KR101278630B1 - Automatic injection method of a vaccine for a fish using a process of shape image - Google Patents

Abstract

PURPOSE: A vaccine automatic inoculation method through morphological image processing of fish is provided to control depth of injection based on the depth of a fish body for vaccine inoculation and to consecutively perform one-way input and automatic output of the fish. CONSTITUTION: A vaccine automatic inoculation method through morphological image processing of fishes comprises the following steps: (S1) injecting anesthetized fish into a fish transfer conveyor; (S2) measuring the depth of the fish while transferring the fish to the fish transfer conveyor; (S3) recording monoscopic images; (S4) determining insertion position and the depth of injection based on the depth of the measured fish while measuring the width of the fish; (S5) inserting an injection needle into the abdominal cavities of the fish; (S6) injecting vaccine into the abdominal cavities of the fish; and (S7) discharging the fish to the outside while returning a cross coordinate robot back to an initial position. [Reference numerals] (S1) Step of anesthetizing and injecting fish; (S2) Step of measuring the depth of the fish while transferring the fish to the fish transfer conveyor; (S3) Step of recording monoscopic images; (S4) Step of analyzing image and determining insertion position; (S5) Step of setting vertical coordinate robot; (S6) Step of injecting vaccine; (S7) Step of restoring system and discharging fish

Description

Automatic injection method of a vaccine for a fish using a process of shape image}

The present invention is to automatically perform the vaccination of fish using a vision system, a rectangular coordinate robot and a vaccine dispenser as a liquid quantitative dispenser, the statistical relationship between the size of the fish and the vaccination location for the battlefield and height By applying the data analysis-based vaccination algorithm that is analyzed and analyzed to follow the vaccination location, it minimizes the database capacity and image processing time required for the vision system, while additionally setting the user's information about the fish. Vaccine through morphological image processing of fish, which enables the administration of the correct amount of vaccine in the correct abdominal position and greatly improves the inoculation speed of the vaccine through continuous one-way fish injection and automatic discharge using the fish transport conveyor. It relates to an automatic inoculation method.

The aquaculture industry is the fastest growing industry in the world, accounting for 50% of world edible fish production. According to the recent statistics of the US Marine Atmosphere Administration, the world aquaculture production increased from 28 million tons in 1998 to 50 It has been reported that it has increased about twice in the last 10 years to 1 million tons.

As described above, as the production of fish by aquaculture increases, the size of the aquaculture is gradually increasing. Due to the nature of the aquaculture industry that breeds a large number of fish at a high density in a narrow space, due to fish scraps and feed residue Not only is the contamination of the breeding water easy, but also the incidence of infectious diseases in the aquaculture is increasing due to the contaminated water.

In order to prevent the infectious diseases of the above-mentioned aquaculture, research and development of vaccines have been actively carried out in the world, and in the case of the Korean government, the National Fisheries Animal Disease Control Law was passed in 2007 by the national government. Since February 2009, some laws have been revised, and since February 2009, in order to strengthen the safety management of aquaculture fish, vaccine supply business for the prevention of diseases of marine animals has been carried out since 2006, have.

However, even if a vaccine for fish is developed, if the automation system for vaccination can not be supplied, it is impossible to overcome the current situation in which 1 to 20,000 grains are inoculated per person with pure manpower, The increase in labor costs due to the inoculation not only brings economic burdens to the aquaculture or consumers, but also causes the damage due to careless vaccination and waste of vaccine.

Therefore, it is required to develop a vaccination automation system to reduce the time and expense required for individual inoculation of fishes and to maximize the efficiency of vaccination. In Norway, Denmark, the Netherlands and the United States, In recent years, a vaccination device has been developed in Norway to follow the vaccination site using the vision processing technology of the vision system.

In Korea, the applicant has been filed as a patent application No. 5715 in 2012 and registered as a patent No. 10-1201430. Pure personnel through automation of vaccination using vision system, Cartesian coordinate robot and vaccine dispenser as liquid quantitative dispenser. To minimize the burden of labor costs and inadvertent vaccination of the vaccination work that depended on the vaccination, and maximize the efficiency of vaccination work by allowing the correct amount of vaccine to be administered at the correct intraperitoneal location regardless of fish species or fish size. BACKGROUND OF THE INVENTION Automated vaccination methods and vaccination devices for fishes are known.

In the vaccination method previously filed by the applicant, the vaccination location is applied by applying a template matching vaccination algorithm so that the vaccination expert can find the vaccination location as much as possible. By displaying the images of the fish of various sizes as template images for each species, extracting the template images with the highest matching rate from the images of the injected fish, and then using the 3-axis Cartesian coordinate robot to follow the vaccination location. Was used.

However, in order to inoculate the vaccine at the correct inoculation site for each fish having various types and sizes using the vaccination algorithm of the template matching method, the template image of the fish showing the vaccination site is converted into a database for each fish body size and species And the template image must be stored as much as possible in order to increase the accuracy of the vaccination.

Therefore, not only a considerable amount of time and money must be invested in the task of database conversion of the template image of a fish, but also a large storage space is required in the vision system for storing the database, and the image of the fish and the matching rate The time required for the image processing for extracting the highest template image is short, resulting in a delay in inoculation time.

In order to shorten the time of image processing by the template matching method as described above, before the vaccination work, additional settings for fish information, such as the operator selecting a fish species or the like and inputting it into the system, must be made. Additional setting work also has a problem of delaying the overall inoculation time, as well as the risk of causing errors in vaccination due to the input of incorrect information.

In addition, in the case of the filing application, the method of injecting fish using the Y-axis robot among the Cartesian coordinate robots and returning the fish back to the feeding direction after the vaccination was applied, there was an inefficient problem in terms of inoculation speed. In some cases, fish were caught in the needle after vaccination, which caused the insertion depth to be unnecessarily deep when the needle was inserted using the Z-axis robot of the Cartesian coordinate robot. This is due to a factor that prevents the instant removal of the needle penetrating the abdominal cavity of the fish after inoculation.

Korea Patent Registration No. 10-1201430 (Publication Date: November 23, 2012) Republic of Korea Patent Application No. 10-2012-132748 (Application Date: November 22, 2012)

The present invention has been made in order to supplement the above problems of the prior application, the automatic vaccine vaccination method through the morphological image processing of the fish according to the present invention of the planar image by the camera in the process of transferring the fish to the fish transport conveyor In addition to taking pictures, the height of the fish is measured so that the injection depth can be adjusted based on the height of the fish during vaccination, as well as continuous one-way feeding and automatic discharge of the fish. The first technical problem is to prevent the damage of the fish body according to the depth of injection and to improve the inoculation speed of the vaccine.

In addition, the automatic vaccine vaccination method through the morphological image processing of the fish according to the present invention, the fish body analysis vaccination algorithm that the data correlation between the body width of the fish and the vaccination location for the total length and height through a statistical analysis By minimizing the vaccination location by applying to the system, the database capacity and image processing time required for the vision system can be reduced to the maximum, and the correct amount of vaccine is administered at the correct peritoneal location even without the user's additional setting of the fish information. The second technical problem is to provide an efficient and economical inoculation method as much as possible.

In addition, the fish transfer conveyor is arranged in two rows so that the height measurement and plane image processing can be performed in parallel on each conveyor, while one orthogonal coordinate robot moves two times while moving the conveyor. By improving the inoculation method so that vaccination can be performed sequentially, it is possible to shorten the time required for vaccination and to make the treatment of the vaccinated fish easier. It is a technical problem.

