear image; Means for erasing a portion overlapping with the non-linear image;
Provided is an image processing device including: a noise removing unit that erases an image included in a region of × 3 pixels as noise; and a unit that calculates the length of the obtained linear image.

Example of Invention

 Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows an embodiment in which the present invention is applied to a sewage treatment process.

In FIG. 1, the supernatant liquid (sewage) of the settling tank 110 and the returned sludge (microorganisms) flow into the aeration tank 100 from the sludge return pipe 160. On the other hand, the blower 120 supplies air through the air pipe 130 and supplies air from the air diffuser 14 to the air tank 100. The returned sludge and waste water supplied into the aeration tank 100 are mixed by stirring. The returned sludge, that is, activated sludge, is an aggregate of microorganisms and has a particle size of 0.1 mm to 1.0 mm (floc).
Although it contains several tens of kinds of microorganisms, they are roughly classified into a flocculating microorganism c and a filamentous microorganism f as shown in FIG.
Activated sludge absorbs oxygen in the supplied air and decomposes organic matter in wastewater into carbon dioxide gas and water. A part of the organic matter is applied to the bacterial growth of the activated sludge. The mixed liquid of activated sludge and wastewater is guided to the settling tank 150, where the activated sludge settles by gravity. The supernatant is usually discharged after chlorine sterilization. On the other hand, the settled sludge is returned to the aeration tank 100 as return sludge from the sludge return pipe 160.

The imaging device 200 is arranged so as to be immersed in the liquid in the air tank 100, and has a function of obtaining an enlarged image of the microorganisms in the air tank 100. Imaging control device 30
0 controls the timing (horizontal and vertical synchronization) of the imaging scan of the imaging device 200, the illumination intensity, and the like, and at the same time, performs the image processing of the enlarged image (grayscale image) of the mixed liquid (including water and microorganisms) obtained by the imaging device 200. To the device 400. The image processing device 400 image-processes the grayscale image, recognizes filamentous microorganisms from the image, and measures the length of the microorganisms. The keyboard 900 inputs numerical values such as image processing parameters used in the image processing apparatus 400. The console display 910 numerically displays the numerical value input by the keyboard 900 and the length of the filamentous microorganism. The monitor television 920 displays an image of the image pickup apparatus 200, an image processing process, and an image processing result.

FIG. 3 shows a detailed configuration of the image pickup apparatus 200.

In FIG. 3, the mixed liquid containing activated sludge and sewage filled in the sample chamber 210 is irradiated with light from the illumination device 230 via the optical image magnifying device 240. The optical image magnifying device 240 includes an objective lens and an eyepiece lens. The irradiation light is irradiated into the sample chamber 210 through the observation window 245 and is reflected by the irradiation mirror 220 in the sample chamber 210. Air is blown from the air tube 270 into the sample chamber 210 so that the mixed liquid is replaced at regular intervals. The supplied air also serves to clean the sample chamber 210 and the observation window 245. The reflected light from the reflection mirror 220 passes through a magnifying lens 250 and an optical image magnifying device 240 and is guided to an industrial television camera (ITV) 260, where a grayscale image of the mixed liquid is recognized and converted into an electric signal. . ITV26
0 has vertical and horizontal image pickup devices (not shown) on the light receiving surface, and is scanned by receiving a vertical direction (vertical direction) synchronization signal VD and a horizontal direction (horizontal direction) synchronization signal HD. The synchronization signals VD and HD are given from the imaging control device 300. The image pickup device generates a signal (video signal) having a different output voltage according to the degree of brightness (luminance) received. In this way, the video signal V is output from the ITV 260. As the imaging device 200, a device that samples the mixed liquid or a device that observes the sampling liquid with a microscope and ITV can be used.

The image pickup apparatus 200 as described above is controlled by the image pickup control apparatus 300. The imaging control device 300 operates the lighting device 230 and the ITV 260, and controls the illuminance of the lighting device 230 as necessary. Further, in the imaging control device 300, the condenser 310 is connected to the sample chamber 2 via the air tube 270.
The air supply frequency and the air supply time of the air supplied to the inside 10 are controlled. Usually, air is sent once every 5 seconds to 20 minutes, and about 2 to 30 seconds once. The air supply frequency and the air supply time can be changed according to the degree of contamination of the observation window 245 and the sample chamber 210.

