JPS63222262A - Water quality abnormality detector - Google Patents

Abstract

PURPOSE:To judge position is present as early as possible, by arranging a camera device to take different kinds of fish raised in a water tank partitioned, a processing means to detect the speed of fish from the resultant image and a processing means to judge position is present from the speed obtained. CONSTITUTION:A water tank 1 for supplying water to be inspected is divided with partitions 10a and 10b and different fishes 9a, 9b and 9c are raised one by one in respective areas. The water tank 1 is lighted with a lighting device 3 through a backscreen 2 to take the fishes 9a, 9b and 9c with a camera device 4 and an output signals is sent to an image processor 5, with which the fishes are recognized on the basis of image information to calculate the center of gravity of each fish. A computing unit 6 computes the speed of each fish from the center of gravity thereof to compute a distribution of center of gravity and speed of each fish. The computing unit 6 which stores distributions of center of gravity and speed of fish in normal condition compares the measured distribution with the normal distributions to judge whether the movement is fish is abnormal or not. The results of judgement is transmitted to an alarm 7.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a water quality abnormality detection device that monitors the inflow of toxic substances even in raw water of a water purification plant and inflowing sewage of sewage treatment.

[Conventional technology]

At water treatment plants, in order to monitor whether or not poisonous substances have been mixed into the raw water, a portion of the raw water is channeled into an aquarium where aquatic animals such as crucian carp, carp, dace, Japanese tanager, rainbow trout, and oysterfish are raised in this aquarium. There is. In other words, if a poisonous substance is mixed into the raw water, the inflow of the poisonous substance into the raw water can be monitored by taking advantage of the phenomenon that the fish exhibit abnormal behavior such as frenzied reversal, nose raising, or death. There is. Additionally, sewage treatment plants need to know whether legally prohibited toxic substances have entered the incoming sewage, and rely on intermittent manual analysis.

As described above, monitoring of toxic substances in water currently relies on human visual observation and analysis, which has the disadvantage that continuous monitoring and early detection are not possible, and countermeasures such as stopping water distribution to customers are delayed.

As a method for monitoring fish, a method has been proposed (Japanese Unexamined Patent Publication No. 46294/1989) in which images of water-based organisms are captured using a video device, and the movements of the organisms are compared with predicted parameters to detect abnormalities. According to this method, in the test circulation of multiple open cuts,
It is described that multiple types of organisms are exposed to water from an aquatic system and the movement of the organisms is observed.

[Problem that the invention seeks to solve]

In the water quality monitoring method described in Japanese Patent Application Laid-Open No. 61-46294, a video device is used to monitor living things.

There is no mention of a method for recognizing living things, a method for calculating motion parameters, or a specific method for determining.

When monitoring multiple fish, an aquarium and a camera are required for each fish, making the device complex. In addition, the @image direction is toward the water surface, and this method cannot detect characteristic abnormal behaviors such as surfacing and nose-down.

Furthermore, the process that shows the predictive parameter calculation method is too complex to be practical.

An object of the present invention is to provide a water quality abnormality detection device that can keep multiple types of fish in one aquarium, continuously monitor them through image recognition, and quickly determine the presence or absence of toxic substances in the water.

[Means for solving problems]

The present invention uses a single imaging device to take images of different types of fish kept in a partitioned aquarium, recognizes the position and speed of the fish from that image, and further identifies abnormalities in the water from this pattern. The main component is a means for determining and monitoring the presence or absence of

[Effect]

Several types of fish are kept in an aquarium divided into multiple areas. The imaging device converts this fish into a luminance signal at every set period and stores it in an image memory. By binarizing the data in the image memory using a set threshold value, fish are binarized and extracted. After calculating the positions of a plurality of fish from this binary image, the speed of each fish is calculated.

In this way, the movements of multiple fish are detected based on their positions and speeds.

These positions and speeds are measured for a predetermined period of time to determine whether or not the behavior is abnormal.