Finally, using the compressed air discharged from the compressor and the dispenser controller as in the previous application to perform the quantitative discharging of the vaccine by the vaccine dispenser, the needle is inserted at an angle through the abdominal cavity of the fish body, using a pneumatic cylinder The fourth technical problem is that the needle can be quickly removed from the abdominal cavity of the fish, thereby preventing the fish from being caught in the needle and preventing it from being damaged or damaged. Shall be.

Vaccine automatic inoculation method through the morphological image processing of the fish according to the present invention as a means for solving the above technical problem, fish anesthesia and input step of injecting the anesthetized fish using an anesthetic agent to the fish transport conveyor, and The fish is transferred to the conveying conveyor, the fish feeding and height measuring step of measuring the height of the fish by the height sensor installed on the side of the fish conveying conveyor, and the plane image of the fish by the camera installed on the upper side of the fish conveying conveyor While measuring the body width and the full length of the fish based on the planar image photographing step and the planar image taken by the camera, the measurement of the needle was performed based on the body width measured from the planar image and the height of the fish measured by the full length and height sensor. Image analysis and inoculation position following step to determine insertion position and depth, X-axis robot, Y-axis robot and Z-axis An orthogonal robot setting step of inserting the needle of the syringe body into the abdominal cavity of the fish by using an orthogonal coordinate robot consisting of a bot, and a vaccine for injecting the vaccine stored in the vaccine storage container into the abdominal cavity of the fish through the syringe body and the needle While the inoculation step and returning the Cartesian coordinate robot to the original position, it is characterized in that it is made through a system restoration and fish discharge step of discharging the completed vaccination fish to the outside using a fish transport conveyor.

More preferably, in the image analysis and inoculation position tracking step, a fish body analysis algorithm for measuring the shape of the fish is applied instead of the existing template matching method, and the fish body analysis algorithm captures an image frame photographed by a camera. Converting the captured image into an 8-bit grayscale image, adjusting the contrast ratio generated in the image conversion step, converting the image with the contrast ratio adjusted to a black and white image by binarization; Morphologically processing the image of the fish using the binarized black-and-white image, and measuring the body width and the full length of the fish with the processed image. The following tracking method of inoculation location is applied by the fish analysis algorithm. Of the preparation compared to the chepok and full-length and the previously stored data by the height of the fish to a one-dimensional regression analysis method for measuring through the height sensor is a way that following the inoculation location from the data.

The pre-stored data by the one-dimensional regression analysis method, the body width, length and height of the fish measured from the images taken by the camera on the upper and side portions of the fish as an independent variable, the abdominal position where the actual vaccination of the fish is made Is a dependent variable, and the correlation between the independent variable and the dependent variable is derived as a regression coefficient, and data on the height of the fish is inoculated with a vaccine confirmed from a two-dimensional plane image taken by a camera from the top of the fish. While the position is displayed as a circular white paper on the surface of the fish, the height value measured by irradiating the beam of the laser displacement sensor to the position indicated by the white paper as a dependent variable, Using the measured fish height as an independent variable, the correlation between the independent and dependent variables is derived as a regression coefficient. The features.

In addition, the process from the fish anesthesia and feeding step to the image analysis and inoculation position tracking step is performed in parallel for each of the two rows of fish transport conveyors arranged in parallel along the direction of fish transport, and the Cartesian coordinates The robot setting step and the vaccination step are carried out sequentially for the fish placed in each fish transport conveyor with one orthogonal coordinate robot, and the vaccination step controls the compressed air discharged through the regulator from the compressor. Vaccine supply and valve operation of the syringe body through the liquid quantitative dispensing method, the process of removing the needle from the abdominal cavity of the fish after vaccination, as a pneumatic cylinder installed on the z-axis robot of the Cartesian coordinate robot with the syringe body It is characterized in that it is performed quickly with a needle adjuster, the orthogonal seat In the robot setting step is characterized in that at an angle such that insertion inclined at a predetermined angle through the peritoneal cavity of the injection needle of the syringe body of fish.

According to the present invention as described above, in addition to the control of the depth of injection based on the height of the fish during the inoculation of the vaccine, as well as by applying a conveyor system capable of continuous one-way injection and automatic discharge of fish, according to the excessive depth of injection It is effective in preventing the damage of the fish and improving the inoculation speed of the vaccine, and by following the vaccination location by the body analysis vaccination algorithm, which minimizes the database capacity and image processing time required for the vision system. At the same time, there is an effect that the correct amount of vaccine can be administered at the correct intraperitoneal location even without the user's additional setting for the information of the fish, thereby providing a more efficient and economical inoculation method than the case of the previous application.

In addition, by rationally improving the vaccination method by using two rows of fish transport conveyors and one orthogonal coordinate robot, the time required for vaccination is further reduced, and the vaccination of fish is easier to process. The needle is inserted at an angle through the abdominal cavity of the fish and by using a pneumatic cylinder to quickly withdraw the needle from the abdominal cavity of the fish, the fish is hooked up to the needle By blocking the phenomenon in advance, it is possible to more reliably prevent damage or damage to the fish, such as to provide a better effect than the previous application.

1 is a flow chart of the automatic vaccine vaccination method through the morphological image processing of the fish according to the present invention.
2 is a flowchart and a photograph of a body analysis algorithm applied to the present invention.
3 is a photograph illustrating the correlation between the measured position of the flounder, the full length (TL) and the height (BH) and the vaccination position (BWx, TLy, BHz) of the flounder;
Figure 4 is a photograph showing the principle of measuring the vaccination position (BHz) in the height (BH) direction using a laser displacement sensor.
5 is a graph showing the correlation between the body width (BW) of the flounder and the distance (BWx) to the vaccination position in the x-axis direction.
6 is a graph showing the correlation between the full length (TL) of the flounder and the distance (TLy) to the vaccination position in the y-axis direction.
7 is a graph showing the correlation between the height of the flounder (BH) and the distance (BHz) to the vaccination position in the z-axis direction.
Figure 8 is a graph showing the image processing time according to the body analysis vaccination algorithm.
9 is a graph showing the inoculation error according to the input angle and size of the flounder.
10 is a plan view of the vaccination device used in the present invention.
11 is a plan view of a fish transport conveyor.
12 is a side view of FIG. 11;
13 is an external perspective view of a rectangular coordinate robot.
14 is a front view of FIG. 13;
Fig. 15 is an enlarged front view of the main portion showing the vaccine dispenser.
Figure 16 is a schematic control and pneumatic circuit diagram of the vaccination device used in the present invention.
17 is a piping diagram showing a separate system of the pneumatic circuit in FIG.
18 is a flow chart of the operation of the vaccination device used in the present invention.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

Similarly, the automatic vaccine vaccination method through the morphological image processing of the fish according to the present invention, taking a plane image of the fish in the process of transporting the fish, and then determine the inoculation position of the vaccine based on the plane image, the Cartesian coordinates By using a robot, a vaccine dispenser including a syringe body and a vaccine storage container is moved to insert a needle into the abdominal cavity of the fish, thereby injecting the vaccine into the abdominal cavity of the fish, and the overall inoculation process is similar to the previous application. It is done.

