KR101106967B1 - Automatic system for measuring character of fishes and method in using same - Google Patents

Abstract

The present invention relates to an automated system and method for measuring fish-measured traits, and the technical problem to be solved is to provide a fish-measurement trait measuring system and method for improving the economics by reducing the time and personnel required to measure the fishes measurement traits have.

To this end, the fish measurement and trait measurement automation system according to the present invention includes a conveyor image photographing unit for photographing and transmitting a front image and a side image of a target fish placed on a conveyor belt, and a weight of the target fish moved from the conveyor image photographing unit. Measuring the measurement characteristics of the target fish by the conveyor weight measuring unit for measuring and transmitting the front image and the side image of the target fish, and measuring the obesity degree of the target fish by the weight of the target fish, image information, measurement of the target fish Characteristic trait measurement unit for storing and managing the trait information, weight information and obesity information in real time.

Fish measurement characteristics, imaging, weighing, obesity, image analysis module

Description

AUTOMATIC SYSTEM FOR MEASURING CHARACTER OF FISHES AND METHOD IN USING SAME

The present invention relates to systems and methods for automated measurement of fish-measurement traits, and to systems and methods for automated measurement of fish-measurement traits.

Conventionally, in measuring the measurement characteristics of fish, the total length, body length, height, etc. were measured using plastic ruler, the weight was measured using a scale, and the data were managed by writing and inputting the measured data on the record sheet. .

Specifically, the number of personnel required for the conventional data input and management method was 12, and the time required to measure 10,000 fish was 333 hours or more, and the required number of measurement traits was measured by manually measuring each fish. There was a problem that the required time is increased.

In addition, the method of transmitting and recording the measured trait measurement information in the conventional word of mouth has a problem that the accuracy of the measurement data is low because the frequency of error of the data is high.

Therefore, there is a need for a fish measurement trait measuring system which can reduce the required number of people and the time required by automating the measurement of fish measurement traits and improve accuracy by automatically database.

The present invention has been made to solve the above problems, and an object of the present invention is to provide a fish measuring trait measuring automation system and method for reducing the time and personnel required for measuring the measuring trait of fish to improve the economics.

Another object of the present invention is to provide an automatic fish measurement and trait measurement system and method for improving the accuracy of data by databaseting image data and measurement data of fish in real time.

In order to achieve the object as described above, the fish measurement trait measurement automation system according to the present invention is a conveyor image pickup unit for photographing and transmitting the front image and the side image of the target fish placed on the conveyor belt, and moving in the conveyor image capture unit A measurement weight of the target fish by a conveyor weight measuring unit for measuring and transmitting the weight of the target fish and the front and side images of the target fish, and measuring the degree of obesity of the target fish by the weight of the target fish And a measurement trait measuring unit configured to store and manage image information, measurement trait information, weight information, and obesity information of the target fish in real time.

The conveyor imaging apparatus may further include a base part including a plurality of support members, a conveyor belt installed on the base part to move the target fish to the conveyor weight measuring unit, and an upper portion of the conveyor belt. The front image photographing unit for photographing the front image of the side and the side of the conveyor belt may include a side image photographing unit for photographing the side image of the target fish.

In addition, the conveyor imaging unit may further include a limit switch for driving the conveyor belt to move the target fish to the conveyor weight measuring unit.

The front image capturing unit and the side image capturing unit may transmit the front image and the side image of the target fish to the measurement trait measuring unit using an IEEE 1394 port, respectively.

In addition, the conveyor weight measuring unit is a base portion consisting of a plurality of support members, an inlet conveyor belt installed on the base portion for introducing the target fish, and installed on the base portion, but continuously installed on the inlet conveyor belt And a weight measuring unit configured to measure the weight of the target fish positioned on the weighing conveyor belt to move the target fish to the weighing position.

The weight measuring unit may transmit the weight of the target fish to the measurement trait measuring unit through RS232 communication.

In addition, the measurement quality measurement unit may include an image analysis module for measuring the measurement characteristics of the target fish by the front image and the side image of the target fish.