〔Example〕

Hereinafter, one embodiment of the present invention will be described based on the drawings.

FIG. 1 is a diagram showing an embodiment of the present invention. 1 is an aquarium for breeding fish, and water is constantly supplied.

The water to be supplied is raw water or wastewater at a water treatment plant, inflow sewage at a sewage treatment plant, or river water when monitoring toxic substances in a river.

Inside the aquarium 1 is a partition plate 10a such as a mesh plate or a perforated plate.

It is divided into three areas by 10b, and one fish 9a, 9b, and 9c are kept in each area. This area is limited only to the number of fish to be kept. Since a plurality of fish are kept, each region is also plural, and in this embodiment, three examples will be explained. The types of fish that live in the water supply are kept, and different types of fish are kept in one aquarium.

Examples of types of fish include crucian carp, carp, dace, and tanago.

These include rainbow trout and oysterfish. Lighting device 3 is aquarium 1
Illuminate the fish 9a, 9b, and 9c inside. Since the present invention applies image processing technology, uniform illumination is required. For this reason,
A translucent back screen 2 corresponding to a light scattering plate made of ground glass, white acrylic, or the like is provided between the lighting device 3 and the aquarium 1. By making the partition plates 10 and 10b the same material as the back screen 2, the influence of the partition plates on the fish school image can be reduced. The back screen 2 also has the role of recognizing the fish 9a, 9b, and 9c with good contrast. That is, back screen 2
By making the background white and the fish black, the fish can be recognized with good contrast.

The imaging device 4 recognizes the fish 9a, 9b, and 9c in the aquarium 1 and converts it into a video signal, and is an industrial television camera (ITV).
is used. In other words, the brightness of the pixel to be recognized w1
Emit electrical signals with different output voltages depending on the degree of
The electrical signal output from the imaging device 4 is sent to the image processing device 5.
sent to. The image processing device 5 outputs horizontal and vertical synchronization signals to the imaging device 4 to control the timing of imaging.

The image processing device 5 takes in the fish image information from the imaging device 4 at every set period Δ. Further, the image processing device 5 recognizes (binarizes) the fish based on the image information of the fish 9a, 9b, and 9c obtained by the imaging device 4, and then calculates the centers of gravity Ga and Gb of each fish.

The configuration and operation of the image processing device 5 for calculating Gc will be explained later.

A monitor 8 is connected to the image processing device 5, and the fish I
Displays TV images and the results of image processing of these images.

The signal of the center of gravity G output from the image processing device 5 is transmitted to the arithmetic device 6 at every set period Δ. The configuration of the arithmetic device 6 will be explained later, but the fish 9a and 9b.

From the center of gravity signals Ga, Gb, and Gc of 9c, the respective velocities,
Calculate V a p V b t V c. As described above, the extracted feature quantities of the fish's center of gravity G and speed V are stored in the arithmetic unit 6 for a set time T every set period Δ, and during the set time T, the center of gravity G and speed V of the fish are stored. Integrate and calculate distribution. The calculation device 6 has a fish 9a.

The normal center of gravity and velocity distributions of 9b and 9c are stored, and these normal distributions are compared with the online measured distributions. That is, if the deviation E between the fish feature quantity distribution and the normal distribution during the set time T exceeds a preset deviation, the movement of the fish is determined to be abnormal. The judgment result is sent to alarm device 7.
Send to.

When the alarm device 7 receives the abnormality detection signal, it sounds an alarm according to the abnormality level, flashes an abnormality lamp, or outputs a message in voice to urge a water quality investigation.

The normal center of gravity and speed distribution of the fish may differ even if the type of fish changes, with the only difference being whether the abnormal reaction is fast or slow, and the shape of the distribution may be the same. Furthermore, although the behavior of fish changes depending on the water temperature, reliable data can always be obtained by keeping the environmental conditions the same, such as by installing a constant temperature device to keep the water temperature constant.