Part of providing a difference between the present invention and the prior application, the fish transfer method using a fish conveying conveyor, and further measuring the height of the fish during the fish transfer process to follow the vaccination location with the plane image taken with the camera Or image processing that tracks the vaccination location.

Vaccine automatic inoculation method through the morphological image processing of the fish according to the present invention, as shown in Figure 1, fish anesthesia and input step (S1), fish transfer and height measurement step (S2), plane image taking step (S3) ), Image analysis and inoculation position following step (S4), Cartesian coordinate robot setting step (S5), vaccination step (S6), system restoration and fish discharge step (S7), each step For ease of understanding, reference will be made to the vaccination apparatus 1 shown in FIGS. 10 to 17 from the following description.

The fish anesthesia and input step (S1) is a start step of vaccination for injecting the anesthetized fish using the anesthetic agent to the fish transport conveyor (3), wherein the anesthetic agent is harmless to the fish, for example, phenoxyethanol (Phenoxyethanol Although chemical anesthetics such as) may be used, natural anesthetics for fish using the bamboo fruit have recently been developed and marketed, and thus it is more preferable to use them.

In addition, the method of anesthetizing fish in the culture tank by dissolving an appropriate amount (about 5 to 10 wt% of the breeding water) in the breeding water of the culture tank rather than anesthetizing the fish one by one using a syringe is more general. Advantageously, by arranging the inoculation waiting table 4 at the inlet side of the fish transport conveyor 3, the anesthetized fish is collected in the inoculation waiting table 4 by a predetermined amount, and then the fish is transferred to the fish conveying conveyor 3. It is preferable to add one by one sequentially.

In addition, the fish transport conveyor (3) used in the present invention, in consideration of the fact that the fish anesthetized in the culture tank is introduced into the fish transport conveyor (3), as shown in Figs. In addition to using the conveyor belt (3a) is formed with a dehydration hole to be discharged smoothly, it is preferable to install a drain chamber (3e) provided with a drain port (3f) on the lower side of the conveyor belt (3a).

After the fish anesthesia and input step (S1) as described above, while passing the anesthetized fish to the fish conveying conveyor (3), by using the height sensor (3b) installed on the side of the fish conveying conveyor (3) The fish go through the height measurement and the height measurement step (S2).

The height sensor (3b) can measure the body height of the fish moving together with the fish conveying conveyor (3), that is, any type of sensor can be used as long as the body thickness of the body can be measured, but it is possible to select a plane image of the fish from the selection source. Instead of the CCD camera (Charge-coupled device camera) used, a laser beam sensor (product name: ZX-LT030, manufacturer name: Omron Corp.) can be used to measure the fish height more accurately by optical method. It is desirable to.

In addition, by comparing the images captured by the laser sensor to detect the Doppler shift light scattered from the surface of the object passing through the intersection of the two laser beams, or by two small cameras in a stereo vision method to compare the height of the fish in three dimensions In addition to the vision sensor to be measured, a general-purpose optical sensor or displacement sensor capable of measuring the size of the object may be used as the height sensor 3b.

After the fish transfer and height measurement step (S2) as described above, the plane image taking step (S3) of taking a plane image of the fish with the camera 6 installed on the upper side of the fish transport conveyor (3). In addition, the camera 6 is preferably installed as a CCD camera on the upper inside of the apparatus main body 2 through which the fish transport conveyor 3 passes, as shown in FIG. 10, and the operator takes an image captured by the camera 6. It is desirable to provide a vision monitor 7 on the front side of the apparatus body 2 so that it can be directly monitored.

While photographing the plane image of the fish with the camera 6 as described above, in order to perform vaccination with the Cartesian coordinate robot 10, the operation of the fish transport conveyor (3) is temporarily suspended in the plane image photographing step (S3). To this end, a proximity sensor (3c) for detecting the transfer state of the fish is installed on the side of the fish transport conveyor 3 corresponding to the lower side of the camera (6).

Therefore, when it is confirmed by the proximity sensor 3c that the fish conveyed along the fish conveying conveyor 3 are in the photographing range θ of the camera 6 as shown in FIG. 16, the main sensor associated with the proximity sensor 3c is connected. The controller 26 determines this and controls the conveyor driving unit 3d to temporarily stop the operation of the fish conveying conveyor 3, and after the vaccination by the rectangular coordinate robot 10 is completed, the fish conveying conveyor 3. Will be restarted.

After going through the planar image photographing step S3 as described above, the body width (BW) and the total length (TL) of the fish are measured based on the planar image photographed by the camera 6. Image analysis and inoculation position tracking to determine the insertion position and depth of insertion of the needle 19 based on the body width measured from the planar image and the body height (BH) measured through the full-length and height sensor 3b. In step S4.

In the present invention, instead of the template matching vaccination algorithm applied in the image analysis and inoculation position following step (S4), the correlation between the body width of the fish and the vaccination position for the total length and height is dataized through statistical analysis. Body analysis vaccination algorithm is applied.

In order to apply the above-described vaccination algorithm of the body analysis, a body analysis algorithm for measuring the body width and the full length of the fish from the two-dimensional plane image of the fish, as well as the body width measured by the body analysis algorithm and It is necessary to identify and database the correlation between the height measured by the full length and height sensor 3b and the actual vaccination location of the fish.

First, as shown in FIG. 2, the body analysis algorithm captures an image frame photographed by the camera 6 (S41), and converts the captured image into an 8-bit gray scale image (S42). ), Adjusting the contrast ratio generated in the image conversion step (S43), converting the image with the contrast ratio adjusted to a black and white image by binarization (S44), and the image of the fish using the binarized black and white image It comprises a step of morphologically processing (S45), and measuring the body width and the full length of the fish in the processed image (S46).

In the photograph attached to FIG. 2, a body analysis process of the flounder is shown, but the method of measuring the body width (BW) and the full length (TL) of the fish from the two-dimensional plane image photographed by the camera 6 is different from the flounder. In addition, the body analysis algorithm of FIG. 2 is proposed to measure the body width (BW) and the total length (TL) by scanning the contour of the fish moving on the fish transport conveyor (3).

As in the body analysis algorithm, the original image photographed by the camera 6 is converted into an 8-bit grayscale image, and then processed into a measurable image by adjusting contrast ratio, binarization, and morphological image processing. The adjustment is to improve the quality of the image, which can be adjusted by contrast with an 8-bit grayscale image with 255 the brightest image information and 0 for the darkest image information.

The above contrast adjustment is for the purpose of binarizing the intensity of the image selectively and binarizing it in the interval, and the binarization operation is to express the contrast of the selected interval in black and white based on a randomly set brightness level. Minimize the influence of the external lighting of the environment and sharpen the outline of the object to be measured.

The morphological image processing is to selectively filter the dorsal, rear and caudal fins of fish, and is a morphological transformation involving structural elements (SE) in two matrices and a matrix. In the furnace, an erosion operation for reducing the components of an image and a dilation operation for enlarging the components of an image are commonly applied.

As the name implies, the erosion operation means that the erosion operation is performed. The erosion operation acts as a minimum value filter that converts the smallest value into a value at the center of the pixel in a predetermined area. In a binary image, an object area is defined. When viewed in white, these areas decrease and black areas increase, while in gray images (or color images), the calculation results change toward decreasing dark areas and increasing dark areas, and expansion operation is the operation that widens the area as opposed to erosion operation. It acts as a value filter.