As described above, the imaging control device 300 is configured so that the imaging device 200
The image processing device 4 receives the analog video signal of the grayscale image sent from the ITV 260, and controls this.
Send to 00.

FIG. 4 shows a detailed configuration of the image processing apparatus 400.

In FIG. 4, the image processing apparatus 400 is an A / D converter 50.
0, an image memory 600, an image processing processor 700 and an arithmetic unit 800. Image memory 600
Is a grayscale image memory 610 and two binary memories 620, 6
It consists of thirty. The image processing processor 700 is for recognizing filamentous microorganisms from a grayscale image, and includes a histogram processing circuit 710, a binarization circuit 720, and a thinning circuit 73.
It consists of zero. The arithmetic unit 800 is an image processing processor 7.
This is for calculating the length from the filamentous microorganisms recognized at 00, and comprises a filamentous microorganism integrating circuit 810, a filamentous microorganism length computing circuit 820, an average length computing circuit 830 and a concentration computing circuit 840.

In this configuration, the A / D converter 510 A / D-converts the analog video signal V into a digital signal.
The brightness of each pixel on the screen is converted into a digital value by the A / D converter 510. The digital value is, for example, 1
28 levels. The grayscale image memory 610 stores all digital signals (grayscale image information of the mixed liquid) having the brightness corresponding to each pixel. The frequency of A / D converting the video signal V and storing the grayscale image information in the grayscale image memory 610 is the same as the frequency of air supply to the sample chamber 210. That is, after the mixed liquid is replaced and a new mixed liquid enters the sample chamber 210 and stands still, this grayscale image is stored in the grayscale image memory 610. In this way, the memory 61
The length of the filamentous microorganism is calculated by image processing based on the grayscale image information stored in 0.

FIG. 5 shows a timing chart of air supply timing, A / D conversion and storage of a grayscale image, and image processing timing. FIG. 5 shows an example in which one cycle is performed in 10 seconds. Aeration for 2 seconds, wait for 2 seconds until the image freezes, A / D
The case where it takes 0.2 seconds for conversion and storage and 2.8 seconds for image processing is shown. This cycle is repeated.

Next, the image processing operation in the image processing processor 700 will be described.

The histogram processing circuit 710 calculates a histogram of the grayscale image information stored in the grayscale image memory 610, that is, a density frequency distribution. The histogram is a characteristic diagram showing how many pixels of a predetermined density (luminance) are present. As shown in FIG. 6, the horizontal axis shows the density (luminance) and the vertical axis shows the frequency (number of pixels). To take. The grayscale image in the grayscale image memory 610 has a bright background (water) b as shown in FIG.
It contains agglutinating microorganisms c, filamentous microorganisms f, and noise n that are darker than this. The histogram of the grayscale image is the sixth
As shown in the figure, it is divided into a portion b corresponding to a bright background and a portion (c + f + n) corresponding to aggregating microorganisms, filamentous microorganisms and noise.

The binarization circuit 720 determines a binarization threshold T based on the histogram of the grayscale image obtained by the histogram processing circuit 710, and grayscale images having various brightnesses (for example, 128 gradations) are white and black. Binarize to and. Pixels brighter than the threshold T are made white (“1” level) and dark pixels are made black (“0” level). The threshold value T is set to a value lower than the peak p of the background portion b by a predetermined value, or set to the inflection point q in FIG.

An example of an image binarized by selecting the threshold value T in this manner is shown in FIG. In FIG. 7, the background b is white (“1” level), and the flocculating microorganism c, the filamentous microorganism f and the noise n are black (“0” level). The "0" level is indicated by hatching. The obtained binary image is stored in the binary memory 620.