The image analysis module may measure the full length, the body length, the head length, the height, the length of disease, and the length of disease of the target fish based on the front image, and the body width of the target fish based on the side image.

In addition, the image analysis module may be composed of a Python PIL or Numpy module.

The fish measurement trait measurement automation system may further include a measurement trait storage unit for storing image information, measurement trait information, weight information, and obesity information of the target fish.

In addition, according to the present invention, there is provided a method for automating measurement of fish traits by measuring image information and transmitting a front image and a side image of a target fish, a weight information transmission step of measuring and transmitting a weight of the target fish, and the target. A measurement trait measuring step of measuring the measurement trait of the target fish by using the front and side images of the fish; an obesity measurement step of measuring the degree of obesity of the target fish by the weight of the target fish; and image information and measurement trait of the target fish Measurement information storage step of storing and managing the information, weight information and obesity information in real time may include.

In addition, the measuring trait measurement step is to analyze the front image to recognize the full length, height, head, head, and disease of the target fish and the front image recognition process to analyze the side image to recognize the body width of the target fish It may include a side recognition image recognition process and a marker recognition process for calculating the actual length value per pixel for calculating the size of the target fish.

In addition, the front image recognition process and the side image recognition process may include a preprocessing process for distinguishing the image and the background portion of the target fish.

In addition, the front image recognition process may include a full length recognition process for recognizing the full length head and full length tail of the target fish.

In addition, the front image recognition process may include a height recognition process for recognizing the height of the upper fin and the height of the lower fin of the target fish.

In addition, the front image recognition process may include a two-head recognition process for recognizing the gills of the target fish.

In addition, the front image recognition process may include a non-sickness recognition process for recognizing the non-sickness of the target fish.

In addition, the front image recognition process may include a diseased bottle recognition process for recognizing the diseased bottle of the target fish.

The side image recognition process may include a body width recognition process of recognizing a body width by recognizing an upper highest point and a lower lowest point of the target fish.

As described above, according to the automated fish measurement and trait measurement system and method according to the present invention, there is an effect that it is possible to reduce the time and personnel required to measure the measurement trait of fish to improve the economics.

In addition, it is possible to increase the accuracy of the data by database the image information and the measurement data of the fish in real time.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. First, it should be noted that the same components or parts in the drawings represent the same reference numerals as much as possible. In describing the present invention, detailed descriptions of related well-known functions or configurations are omitted in order not to obscure the gist of the present invention.

1 is a block diagram of an automated fish measurement trait measurement system according to an embodiment of the present invention.

As shown in FIG. 1, the fish measurement trait measuring system according to an embodiment of the present invention includes a conveyor imaging unit 100, a conveyor weight measuring unit 200, and a measuring trait measuring unit 300. .

The conveyor image photographing unit 100 photographs and transmits a front image and a side image of a target fish placed on a conveyor belt.

Figure 2a is a front view of the conveyor imaging unit of the fish measurement trait measurement automation system according to an embodiment of the present invention, Figure 2b is a plan view.

As shown in FIGS. 2A and 2B, the conveyor image capturing unit 100 includes a base 110, a conveyor belt 120, a front image capturing unit 130, and a side image capturing unit 140. Include.

The base unit 110 is composed of a plurality of support members.

The conveyor belt 120 is installed on the base unit 110 to move the target fish to the conveyor weight measuring unit 200. At this time, the target fish is preferably placed on the conveyor belt 120 in an anesthetic state.

The front image capturing unit 130 is installed on the conveyor belt 120 to photograph the front image of the target fish.

In this case, the front image capturing unit 130 may be an image camera, and may transmit the front image of the target fish to the measurement trait measuring unit 300 using an IEEE 1394 port.

The side image photographing unit 140 is installed at the side of the conveyor belt 120 to take a side image of the target fish.

In this case, the side image capturing unit 140 may be formed of an image camera in the same manner as the front image capturing unit 130, and the measurement trait measuring unit 300 uses the IEEE 1394 port for the side image of the target fish. Can be sent to.

Meanwhile, the conveyor image photographing unit 100 may further include a limit switch 150 as shown in FIG. 2B.