The erosion operation may cause small objects to disappear according to the size and the number of times of use of the filter. In contrast, the expansion operation may cause disappearance of small holes inside the object. It is mainly used for removal, and the proper mixing of erosion and expansion operations can remove the desired clutter with little change in the size of the object.

As described above, the morphological image processing using erosion and expansion operations as a basic operator includes performing an erosion operation first and performing expansion operation to remove miscellaneous operations, and performing expansion operation first and then performing erosion operation. Closing operation is performed to remove the holes, and the results of these two operations are combined and used as an image for measuring the body width (BW) and the full length (TL) of the fish.

In the morphological treatment of fish, the hot operation is to soften the contour of the fish individual and break the narrow isthmus, while removing the slender protrusion, and the close operation also provides the function of softening the contour, but Contrary to the calculations, it adds narrow breaks and long thin gaps in the contour, removes small holes and fills the gaps.

The image of the flounder processed according to the above contrast adjustment, binarization image processing and morphological image processing, respectively, is shown as a representative photograph on the right side of the fish analysis algorithm in the drawing. The size and contrast ratio of the 8-bit grayscale image set in the processing and the morphological image processing, and the matrix value of the binarization region and the structural element SE are shown.

8-bit grayscale image Contrast ratio Binarization Zone Matrix Value of Structure Element 1600 × 1200 pixels 1.5 0-160 9 × 9 matrix

In addition, the body width (BW) of the fish is to measure the maximum width of the meat portion excluding the dorsal fin and the rear fin, and to obtain the center of gravity of the fish, scan from the left and right sides of the fish except the fin portion to the center of gravity Detecting the maximum width point for each right side, detecting a straight distance passing through the head and tail of the fish using edge pattern matching and edge point detection. The left and right repair distances (distances corresponding to the left and right maximum widths based on the straight line distance) are measured using the right maximum width and the straight line distance. The fish width is derived from the sum of the left and right repair distances. do.

On the other hand, the total length (TL) of the fish is to measure the maximum end of the head and tail fin of the fish, the step of recognizing the position of the head and tail of the fish with the edge pattern matching, the fish detected by the edge position detection Measuring the straight line distance from the end of the head to the measurement position in the tail sack pattern by edge pattern matching, and measuring the straight distance from the measured position in the sack pattern and the tail end position of the fish detected by edge detection The total length of the fish is derived from the sum of the two straight distances. Thus, the reason for measuring the two straight distances is that the tail fins of the fish are often not in line with the body of the fish.

The morphological image processing of the fish is possible through the above-described body analysis algorithm. The algorithm itself for extracting and acquiring the data of the required area from the image captured by the CCD camera is widely used in the fields such as face recognition and pattern recognition. In addition to being used, since it can be implemented in a wide variety of forms according to the type of data (information) required, it is clear that the body analysis algorithm applied to the present invention is not limited to the method shown in FIG. The body weight (BH) of the fish is measured by the height sensor (3b) and used to determine the vaccination location together with the body width (BW) and the full length (TL) measured by the fish body analysis algorithm.

The most important part of the phylogenetic vaccination algorithm applied to the present invention is the fish's body width (BW), the total length (TL) and the body height of the fish measured through the body height sensor (3b) BH) is compared with the pre-stored data by 1-D regression analysis, and the inoculation position is followed from the corresponding data.

In order to collect data by one-dimensional regression analysis as described above, the measurement of the shape of the fish and the investigation of the vaccination location based on the same should be made. In the present invention, the flounder of the flounder using a body measurement system installed in a laboratory scale space As a representative example, a process of constructing a database for halibut vaccination by analyzing the body and identifying the correlation with the vaccination location based on the same will be described.

The body measurement system may be divided into a measurement case, an image acquisition unit and a weight measurement unit of a fish, and the image acquisition unit is a vision system (Panasonic, PV200), a CCD camera (upper: 1, side: 1), lighting device (Bottom: 1 unit, side: 1 unit), the CCD camera installed on the upper part of the measuring case measures the body width and the full length of the fish, and the CCD camera installed on the side of the measuring case measures the height of the fish. Done.

In addition, the lighting device is also installed on the lower and side of the measuring case, respectively, and relatively brighter than the external lighting to clear the outline of the fish, and white LED and transparent glass plate to facilitate filtering of the fins of each part The lighting device is combined with each other, and the weight measuring unit is composed of a load cell and an indicator for displaying the weight.

The measuring case is preferably manufactured to have dimensions similar to the apparatus body 2 of the vaccination apparatus 1 in FIG. 10, and the CCD camera used for the measuring case is the camera 6 of the vaccination apparatus 1. In the case of the height sensor 3b, it is possible to measure the height of the flounder, but it is impossible to compare the flounder with the actual vaccination position (abdominal position) by outputting the side shape of the flounder on the screen. The CCD camera is installed on the side of the measurement case instead of the height sensor 3b.

After the flounder is put into the inside of the measuring case equipped with the lighting apparatus as described above, the flat image of the flounder is taken by a CCD camera installed on the upper side of the measuring case, and the flounder is carried out by the CCD camera installed on the side of the measuring case. The body image, height, and height of the flounder were measured by using the CCD camera and the image analysis algorithm. The measurement errors were 0.40mm and 0.49, respectively. It became within mm and 0.01 mm.

Measure the body width and the total length and height of the flounder in the same manner as above, but the body width (BW) and full length (TL) by size of the flounder (100g: 200, 200g: 200, 300g: 200) as shown in FIG. The relationship between the type of flounder and the location of vaccination was investigated by the vaccination expert directly indicating the location of vaccination on the xyz coordinate plane of the 2D plane image.

In Figure 3, the body width (BW) is the width of the fleshy part except the rear fin (Anal fin) of the flounder and Dorsal fin (Dorsal fin) of the flounder, BWx is the flesh portion except the rear fin of the flounder on the x-axis line By reference, the distance from the largest end of the dorsal fin to the injection site indicated by the expert.

And, in the case of the full length (TL) is the length across the maximum end of the head and the tail fin (Caudal fin) of the flounder in Figure 3, TLy is displayed by the expert at the maximum end of the head of the flounder on the y-axis line The distance from one vaccination site, BHz, represents the abdominal thickness from the bottom of the z-axis line based on fish height (BH) to the vaccination site indicated by the expert.

Here, the fish body analysis algorithm for the measurement of the flounder shape is to measure the flounder shape through the two-dimensional plane image. BWx and TLy can be measured based on the image, but exist in the three-dimensional space like BHz. It is difficult to accurately measure the components by the 2D plane image alone.

Therefore, the inoculation position of the vaccine identified from the two-dimensional plane image taken by the CCD camera from the top of the flounder as shown in Figure 4 on the surface of the fish as a circular white paper, while the laser displacement at the position indicated by the white paper BHz was measured by irradiating the beam of a sensor (product name: LK-501, a manufacturer name: Keyence Corp.).

The body analysis-type vaccination algorithm established through the above investigation process requires an image processing time for measuring the shape of the flounder and a time for calculating the vaccination position from the data measuring the shape of the flounder, and many flounder in a short time. In order to inoculate, the correlation between the shape of the flounder and the location of vaccination should be simplified, as well as the analysis algorithm of the flounder.