The binary image signal in the binary memory 620 is the thinning circuit 7
Added to 30. The thinning circuit 730 is a binary memory 6
The binary image in 20 is thinned multiple times to extract filamentous microorganisms. As shown in FIG. 2, the filamentous microorganism f corresponds to the skeleton of the aggregating microorganism c and is also contained inside the aggregating microorganism c. Since the filamentous microorganism f contained in the aggregating microorganism c overlaps the aggregating microorganism c, the binarized image of FIG. 7 includes the filamentous microorganism f and the aggregating microorganism. From the binarized image of FIG. 7, the filamentous microorganism f
Perform thinning process multiple times to extract only
The thinning process is performed when the binarized image data in the binarized memory 620 is updated.

If the binarized image of FIG. 7 is thinned only once, it becomes as shown in FIG. The thinning circuit 730 performs thinning processing by a known thinning method. By the thinning process, the pixels of the filamentous microorganisms f and the flocculent microorganisms c (hatched portion in FIG. 7) are cut off by one pixel from the contour. At this time, the number of pixels is 1
It is not deleted where there are only pixels. For example, the thickness is 3
For pixels, one pixel is trimmed from both sides to make the thickness one pixel, and for two pixels, one pixel on one side is trimmed to one pixel.

By such a thinning process, the filamentous microorganism has a thickness of 1
Or, a pixel having 3 pixels has 1 pixel, and a pixel having a thickness of 4 pixels or more has 2 pixels or more. The binary image thus thinned for the first time is stored in the binary image memory 630. The binarized image stored in the binarized image memory 630 is further subjected to the thinning process a plurality of times in order to make one pixel a portion having two or more pixels in the first thinning process. By repeating this, the filamentous microorganism f becomes an image in which the thickness is all 1 pixel. At that time, the filamentous microorganisms f overlapping with the aggregating microorganisms c cannot be trimmed to more than one pixel by the thinning process, so that an image as shown in FIG. 9 is obtained. In this way, even if the filamentous microorganism has a thickness of 2 to 3 pixels, the thickness is all 1
It can be extracted as a pixel. The number of repetitions of the thinning process can be determined by the number of pixels of the diameter of the aggregating microorganism. For example, a coagulative microorganism having a diameter of 50 pixels can be thinned 25 times to 1 pixel. But,
Since the object is not thinned to 1 pixel or less in the thinning processing, there is no problem even if it is performed 25 times or more. However, in order not to repeat more than necessary, the binarized images before and after the thinning process are compared (difference) to determine whether or not they are the same, and when they are the same (images that cannot be further thinned), It is also possible to stop the conversion.

As shown by n in FIG. 9, the binarized image extracted by such a thinning process also contains small aggregating microorganisms (small in quantity) that do not contain filamentous microorganisms. Therefore, it is desirable to remove the microorganism n when calculating the length from the recognized filamentous microorganism.

Returning to FIG. 4, the operation of the arithmetic unit 800 for calculating the length from the pixel number of the filamentous microorganism will be described.

The filamentous microorganism integrating circuit 810 integrates the pixel number of the filamentous microorganisms extracted as one pixel in thickness in the grayscale image at time t. Specifically, in the binarized image obtained in FIG. 9, the number of pixels only in the hatched portion (pixels having a value of “0” level) is integrated. The integrated number of pixels is F (t). The filamentous microorganism length calculation circuit 820 calculates the filamentous microorganism length L (t) from the integrated pixel number F (t) of the filamentous microorganism by the following formula.

L (t) = A · F (t) −B ... (1) Here, A and B are coefficients, and B corresponds to n minutes of the microorganism image in FIG. 9 to be subtracted as an error. The coefficient A is 1
When the length of one side of the pixel corresponds to, for example, 10 μm (0.01 mm), it becomes 0.01. By the way, when the filamentous microorganisms are arranged horizontally or vertically as shown in FIGS. 10 (a) and (b), the coefficient A is 0.01.
As shown in (b), the values of the coefficient A are become. If the filamentous microorganisms are evenly distributed horizontally and vertically and to the right and upward to the left at an angle of 45 degrees, the number of horizontal and vertical pixels is the same as the number of pixels to the right and upward to the left, and 45 degrees diagonally. it is conceivable that. In this case, the value of A is 0.
01 and It becomes 0.012 of the average value of.