The limit switch 150 drives the conveyor belt 120 to move the target fish to the conveyor weight measuring unit 200.

Figure 3a is a front view of the conveyor weight measuring unit of the fish measurement trait measurement automation system according to an embodiment of the present invention, Figure 3b is a plan view.

The conveyor weight measuring unit 200 measures and transmits the weight of the target fish moved by the conveyor image capturing unit 100.

3A and 3B, the conveyor weight measuring unit 200 includes a base 210, an incoming conveyor belt 220, a weighing conveyor belt 230, and a weighing unit 240. do.

The base portion 210 is composed of a plurality of support members.

The inlet conveyor belt 220 is installed on the base portion 210 to introduce the target fish.

In detail, the inlet conveyor belt 220 introduces the target fish moved through the conveyor belt 120 of the conveyor image photographing unit 100 into the conveyor weight measuring unit 200.

The metering conveyor belt 230 is installed on the base portion 210, and is continuously installed on the inlet conveyor belt 220 to move the target fish to the weighing position.

The weight measuring unit 240 measures the weight of the target fish located on the weighing conveyor belt 230.

In this case, the weight measuring unit 240 may transmit the weight of the target fish to the measurement trait measuring unit 300 using RS232 communication.

The measurement trait measuring unit 300 includes an image analysis module (not shown) for measuring the measurement trait of the target fish by the front image and the side image of the target fish.

The image analysis module (not shown) may measure the full length, the body length, the head length, the height, the length of the disease, and the length of the diseased fish using the front image, and the body width of the fish based on the side image. .

In this case, the image analysis module may be composed of a Python PIL or Numpy module.

Figure 4 is a program screen configuration of the measurement trait measurement unit of the fish measurement trait measurement automation system according to an embodiment of the present invention.

In addition, the measurement trait measuring unit 300 may measure the degree of obesity of the target fish by the weight of the target fish, as shown in Figure 4, the image information, the measurement trait information and the weight of the target fish Information and obesity information can be stored and managed in real time.

Meanwhile, although not shown, the automated fish measurement trait measurement system according to an embodiment of the present invention may further include a measurement trait storage unit (not shown).

The measurement trait storage unit (not shown) may store the image information, measurement trait information, weight information and obesity information of the target fish in real time by the measurement trait measurement unit 300 to increase the accuracy of the data.

Specifically, the accuracy of the repeatedly measured data can be verified in the following [Table 1].

Hereinafter, a method for automating fish measurement trait measurement according to an embodiment of the present invention will be described in detail.

5 is a block diagram of a method for automating fish measurement trait measurement according to an embodiment of the present invention.

As shown in FIG. 5, the method for automating fish measurement trait measurement according to an embodiment of the present invention includes an image information transmission step S100, a weight information transmission step S200, a measurement trait measurement step S300, and Obesity measurement step (S400) and the measurement information storage step (S500).

The image information transmitting step (S100) is a step of photographing and transmitting the front image and the side image of the target fish.

Specifically, in the image information transmitting step (S100), as shown in Figures 1 to 2b, the conveyor belt through the front image capture unit 130 and the side image capture unit 140 of the conveyor image capture unit 100. After photographing the target fish placed on the 120 to obtain a front image and a side image of the target fish, the front image and the side image is transmitted to the measurement trait measuring unit 300.

The weight information transmitting step (S200) is a step of measuring and transmitting the weight of the target fish.

Specifically, in the weight information transmitting step (S200), as shown in Fig. 1, 3a and 3b, the conveyor imaging unit 100 through the weight measuring unit 240 of the conveyor weight measuring unit 200 Measure the weight of the target fish moved in and transmits to the measurement trait measuring unit (300).

The measurement trait measurement step (S300) is a step of measuring the measurement trait of the target fish by the front image and the side image of the target fish.

Figure 6 is a block diagram of the measurement trait measurement step of the automated fish measurement trait measurement method according to an embodiment of the present invention.

The metrology trait measurement step S300 includes a front image recognition step S310, a side image recognition step S320, and a marker recognition step S330 as shown in FIG. 6.