Therefore, in the present invention, in order to derive a correlation between the shape of the flounder and the vaccination position measured at each reference point, a simple linear regression analysis is performed from the measured data of the flounder. We modeled the first-order polynomial with the y-intercept as zero as (1), (2), and (3).

BWx = αBW---------(1)

TLy = βTL---------(2)

BHz = γBH---------(3)

Based on the above formula, the actual vaccination of the flounder, with the body width (BW), total length (TL) and body height (BH) of the flounder measured from images taken by CCD cameras at the upper and side portions of the fish as independent variables. Using the intraperitoneal position BWx, TLy, and BHz as the dependent variable, the correlation between the independent variable and the dependent variable is identified to derive respective regression coefficients α, β, and γ.

By using the regression coefficients (α, β, γ) derived as described above, data that can be followed by the vaccination position (x, y, z) as shown in (4), (5), (6) can be programmed. Where x _anal and y _head are the initial points (reference points) at which measurements of body width (BW) and full length (TL) are measured in the 2D planar image coordinate system.

x = x _anal + αBW---------(4)

y = y _head + βTL---------(5)

z = γ BH------------(6)

FIG. 5 is a graph showing the correlation between the measured body width (BW) of the flounder and the distance from the maximum end of the meat portion except the rear fin of the flounder to the vaccination position (BWx) in the x-axis direction. As can be seen, the regression coefficient α in Equation (1) is 0.195, and the correlation coefficient (coefficient representing the correlation between two variables) is 0.988, and based on this, the meat part of the flesh except for the rear fin It can be seen that about 20% of the halibut body width (BW) in the x-axis at the maximum end (x _anal ) is the vaccination position in the x-axis.

6 is a graph showing the correlation between the measured total length (TL) of the flounder and the distance (TLy) from the head end of the flounder to the vaccination position in the y-axis direction, as shown in Equation (2) The regression coefficient β at is 0.296, and the correlation coefficient is 0.996. Based on this, about 30% of the total length of the flounder in the y-axis from the y _head of the flounder is in the y-axis direction. It can be seen that the vaccination site.

FIG. 7 is a graph showing the correlation between the measured height of the flounder (BH) and the abdominal thickness (BHz) of the flounder from the bottom surface based on the z-axis direction, as shown in the graph (3). The regression coefficient γ is 0.599, and the correlation coefficient is 0.988. Based on this, it can be seen that about 60% of the flounder height in the z-axis direction from the bottom surface is the vaccination position in the z-axis direction.

In this way, the correlation (regression coefficient) between the body widths of various fish as well as the actual length of vaccination for the full length and body height (regression coefficient) is calculated in the same manner as in (1), (2) and (3). After identifying by the analysis method, it is formulated in the same manner as the formulas (4), (5), (6), and the database can be programmed by programming the formula according to the expression following the inoculation position.

Accordingly, the body width and the full length of the fish are measured from the two-dimensional plane image photographed by the camera 6, and the height of the fish is measured using the height sensor 3b, and the data is converted into a regression coefficient assigned to the fish. When it is inserted into the inoculation position tracking formula programmed together, the inoculation position in the X-axis direction and the Y-axis direction and Z-axis direction required for the operation of the Cartesian coordinate robot 10 is followed.

When applying the phage analysis vaccination algorithm as described above, it is possible to program and store only the data and formula necessary for tracking the inoculation location by the fish species and size, that is, the representative database for each fish species. Tracking of vaccination sites for almost all fish species is possible.

As a result, when compared to the task of securing a large number of template images for each type of fish and the size of the fish and making them into a database like the existing template matching vaccination algorithm, the time and cost of constructing the vaccination algorithm are maximized. This can reduce the amount of database storage required for vaccination algorithms.

In addition, it is possible to follow the inoculation time of the vaccine by simple image processing that analyzes only the type of the injected fish, rather than extracting the image with the highest matching rate from the injected fish image. Not only can it contribute greatly to the aspect, it also enables rapid and smooth image processing and vaccination without the need for additional setting of fish information such as the selection of fish and the like into the system before vaccination. will be.

Table 2 shows the results of experiments comparing the inoculation rate by the automatic vaccine vaccination method according to the present invention with the conventional manual inoculation speed, Table 3 is a filed inoculation rate by the automatic vaccine inoculation method according to the present invention The experimental results compared with the case of vaccination method are shown.

Halibut weight Inoculation rate of the present invention
(1 person / 1 hour) Manual vaccination speed
(4 people / 1 day) 150g ~ 300g 1000 or more 22,000 to 30,000 300g ~ 500g 850 or more About 22,000 500g ~ 1,000g More than 800 About 18,000

type Vaccine Algorithm Inoculation rate (per hour) Inoculation error (mm) Servo type Template matching 2,000 Within 0.6 Conveyor type Body analysis 2,800 Within 1.9

According to the experimental results of Table 2, the conventional manual inoculation method showed a difference in inoculation speed by weight, and the vaccine automatic inoculation method according to the present invention showed a maximum inoculation speed up to three times or more than manual inoculation, as shown in Table 3 It can be seen that the inoculation speed of the conveyor type to which the fish-body type vaccination algorithm according to the present invention is applied is improved by about 40% than the servo type to which the template matching vaccination algorithm is applied.

And, as the image processing time closely related to the inoculation speed was measured for each length of the flounder as shown in the graph of FIG. 8, if there was no significant difference in the total length of the flounder from the experimental results, the body analysis It can be seen that the average image processing time required for the expression vaccination algorithm is about 25 ms, which is about one third of the image processing time by the template matching method.

For reference, in the case of the template matching vaccination algorithm, since the time required for analyzing the matching rate by detecting the feature points for the entire area of the flounder template images and the flounder image that is databased in the search group, the area of the flounder At about 475cm ² , the time required for image processing was up to 72ms.

In addition, as a result of analyzing the inoculation error for each angle of the flounder as shown in Figure 9 by applying a fisheye analysis vaccination algorithm to the conveyor vaccination device (1) used in the present invention, the maximum error range is about 1.9 mm, which is about twice as large as the case of applying the template matching vaccination algorithm to the pre-served servo vaccination device as shown in Table 2.

However, the error range of the above level is not a problem in consideration of the abdominal size of the fish, the inoculation method according to the present invention in consideration of the various aspects such as the inoculation speed except the inoculation error It can be seen that it is much more efficient and economical than the inoculation method.

After the image analysis and inoculation position following step (S4) using the above-described body analysis vaccination algorithm, the rectangular coordinate consisting of the X-axis robot 11, Y-axis robot 12 and Z-axis robot (13) The robot 10 is subjected to an orthogonal coordinate robot setting step (S5) of inserting the needle 19 of the syringe body 15 for the vaccine dispenser 14 into the abdominal cavity of the fish.

As shown in FIGS. 13 and 14, the Cartesian coordinate robot 10 moves along a guide rail 11a of the X-axis robot 11 in a straight line (left and right on the drawing). 11b), the Y-axis robot 12 is mounted, the Z-axis on the Y-axis feed table 12b moving along the guide rail 12a of the Y-axis robot 12 in a straight line (before and after drawing) direction The robot 13 is mounted, and the vaccine dispenser 14 is installed on the Z-axis feeder 13b moving along the guide rail 13a of the Z-axis robot 13 in the vertical (up and down) direction.