In this way, when the filamentous microorganism length L (t) is calculated for the grayscale image of one screen, as described in FIG. 5,
Perform the second sampling and calculation. That is, the imaging control device 300 controls the compressor 310 to supply air to the sample chamber 210 through the air tube 270,
Wash and replace the sample (mixed solution). Next, the grayscale image is A / D converted by the same operation as described above, stored in the grayscale image memory 610 at time t + h (where h is a sampling period), and the same processing is repeated in the image processing processor 700. The filamentous microbial length obtained by this second treatment is L (t + h). In this way, the filamentous microorganism length calculation circuit 820 sequentially lengths L (t + h), L (t + 2) of the filamentous microorganisms of each image at time t, t + h, t + 2h, ... T + Kh.
h), ... L (t + Kh) is calculated one after another at each time h.

The timing control of the imaging device 200 by the imaging control device 300 is performed in conjunction with the image processing device 400.

Upon receiving the signal from the filamentous microorganism length computing circuit 820, the average computing circuit 830 computes the average length of the filamentous microorganisms per screen. The average length calculation circuit 830 is a signal output from the filamentous microorganism length calculation circuit 820. L (t +
From h), L (t + 2h), ... L (t + Kh), the average length Lm of the filamentous microorganism per screen is calculated by the following equation.

Where (K + 1) is the average number of times, 10 to 1000
It is about once.

In response to the signal Lm from the average length calculation circuit 830, the concentration calculation circuit 840 calculates the filamentous microorganism length per unit volume (that is, filamentous microorganism concentration) Lv by the following equation.

Lv = Lm / (K + 1) / v ... (3) Here, v is the volume of the mixed liquid imaged.

Through the above processing, the filamentous microorganism length (concentration) per unit volume can be obtained.

As described above, the concentration (amount) of filamentous microorganisms is obtained. In the present invention, binarization is performed based on the histogram of the concentration image, and the binarized image is thinned and then the filamentous microorganisms Since the length is calculated, the length of the filamentous microorganism contained in the aggregating microorganism can also be accurately measured.

FIG. 12 shows another embodiment of the present invention of the image processing apparatus.

The embodiment of FIG. 12 is configured by adding a line segment direction code processing circuit 780 to the embodiment of FIG. 12th
In the embodiment shown in the figure, the length is accurately obtained in consideration of the connecting direction of the filamentous microorganism pixels.

In FIG. 12, the line segment direction code processing circuit 780 receives the signal from the image memory 630 and calculates the line segment direction code of the pixel of the filamentous microorganism. The line direction code is the 13th
When considering a 3 × 3 pixel as shown in FIG. 6A, the first direction is to determine which direction the pixel of interest is connected from the central 9 pixels.
The code numbers are shown in FIG. 3 (b). In order to obtain the line segment direction code, it is sequentially examined in which direction the filamentous microorganisms are connected, and it is counted which code number of the line segment direction code each pixel is.

As a result, the total number of pixels corresponding to each code number can be counted. The filamentous microorganism integrating circuit 810 receives the signal from the line direction code processing circuit 780 and calculates the pixel number F ₁ (t) with an odd code number and the pixel number F ₂ (t) with an even number. Upon receiving the signals of F ₁ (t) and F ₂ (t), the filamentous microorganism length computing circuit 820 computes the filamentous microorganism length L (t). When the line segment direction code is odd, the value of the coefficient A of F ₁ (t) is set to 0.01, while when the line segment direction code is even, the value of the coefficient A of F ₂ (t) is set to 0.014. To do.
That is, in the filamentous microorganism length calculating step 820, the following calculation is performed.

L (t) = 0.01F ₁ (t) + 0.0141F ₂ (t) -B ... (4) The average length calculation circuit 830 inputs the signal L (t) and the filamentous microorganisms per screen are input. Calculate the average length Lm. Then, the arithmetic unit 800 performs the same operation as in the embodiment of FIG.
The filamentous microorganism length Lv per unit volume is determined.