Although not shown, a display and a keyboard are connected to the arithmetic unit 6, and internal calculation results such as fish feature values can be displayed, and initial setting values such as a set period Δt and a set time T can be operated.

FIG. 2 shows the configuration of the image processing device 5.

The image processing device 5 determines the centers of gravity Qa and Gb of the fish 9a, 9b, and 9c from the fish images obtained by the imaging device 4.

It has the function of detecting Gc. The timer 501 outputs a trigger to the A/D converter 502 every set period Δ. A/
The D converter 502 A/D converts the image luminance signal from the imaging device 4 every Δ in synchronization with this trigger, and stores the point image as a digital quantity in the multivalued image memory 503. For example, the multivalued image memory 503 has 256 pixels x 256 pixels x
It has a capacity of 8 bits (256 gradations for each pixel). Point images are taken into the multivalued image memory 503 at every cycle Δ.

The point image stored in the multivalued image memory 503 is sent to the binarization circuit 504. The binarization circuit 304 binarizes the fish part according to an initially set threshold. The point image in the multivalued image memory 503 at time t is converted into two by the binary circuit 504.
The fish parts are digitized and stored in binary memory 505.

For example, the binary memory 505 has 256 pixels x 236 pixels x
It has a capacity of 1 bit. The center of gravity calculation circuit 506 calculates the center of gravity Ga (ly j+ j)y Gb (itj,
t) Calculate Gc (it J* t). The input/output device 507 is an A/D converted point image, a multivalued image memory 5
03. Binary memory 505 and extracted centroids Ga and Gb
, Gc is output to the monitor 8. Furthermore, the center of gravity of the fish is also output to the calculation device.

FIG. 3 is a diagram showing a method of binarizing fish.

FIG. 3(a) is an example of a point image stored in the multilevel image memory 503. W in FIG. 3(a) is a portion of water in the background. Further, FIG. 3CQ) shows the luminance distribution on line A-A in FIG. 3(a).

As shown in this figure, in the point image, the brightness of the dot main body portion is the lowest, and the brightness increases in the order of the fish fin portion and the background water portion W. For this brightness distribution, by setting a threshold value that is smaller than the brightness W of the background water and larger than the brightness of the point itself, pixels with brightness above the threshold value are treated as background, and pixels with brightness below the value are treated as fish. become That is, for the point image S(i, jt t) at time t stored in the multivalued image memory 503, the binary circuit 504 performs (1)
, calculates equation (2) and stores binary memory B (it j+
t).

When L≦S (19,) + t), B(l t J * t) = O... (1) L>S(
i, j, t), then B (i, Jj t)=1...
(2) Figure 3(b) is a binary image of a fish. In this binary image, the blacked-out part means that it has a value of 1". A set of pixels with this value of 1" is regarded as a fish. Since the luminance of the partitions 10a and 10b has a value close to the background, the partitions are not extracted during binarization.

FIG. 4 shows the configuration of the arithmetic unit 6.

The center of gravity Ga of the fish calculated by the image processing device 5.

Gb and Gc are sent to a branching circuit 601. Branch circuit 60
1 is 3 from the horizontal position of the three centers of gravity Ga, Gb, and Gc.
Which of the three areas does the fish belong to?
The six centers of gravity that have the function of associating the center of gravity of the fish 9a with Ga, 9b with Gb, and 9c with Gc are stored in center of gravity memory circuits 602, 603, and 604, respectively, through this branch circuit. The speed calculation circuit 605 calculates the center of gravity a (it
Based on the signal of jp t), the speed V (t) is calculated using equation (3).

v (t) = lQ (it js t) −a (i
t jtt+Δt) 1/Δt...(
3) Thus, time 1.1+Δt, t+2Δt.

..., the speed of the fish at t+nΔt is V (t)
TV (t+Δt), V (t+2Δt)..., V
(t+nΔt) is calculated for each Δ. Fish 9a.