Therefore, the X-axis robot 11, the Y-axis robot 12, and the Z-axis robot according to the inoculation position and the scanning depth of the vaccine followed through the fisheye analysis vaccination algorithm in the image analysis and inoculation position following step (S4). Vaccine installed on the Z-axis feeder 13b of the Z-axis robot 13 by moving the feeders 11b, 12b, 13b along the X-axis direction, the Y-axis direction, and the Z-axis direction. The needle 19 of the syringe body 15 for the dispenser 14 is inserted into the abdominal cavity of the fish, which is the following vaccination position.

The X-axis robot 11, the Y-axis robot 12 and the Z-axis robot 13 is a known technology widely used in the field of automation equipment named as Linear Motion Guide (LM Guide), reducer Screw shafts which are rotated by a driving motor with a shaft are built in each of the guide rails 11a, 12a, 13a, and the feed shafts 11b, 12b, 13b are nuts on the screw shaft. In this way, when the screw shaft is rotated by the drive motor, each of the carriages 11b, 12b, 13b is along the guide rails 11a, 12a, 13a of the robot, and the X, Y, and Z axes. It will move in the axial direction.

In addition to the orthogonal coordinate robot 10 of the LM guide system using a drive motor and a screw shaft as described above, a cylindrical orthogonal tube in which each feed table 11b, 12b, 13b is mounted on a piston rod of a hydraulic cylinder or a pneumatic cylinder. It is to be understood that various robotic mechanisms such as a geared rectangular coordinate robot can be applied to coordinate robots or rack gears moving in a linear direction by engaging a drive gear. .

After the orthogonal coordinate robot setting step (S5) as described above, the vaccine dispenser used in this step is subjected to the vaccination step (S5) of injecting the vaccine into the abdominal cavity of the fish using the vaccine dispenser (14). As shown in FIG. 15, the vaccine storage container 16 in which a predetermined amount of vaccine is stored and the vaccine introduced from the vaccine storage container 16 are injected into the abdominal cavity of the fish through the needle 19. It is composed of a syringe body 15 to be injected, the dose of vaccine (one) using the vaccine dispenser 14 is preferably about 0.1ml.

As shown in the figure, it is preferable to mount the small vaccine storage container 16 directly to the syringe body 15 in terms of simplifying the structure of the vaccine dispenser 14 and preventing unnecessary waste of vaccine. While separating the storage container 16 from the syringe body 15, it is also possible to connect the vaccine storage container 16 and the syringe body 15 using a tube or pipe for vaccine supply.

In addition, the syringe body 15 may be used as long as it can inject the vaccine stored in the vaccine storage container 16 in a liquid quantitative dispensing method, and the vaccine storage container 16 and the syringe body 15 If is separated, the vaccine stored in the vaccine storage container 16 should be supplied to the syringe body 15 using compressed air or a small pump, of course.

More preferred vaccine automatic vaccination method that can be applied to the present invention, while the injection operation of the vaccine by the vaccine dispenser 14 in the vaccination step (S6) is carried out in a liquid quantitative discharge method using compressed air, while The process of removing the needle 19 from the abdominal cavity of the fish immediately after the vaccination step (S6) is also orthogonal with the syringe body (15) so that the phenomenon of being caught on the needle 19 can be prevented. It is to be performed quickly by the needle regulator 17 as a pneumatic cylinder installed on the z-axis robot 13 of the coordinate robot 10.

To this end, as shown in FIG. 15, the syringe body 15 and the vaccine storage container 16 of the vaccine dispenser 14 are connected to pneumatic lines 15a and 16a, and the syringe body of the vaccine dispenser 14 ( 15) is provided with a pneumatic discharge port (15b) for the discharge of the compressed air used for vaccination, the inside of the syringe body (15) is a pneumatic type valve operation is performed by the compressed air introduced through the pneumatic line (15a) Valve mechanism is installed.

As shown in FIGS. 16 and 17, the pneumatic line extends from the discharge port of the compressor 20 as a pneumatic mechanism for injecting the vaccine into the abdominal cavity of the fish in a quantitative manner using compressed air as described above. 29 is connected to the regulator 21, which is a pressure regulator, and the pneumatic line 29 extending through the regulator 21 is branched into two pneumatic lines 15a and 16a, so that the syringe body 15 and the vaccine It is connected to the storage container 16, respectively, and the pneumatic lines 29 and 15a connecting the regulator 21 and the syringe body 15 are provided with a dispenser controller 9 for controlling the discharge amount of the vaccine.

In addition, the air filter (F) is installed on the suction side of the compressor 20, the solenoid valve (SV) is installed on the pneumatic line 29 corresponding to the outlet side of the regulator 21, the regulator 21 and the vaccine storage A relief valve (safety valve) RV is installed in the pneumatic line 16a on the vessel 16 side, and a relief valve RV is also provided in the pneumatic line 15a connecting the dispenser controller 9 and the syringe body 15. Solenoid valves SV are provided respectively.

Therefore, the discharge amount of the vaccine is controlled by the air pressure applied to the vaccine storage container 16 and the syringe body 15 via the regulator 21 and the opening time of the solenoid valve SV which performs the valve operation of the syringe body 15. By controlling the operation of the regulator 21 and the solenoid valve SV using the input means provided to the dispenser controller 9, the discharge amount of the vaccine can be arbitrarily adjusted by the user.

The dispenser controller 9 is also preferably connected to the main controller 26 by a cable so as to be included in the overall control range of the vaccination device 1, if necessary, the operation of the regulator 21 and the syringe body 15 It may be controlled directly by the main controller 26 without passing through the dispenser controller (9).

In this case, the valve mechanism installed in the pneumatic line of FIG. 17 should be connected to the main controller 26, and as shown in FIG. 10, the operation switch 8 or the touch screen as an input means installed on the front surface of the apparatus main body 2, 8a) should be connected to the main controller 26 in a state where an operation function of the vaccination device 1 and a function of controlling the discharge amount of the vaccine are additionally provided.

Since compressed air is used for vaccination by the vaccine dispenser 14 as described above, it is advantageous to install the needle adjuster 17 for quickly removing the needle 19 from the abdominal cavity of the fish as a cylinder mechanism using compressed air. To this end, as shown in FIG. 15, the needle adjuster 17 is installed on the Z-axis feeder 13b together with the syringe body 15 for the vaccine dispenser 14.

The needle adjuster 17 preferably uses a flat plate-shaped pneumatic cylinder for firmly mounting the syringe body 15, and the needle adjuster 17 also has a pair of pneumatics for inlet and outlet of compressed air. A line 17a is connected and installed, and each pneumatic line 17a extends from the discharge port of the compressor 20 through the regulator 21 and the pneumatic circuit 24 as shown in FIGS. 16 and 17.

As described above, a pneumatic distributor 22 is installed to distribute the compressed air supplied from the compressor 20 through the regulator 21 to the vaccine dispenser 14 and the needle regulator 17, and the regulator 21 Pneumatic distributor 22 is installed on the relay board 23, the relay board 23 is also to perform the function of electrically connecting the regulator 21 and the dispenser controller (9).

On the other hand, the pneumatic circuit 24 is to control the inlet and discharge path of the compressed air through the needle regulator 17 to move the syringe body 15 in the up and down direction, the main controller 26 by the cable It is preferable to connect the needle 19 to the abdominal cavity of the fish, as well as to insert the needle 19 into the abdominal cavity of the fish can be performed by the needle adjuster (17).