By the above calculation, this length can be accurately obtained regardless of the direction in which the pixels of the filamentous microorganism are connected. That is, it is possible to realize the same operation as tracing the filamentous microorganisms with a human quilvimeter.

The filamentous microbial concentration (amount) is obtained as described above. In the embodiment of FIG. 12, the same effect as that of the embodiment of FIG. 4 can be obtained, and further considering the connecting direction of the pixels. Since the concentration of the filamentous microorganisms is obtained, the length can be accurately measured even when the filamentous microorganisms are connected in the horizontal and vertical directions and the diagonal direction.

FIG. 14 shows another embodiment of the present invention.

In the embodiment shown in FIG. 14, only the filamentous microorganisms protruding from the aggregating microorganisms are extracted.

The embodiment shown in FIG. 14 is similar to the embodiment shown in FIG.
0, an expansion processing circuit 750 and a mask processing circuit 760 are added.

First, the processing of the reduction processing circuit 740 and the expansion processing circuit 750 will be described. The reduction processing circuit 740 is 2 in one reduction processing.
Eliminate filamentous microorganisms and noise within the pixel size. Next, only the aggregating microorganisms reduced by the expansion processing circuit 750 are restored to the original size.

The reduction processing circuit 740 takes in the binarized image (shown in FIG. 7) stored in the binarized memory 620, reduces it, and stores it in the binarized memory 640. In the reduction processing, the contour of the "0" level portion indicated by hatching is deleted pixel by pixel. Unlike the thinning processing, the reduction processing simply deletes each pixel from the contour. For this reason, the filamentous microorganisms having a width of 2 pixels are deleted because the contours are deleted by 1 pixel from both sides. Filamentous microorganisms having a thickness of 2 pixels are eliminated by one reduction process, but at the same time, one pixel is deleted from the contour of a flocculating microorganism (flock).

Thus, if the binarized image shown in FIG. 7 is reduced once, the image shown in FIG. 15 is obtained. The reduced binary image is stored in the binary memory 640 again. By this reduction treatment, filamentous microorganisms can be almost eliminated. Since the outline of the aggregating microorganism is cut off pixel by pixel due to the reduction processing, this must be restored to the original size. The dilation step 750 dilates the signal of the binarization memory 640, that is, the binarized image of FIG. In the expansion process, one pixel is added around the "0" level portion shown by hatching. That is, one pixel is added to the contour of the cohesive microorganism c in the image shown in FIG. As a result, the image becomes as shown in FIG. By this operation, noise n
Although some remain, coagulable microorganisms can be extracted. The result is stored again in the binarization memory 640.
The mask processing circuit 760 projects from the aggregating microorganism based on the binary image (shown in FIG. 9) stored in the binary memory 630 and the binary image (shown in FIG. 16) stored in the binary memory 640. Extract the length of the filamentous microorganisms. Specifically, the portion of the aggregating microorganism c in FIG. 16 (the value of the “0” level shown by hatching) is masked, and other than the aggregating microorganism c shown in FIG. 9 (the white portion indicates the “1” level). Value) to extract the filamentous microorganisms. As a result of this mask processing, the first
Filamentous microorganisms f as shown in FIG. 7 are extracted and stored in the binarization memory 650. From the binary image (FIG. 17) stored in the binary memory 650, the filamentous microorganism concentration Lv
The operation for calculating is the same as the operation described in the embodiment of FIG. Although not shown in FIG. 14, it goes without saying that the line direction code processing circuit 780 described in the embodiment of FIG. 12 may be added.

As described above, the filamentous microorganism concentration Lv is calculated. In the embodiment of FIG. 14, the same effect as that of the embodiment of FIG. 4 is obtained, and in addition, as shown in FIG. Since only filamentous microorganisms that are not included in the microorganisms are extracted, only filamentous microorganisms that have been conventionally subjected to visual measurement can be extracted.

FIG. 18 shows another embodiment of the present invention.

The embodiment shown in FIG. 18 is an embodiment in which the filamentous microorganism length is obtained with higher accuracy. Fig. 18 shows 14
It is configured by adding a noise removal circuit 770 of the embodiment shown. The noise removal circuit 770 erases the noise n from the binarized image (shown in the seventeenth diagram) stored in the binarized memory 650. In this noise n, a single microorganism or a small aggregating microorganism remains as a line in the thinning processing step 730.