The speeds of 9b and 9c are Va (t) and Vb (t), respectively.
, Vc (t) as velocity storage devices 606°607.
608. The determination circuit 607 has the fish 9 a
? The center of gravity and velocity distribution of 9 b t 9 c during normal operation are stored. Storage device 602, 603゜604.60
6, 607, and 608 store the feature quantities of the fish's center of gravity and speed input online in a time series manner, and at the same time calculate the center of gravity distribution and speed distribution during the initial setting time T2.
It is output to the determination circuit 609. The determination circuit 609 compares the center of gravity distribution and velocity distribution of the fish, which are input online, with the normal distribution, and calculates the deviation E. The deviation E is stored in the storage device 610.
At the same time, it is output to the alarm device 7. The alarm device 7 outputs an alarm when the value of the deviation E is larger than the initial setting value.

FIG. 5 is a diagram showing the distribution when focusing on the vertical direction of the center of gravity of the fish.

FIG. 6 is a diagram showing the distribution of fish speed and frequency of appearance.

The characteristic value distribution of the fish shown in Figures 5 and 6 above is compared with the normal state C1 to determine whether the fish is normal or abnormal. In the case of C2 at the time of frenzy, the center of gravity,
Since abnormalities are detected in both speed and speed, a strong warning is output even if there is only one animal.

Also, if three animals behave abnormally at the same time, a strong warning will be output.

FIG. 7 shows an example of status determination for a type 1 animal. Measures the fish's center of gravity and velocity distribution, and determines whether the fish is normal or abnormal.

It can be divided into four patterns: center of gravity abnormality and speed abnormality. By displaying the center of gravity and speed distribution data for each fish on the monitor, the observer can always monitor the condition of multiple fish numerically. This value is displayed in blue when it is normal and in red when it is abnormal.

Figure 8 shows an example of alarm output when three fish are monitored.

Since the speed distribution of fish changes less than the center of gravity distribution, weight is given to changes in the center of gravity distribution when outputting abnormality warnings. The status of each fish as shown in Fig. 7 is input, and if one or more fish is abnormal, a strong warning is output. The center of gravity 2 also outputs a strong warning when two or more fish exhibit an abnormality in only one speed.

If only one animal exhibits an abnormality in the center of gravity, a weak warning will be output, and if only one animal exhibits a speed abnormality, no warning will be output.

〔Effect of the invention〕

According to the present invention, multiple types of fish can be recognized with a single camera, and fish behavior can be detected using the same algorithm regardless of the type of fish, so a low-cost system can be constructed.

Furthermore, since abnormalities in fish can be detected while minimizing the influence of individual differences, the presence or absence of toxic substances in water can be determined more accurately. Therefore, monitoring of inflow water at water purification plants and sewage treatment plants can be carried out quickly and labor-savingly, and the safety of water can be ensured.

[Brief explanation of the drawing]

FIG. 1 is an embodiment of the present invention, FIG. 2 is an embodiment of the image processing apparatus of the present invention, and FIG. 3 is a side view illustrating processing of an image of a fish.
Figure 4 shows the implementation side of the arithmetic device of the present invention, Figures 5 and 6.
The figure shows an example of changes in the center of gravity and speed in abnormal fish behavior, and Figures 7 and 8 are the 5-report processing side entrance. 1...Aquarium, 5...Image processing device, 6...Arithmetic device.

Claims (1)

[Claims] 1. An aquarium in which a plurality of aquatic animals are raised in order to detect the inflow of toxic substances into the water, an imaging device that images the inside of the aquatic animal and obtains image information, and distinguishes the aquatic animals from the background from the image information from the imaging device. a first processing means for obtaining the binarized information, a second processing means for detecting the position of the aquatic animal from the binarized information, and detecting the moving speed of each aquatic animal from the detected position; A water quality abnormality detection device comprising: a third processing means for determining the presence of poisonous substances in water based on the movement speed of each aquatic animal.