In addition, when the needle 19 is inclined at an angle to be inserted into the abdominal cavity of the fish, even if the descending width of the Z-axis feed table 13b becomes somewhat large, the actual insertion depth of the needle 19 may not be so deep. This can minimize or prevent the situation where the needle 19 penetrates the body in the vertical direction and damages the body, and is required to move the needle 19 to the needle adjuster 17. The width can also be sufficiently secured to provide convenience according to the operation or control of the needle adjuster 17.

To this end, as shown in FIG. 15, the syringe body 15 and the needle regulator 17 of the vaccine dispenser 14 are installed at an angle to the Z-axis feeder 13b at an inclined angle by a predetermined angle, and the inclination angle is adjusted. An angle adjusting plate 18 is provided between the needle adjuster 17 and the Z-axis feeder 13b.

The angle adjusting plate 18 is fixedly installed with the needle adjuster 17 while being able to rotate relative to the Z-axis feeder 13b by a rotating shaft (not shown), and one side of the angle adjusting plate 18 (upper side in the drawing). An arc-shaped angle adjustment hole is formed at the end, and a stopper is installed at the angle adjustment hole so as to be engaged with the Z-axis feeder 13b and set according to an angle requiring an inclination of the angle adjustment plate 18.

Therefore, by angularly moving the angle adjusting plate 18 about an axis of rotation not shown in the drawing, the stopper 19 of the syringe body 15 for the vaccine dispenser 14 is set to the required angle, and then the stopper. When the angle adjusting plate 18 and the Z-axis feeder 13b are brought into close contact with each other by tightly tightening to the Z-axis feeder 13b, the insertion angle of the needle 19 can be set to the required angle. .

After the vaccination step (S6) as described above, while restoring the Cartesian coordinate robot (10) to the original position, using the fish transfer conveyor (3) to restore the system to discharge the fish vaccination completed to the outside and fish By going through the discharge step (S7), the automatic vaccine inoculation method according to the present invention is completed, and the telescopic type to safely discharge the fish into the water tank and the like not shown on the outlet side of the fish transport conveyor (3). It is preferable to connect and use the collapsible fish discharge plate 5.

All processes described above are automatically controlled by the main controller 26 forming the device controller 28 as shown in FIG. 16, and the main controller 26 controls the fish feed conveyor 3 and the Cartesian coordinate robot 10. It is connected to the robot controller 25 for controlling and the image processor 27 for implementing the body analysis vaccination algorithm, respectively, the operation switch 8 and the touch screen for setting the conditions necessary for the operation of the vaccination device (1) 8a and the dispenser controller 9 are connected to the main controller 26.

On the other hand, the robot controller 25 is connected to the conveyor driving unit 3d and the rectangular coordinate robot 10 of the fish transport conveyor 3 by a cable, the image processor 27 is the height sensor 3b and the camera 6 The planar image and the height data are transmitted from the screen, the image is displayed on the screen through the vision monitor 7, and the inoculation position followed by the body analysis vaccination algorithm is transmitted to the main controller 26. The robot controller 25 causes the cartesian robot 10 to operate.

10 to 17 shows a vaccination device (1) that can be optimally applied to the automatic vaccine vaccination method according to the present invention, most of the configuration is the automatic vaccine vaccination method according to the present invention, respectively It is mentioned in the step-by-step description, and the rest of the detailed configuration is based on the automated vaccination device for fish filed by the applicant of the Republic of Korea Patent Application No. 132748 (application date: November 22, 2012) in 2012 Please note.

The prior application is the first vaccination apparatus (1) only before the filing of the present application in consideration of the time according to the development of the fisheye analysis vaccination algorithm in the automatic vaccine vaccination method through the morphological image processing of the fish according to the present invention As filed, the device part for implementing the automatic vaccine inoculation method according to the present invention is mentioned in detail, and it is noted that this prior application was not disclosed before the present application.

Finally, as shown in Figure 10, when the fish conveying conveyor (3) is arranged in parallel in two rows along the feed direction of the fish, from the fish anesthesia and input step (S1) of the present invention The process up to the image analysis and inoculation position following step S4 is performed in parallel for each fish transfer conveyor 3 using the height sensor 3b and the camera 6 provided for each fish transfer conveyor 3. Separately), and the orthogonal coordinate robot setting step (S5) and vaccination step (S6) of the present invention, one orthogonal coordinate robot (10) in sequence with respect to the fish placed in each fish transfer conveyor (3) It is desirable to be carried out as.

The overall process of vaccination using the two rows of fish transport conveyor 3 as described above is carried out by the algorithm of the flowchart shown in FIG. 18, and the algorithm shown in FIG. 18 is also applied to one representative example of the present invention. It is not merely meant that the automatic vaccine vaccination method according to the present invention is performed by being limited to the algorithm of the flowchart shown in FIG. 18.

Briefly describing the 2-line vaccination process based on the flow chart of FIG. 18, first, the fish is subjected to a shallow anesthesia by immersion by using an anesthetic agent that is harmless to the fish. Put fish on the conveyor belt (3a) of the). -S1-

As described above, the fish is fed to the conveyor belt 3a of the 2-line fish conveying conveyor 3 to transport the fish, and in the process, each fish passes through the permeable laser displacement sensor, which is the height sensor 3b. The fish height is measured. -S2-

Then, by moving the fish to the position where the proximity sensor 3c is installed, the proximity sensor 3c of the No.1 fish conveying conveyor 3 and the proximity sensor 3c of the No.2 fish conveying conveyor 3 are moved. While recognizing the input state of the fish, and captures and captures the image of the fish with the CCD camera (6) installed on each of the fish conveying conveyor (3). -S3-

After measuring the body width and the full length of the fish through the captured image as described above, equations (4) and (5) based on the measured body width and the height data of the fish measured from the full-length data and the height sensor 3b. Calculate the coordinates corresponding to the inoculation position of the vaccine using, (6). -S4-

After the 3-axis Cartesian coordinate robot 10 is followed along the vaccination position with the coordinates of the vaccination position calculated as described above, the needle 19 is inserted into the abdominal cavity of the fish, and the vaccine dispenser 14 and the dispenser controller. Activate (9) to inject 0.1 ml of the vaccine into the abdominal cavity of the fish. -S5-and-S6-

Here, when fish are simultaneously introduced through the 2-line fish transport conveyor (3), the algorithm is designed so that vaccination can be preferentially performed to the fish transport conveyor (3) designated as No. 1, the inoculation is Upon completion, the Cartesian coordinate robot 10 returns to the origin, and each fish conveying conveyor 3 is driven to discharge the fish. -S7-

According to the vaccine automatic inoculation method of the present invention made in the above manner, it is possible to control the injection depth based on the height of the fish at the time of inoculation of the vaccine, as well as the conveyor system that can continuously feed and automatically discharge the fish in one direction By application, it is possible to prevent the damage of the fish body due to excessive injection depth and to improve the inoculation speed of the vaccine.

In particular, the vaccination location is tracked by a fisheye analysis vaccination algorithm, which reduces the capacity and image processing time of the database required for the vision system as much as possible, and allows accurate abdominal cavity even without additional user setup of fish information. The correct amount of vaccine can be administered at the site, thereby providing a more efficient and economical inoculation method.