The front image recognition process (S310) is a process of analyzing the front image to recognize the full length, height, head, tail, and tail of the target fish.

7 is a block diagram of the front image recognition process of the measurement trait measurement step in the automated fish measurement trait measurement method according to an embodiment of the present invention.

As illustrated in FIG. 7, the front image recognition process S310 may include a preprocessing process S311 for distinguishing an image of the target fish from a background portion.

9A and 9B are diagrams illustrating a preprocessing process of a front image recognition process.

In this case, (1) of FIG. 9A is a photographed image, (2) of FIG. 9A and (1) of FIG. 9B are clear images, and (2) of FIG. 9B is an image cropped on the basis of an outer region.

In the preprocessing process (S310), as illustrated in FIG. 9A, the image may be clearly manipulated by manipulating data to clearly distinguish the target fish and the background portion of the front image.

Specifically, first, the image of the outer region that is not necessary for the target fish recognition is removed, converted to a gray image, and then the gray average T of the predetermined region A that does not overlap with the target fish image is obtained in the left region of the image. The gray values larger than the gray mean (T) are converted to 255 gray values over the entire image.

Then, by performing the same process for the upper (B), the lower (D), the right (C) as in A, the image can be clearly distinguished from the target fish image and the background image as a whole.

In addition, in the preprocessing step S311, as shown in FIG. 9B, the farthest side image of the target fish may be recognized and the image may be cut based on the area for use in actual image processing.

In detail, first, the boundary coordinates of the fish image and the background image of the outer target fish are scanned and scanned from left to right within the area A, and then the smallest y value is selected among the coordinates, Scanning from the top to the bottom, finding and saving the boundary coordinates of the fish image and the background image, and selecting the largest value among the coordinates.

In addition, after scanning from the left to the right within the C region range, the boundary coordinates of the fish image and the background image of the outer shell are found and stored, and then the value of the largest y value is selected among the coordinates. After scanning with the border coordinates of the fish image and the background image of the outer target, the x value is selected among the coordinates.

Thereafter, the preprocessing process may be terminated by cutting the image around the recognized outer region.

Meanwhile, as illustrated in FIG. 7, the front image recognition process S310 may include a full length recognition process S312 for recognizing the full length head and the full length tail of the target fish.

10A and 10B illustrate a full-length recognition process of the front image recognition process.

10A is a diagram illustrating a state of recognizing the full length head, and FIG. 10B is a diagram illustrating a state of recognizing the full length tail.

In the full-length recognition process (S312), as shown in Figure 10a, after processing the image for accurate head recognition of the target fish, the head of the full length is recognized.

Specifically, first, the head image (A) of 10% of the total length is received from the target fish head, and then the head vertical portion is cut out as shown in B. FIG.

Then, binarization is performed to produce a black-and-white image like C and scan from top to bottom to find the coordinates of the leftmost boundary area.

In addition, in the full-length recognition process (S312), as shown in Figure 10b, after processing the image for accurate fish tail recognition, the tail of the full length is recognized.

Specifically, first, after receiving the tail image (A) of 10% of the entire length from the target fish tail, and cut off the vertical portion of the tail as shown in B.

Then, binarization is performed to make a black and white image as in C. Binary erosion is performed to fill the white part of the tail part with black, and scan from top to bottom to obtain the coordinates of the rightmost boundary area.

Meanwhile, as shown in FIG. 7, the front image recognition process S310 may include a height recognition process S313 for recognizing a height above a fin boundary and a height below a fin boundary of the target fish.

11A and 11B illustrate a height recognition process in the front image recognition process.

At this time, Figure 11a is a diagram showing a state of recognizing the height of the upper fin border, Figure 11b is a diagram showing a state of recognizing the height of the upper fin.

In the height recognition process (S313), as shown in FIG. 11A, an image of 20% from the upper end of the target fish is cut out in (1), and the start (A) and the end (B) of the fin in (2). ), And binarization is performed based on the gray mean value of the yellow area in (3).