The noise removing circuit 770 removes noise as follows. Noise n due to thinning is shown in Fig. 19 (a) ~ (g)
Take the pattern shown in. These are included in an area of 3 × 3 pixels. Those contained in 4x4 pixels are extracted as coagulant microorganisms and are not counted in filamentous microorganisms.
You do not have to consider it as noise. (a) is noise of one pixel h, which is a binarized image of the binarized memory 650 (the 17th pixel).
This is removed by extracting the independent 1 pixel h existing in the figure). A known technique can be applied to the method for removing one independent point.

Next, the patterns shown in (b) to (g) are h ₁ , h ₂ , h ₃ ,
Clearing the h ₄ and h ₅ as the end point (in chopsticks pixels), leaving a pixel h center adjacent to the end point. Center pixel h
Is an independent point, and is removed as noise.

FIG. 20 shows a result of removing the noise n by performing the above processing in the noise removing circuit 770. Fig. 17 shows an example with little noise, but in reality, the mixed liquid of activated sludge contains many minute dusts and microorganisms, so removing this as noise reduces the error. It is effective in improving accuracy.

The arithmetic unit 800 receives the binary image (FIG. 20) of the binary memory 650 and receives the length L of the filamentous microorganism per unit volume L.
The method of calculating v is the same as in the embodiment of FIG.

The embodiment shown in FIG. 18 has the same effect as that of FIG. 4, but since noises other than filamentous microorganisms are removed,
Only filamentous microorganisms can be accurately extracted.

Needless to say, the noise removal circuit 780 can be applied to the embodiment of FIGS. 12 and 14 to accurately calculate the concentration of filamentous microorganisms.

〔The invention's effect〕

According to the present invention, required image information and image information that is not required can be discriminated from each other, and the length of the required thread-shaped information can be selectively obtained easily.

In addition, although the above-mentioned example explained the example applied to the sewage treatment process, it goes without saying that the present invention can be applied to the measurement of the concentration of filamentous microorganisms in other bioprocesses for culturing the filamentous shape.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a diagram showing a grayscale image, FIG. 7, FIG. 8, FIG. 9, FIG. 15, FIG. 16, FIG. Figure 20 shows a binary image,
FIG. 3 is a diagram for explaining the image pickup apparatus 200 in detail, FIG.
12, 14, and 18 are views for explaining the configuration of the image processing apparatus of the present invention, FIG. 5, FIG. 6, FIG.
FIG. 11, FIG. 13 and FIG. 19 are diagrams for explaining the image processing operation of the present invention. 100 ... Air tank, 200 ... Imaging device, 300 ... Imaging control device, 400 ... Image processing device, 500 ... A / D converter, 600 ... Image memory,
700 ... Image processing processor, 800 ... Computing device.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Mikio Yoda 5-2-1 Omika-cho, Hitachi City, Ibaraki Prefecture Hitachi Ltd. Omika Plant, Hitachi Ltd. (72) Naoki Hara 5-chome, Omika-cho, Hitachi City, Ibaraki Prefecture No. 1 Incorporated company Hitachi Ltd. Omika factory (56) References JP 57-13581 (JP, A) JP 57-17085 (JP, A) Translated by Shin Nagao "Digital Image Processing" (SHO 53- 12-10) Modern Science P. 385-390

Claims (1)

[Claims] 1. A binarization processing means for binarizing an image converted into luminance information by an image pickup device, and thinning the obtained binarized image to extract a line-shaped image having a thickness of all 1 pixel. Thinning processing means for performing the reduction processing and the expansion processing for the obtained binarized image to extract the non-linear image, and the line-shaped image overlapping the non-linear image. Means for erasing the existing portion, noise removing means for erasing the image included in the 3 × 3 pixel area from the linear image from which the overlapping portion has been eliminated as noise, and the length of the obtained linear image. An image processing apparatus comprising: a calculating unit.