In addition, the fish transport conveyor (3) is arranged in two rows to perform the image processing in parallel for each fish transport conveyor (3), the vaccination operation is to be performed sequentially by one Cartesian coordinate robot (10) In this case, unlike the servo-type inoculation method of the prior application, it is possible to save about 0.5 seconds for discharging the fish in the feeding direction after inoculation, and eliminate unnecessary work such as reoperation of the inoculation button after discharging. Therefore, the time required for vaccination can be further shortened, and the treatment of vaccinated fish can be performed more easily.

In addition, the needle 19 of the syringe body 15 can be inserted at an angle through the abdominal cavity of the fish, and at the same time, the needle 19 can be quickly removed from the abdominal cavity of the fish using a needle regulator 17, which is a pneumatic cylinder. By doing so, it is possible to provide an additional function, such as to prevent the damage or damage of the fish body more reliably by blocking the phenomenon that the fish is caught up in the state of being plugged into the injection needle (19).

1 Vaccination vaccination device 2 Device main body 3 Fish transport conveyor
3a: Conveyor belt 3b: Height sensor 3c: Proximity sensor
3d: Conveyor Drive 3e: Drain Chamber 3f: Drain
4: inoculation waiting table 5: fish discharge plate 6: camera
7: Vision monitor 8: Operation switch 8a: Touch screen
9: dispenser controller 10: Cartesian robot 11: X-axis robot
11a: guide rail 11b: X-axis feeder 12: Y-axis robot
12a: guide rail 12b: Y-axis feeder 13: Z-axis robot
13a: guide rail 13b: Z-axis feeder 14: vaccine dispenser
15: syringe body 15a: pneumatic line 15b: pneumatic discharge port
16: Vaccine storage container 16a: pneumatic line 17: needle regulator
17a: pneumatic line 18: angle adjustment plate 19: needle
20: compressor 21: regulator 22: pneumatic distributor
23: relay board 24: pneumatic circuit 25: robot controller
26: main controller 27: image processor 28: device control unit
29: Pneumatic line SV: Solenoid valve RV: Relief valve
F: Air filter θ: Shooting range

Claims (9)

Fish anesthesia and input step (S1) for injecting the anesthetized fish using an anesthetic using a fish transport conveyor,
After the fish anesthesia and input step (S1), while transferring the fish to the fish transport conveyor, the fish transport and height measuring step (S2) for measuring the height of the fish with a height sensor installed on the side of the fish transport conveyor,
After the fish feed and height measurement step (S2), the plane image taking step (S3) for taking a plane image of the fish with a camera installed on the upper side of the fish conveying conveyor, and
After the plane image photographing step (S3), the body width and the full length of the fish are measured on the basis of the plane image taken by the camera, and the body height measured from the plane image and the height of the fish measured by the total length and the height sensor. Image analysis and inoculation position following step (S4) to determine the insertion position and the insertion depth of the injection needle,
After the image analysis and inoculation position following step (S4), using an orthogonal coordinate robot consisting of the X-axis robot, Y-axis robot and Z-axis robot to set the Cartesian coordinate robot to insert the needle of the syringe body into the abdominal cavity of the fish Step S5,
After the orthogonal coordinate robot setting step (S5), the vaccination step (S6) for allowing the vaccine stored in the vaccine storage container is injected into the abdominal cavity of the fish through the syringe body and the needle,
After undergoing the vaccination step (S6), while returning the Cartesian coordinate robot to its original position, using a fish transfer conveyor to restore the fish vaccination completed to the outside of the system and the fish discharge step (S7) that is made Automatic vaccine inoculation method through the morphological image processing of the fish characterized by. The method of claim 1, wherein the image analysis and inoculation position following step (S4) is applied to the fish body analysis algorithm for measuring the shape of the fish,
The body analysis algorithm may include: capturing an image frame photographed by a camera (S41), converting the captured image into an 8-bit grayscale image (S42), and adjusting a contrast ratio generated in the image conversion step. (S43), converting the contrast-adjusted image into a black and white image by binarization (S44), and processing the image of the fish morphologically using the binarized black and white image (S45); Automated vaccine inoculation method through the morphological image processing of the fish, characterized in that comprising the step (S46) of measuring the body width and the full length of the fish as an image. The method of claim 2, wherein the following method of inoculation position is applied to the image analysis and inoculation position following step (S4), and the body width of the fish measured by the fish body analysis algorithm and the height of the fish measured by the total length and the height sensor. It is a method of following the inoculation position from the corresponding data in comparison with the pre-stored data by the one-dimensional regression analysis method,
The pre-stored data by the one-dimensional regression analysis method, the body width, length and height of the fish measured from the images taken by the camera on the upper and side portions of the fish as an independent variable, the abdominal position where the actual vaccination of the fish is made As a dependent variable, the automatic vaccine vaccination method through the morphological image processing of the fish, characterized in that the correlation between the independent variable and the dependent variable is derived as a regression coefficient. According to claim 3, wherein the data about the height of the fish among the data by the one-dimensional regression analysis method, the inoculation position of the vaccine confirmed from the two-dimensional planar image taken by the camera from the top of the fish to the circular surface of the fish body The white paper is displayed as a dependent variable, and the height value measured by irradiating the beam of the laser displacement sensor to the position indicated by the white paper is used as a dependent variable. The automatic vaccine vaccination method through morphological image processing of fish, characterized in that the correlation between the independent variable and the dependent variable is derived as a regression coefficient. The process according to any one of claims 1 to 4, wherein the process from the fish anesthesia and feeding step (S1) to the image analysis and inoculation position following step (S4) is arranged in parallel along the transport direction of the fish. It is carried out in parallel for each fish transfer conveyor using a two-row fish transfer conveyor, and a height sensor and a camera provided for each fish transfer conveyor,
The orthogonal coordinate robot setting step (S5) and the vaccination step (S6) is carried out in sequence with respect to the fish placed in each fish transport conveyor with one orthogonal coordinate robot through the morphological image processing of the fish Automatic vaccine vaccination method. The method of any one of claims 1 to 4, wherein in the vaccination step (S6) through the control of the compressed air discharged from the compressor through the regulator, the supply of the vaccine and the valve operation of the syringe body in a liquid quantitative discharge method Vaccine automatic inoculation method through the morphological image processing of the fish, characterized in that to be carried out. According to claim 6, After the vaccination step (S6), the process of removing the needle of the syringe body from the abdominal cavity of the fish, is performed by a needle adjuster installed on the z-axis robot of the Cartesian coordinate robot with the syringe body ,
The needle regulator is a plate-type pneumatic cylinder using the compressed air discharged through the regulator from the compressor, the system restoration and fish discharge step (S7) is performed after the needle exits the abdominal cavity of the fish by the needle regulator Vaccine automatic inoculation method through the morphological image processing of the fish, characterized in that. The method of claim 6, wherein the Cartesian coordinate robot setting step (S5) is automatic injection of the vaccine through the morphological image processing of the fish, characterized in that the needle of the syringe body is inserted obliquely inclined at a predetermined angle through the abdominal cavity of the fish Inoculation method. [8] The method of claim 7, wherein the rectangular coordinate robot setting step (S5) is automatic injection of the vaccine through the morphological image processing of the fish, characterized in that the needle of the syringe body is inserted obliquely inclined at a predetermined angle through the abdominal cavity of the fish Inoculation method.

Images