Then, find the central coordinates (C) of the two points A and B in the binarized image, scan them from the point C to the left and read the upper fin boundary line coordinates, and read the upper fin boundary line coordinates from the C point to the right. Save it.

Then, the distance between the previous coordinate and the next coordinate among the stored coordinates is obtained and aligned, and some of the largest and smallest distances are removed.

Then, the average value y1 of the largest value y2 and the smallest value y3 of the stored coordinates is obtained, and the difference between the boundary coordinates of the three values y1, y2, and y3 is the most. Choose a small value.

In addition, in the height recognition process (S313), as shown in Figure 11b, in (1) to cut the image up to 20% from the lower end of the target fish, in (2) the start (A) and the end of the fin Find (B) and perform binarization on the basis of gray mean value of yellow area in (3).

Then, find the central coordinates (C) of the two points A and B in the binarized image, scan them from the C point to the left and read the lower fin boundary line coordinates, and read the lower fin boundary line coordinates to the right from the C point. Save it.

Then, the distance between the previous coordinate and the next coordinate among the stored coordinates is obtained and aligned, and some of the largest and smallest distances are removed.

Then, the average value y1 of the largest value y2 and the smallest value y3 of the stored coordinates is obtained, and the difference between the boundary coordinates of the three values y1, y2, and y3 is the most. Choose a small value.

On the other hand, the front image recognition process (S310) may include a two-head recognition process (S314) for recognizing the gills of the target fish, as shown in FIG.

12 is a diagram illustrating a two-head recognition process of the front image recognition process.

In the two-head recognition process (S314), as shown in FIG. 12, a histogram equalization is performed as shown in (2), and binarization is performed as shown in (3). Perform binary erosion as in (4).

Then, calculate the coordinates of the cx point by obtaining the coordinates of the sy1, sy2 points at the left most, but calculate the number of x1 coordinates and the white points with the most white points starting from the center of cx to the L point to the left, and starting the cx center. Calculate the number of x2 coordinates and the number of white points with the most white points from point R to the right.

If the white area is not found between L and R, the cx value is finally selected. If x1 coordinates are found, the x1 value is finally selected. If x2 coordinates are found, the x2 value is finally selected. If both are present, the x1 value is finally selected.

Meanwhile, as illustrated in FIG. 7, the front image recognition process S310 may include a non-sickness recognition process S315 for recognizing a non-sickness of the target fish.

FIG. 13 is a diagram illustrating an unsigned disease recognition process in the front image recognition process.

In step S315, as shown in FIG. 13, the tail image shown in (1) is binarized as shown in (2), and then binary erosion is performed as shown in (3).

Then, after saving the boundary coordinates of the part A, select the value of the largest y value at the point A as the H1 point coordinates, and save the boundary coordinates of the part B, and then change the y value at the B point. Select the smallest value as the H2 point coordinate. Here, H2-H1 is non-ill.

Meanwhile, as illustrated in FIG. 7, the front image recognition process S310 may include a disease free bottle recognition process S316 for recognizing the disease free bottle of the target fish.

Since the diseased portion is difficult to calculate due to the characteristics of the image, after calculating the ratio of the total length and the disease free on the basis of the previously measured data, after calculating the proportional constant value based on the calculation [Equation 1] And [Equation 2] can predict the point.

Armed Forces Location = 0.867290183093 × Battlefield

Armory Value = Battlefield-Armor Location

The side image recognition process (S320) is a process of recognizing the body width of the target fish by analyzing the side image.

8 is a block diagram of the side image recognition process of the measurement trait measurement step in the automated fish measurement trait measurement method according to an embodiment of the present invention.

As shown in FIG. 8, the side image recognition step S320 may include a preprocessing step S321 for distinguishing an image of the target fish from a background part.

14 is a diagram illustrating a preprocessing process of a side image recognition process.

In the preprocessing step S321, as illustrated in FIG. 14, the image may be clearly manipulated by manipulating data to clearly distinguish the target fish and the background portion of the side image.

Specifically, first, the data of the upper boundary line of the target fish is read and the gray value is averaged, and then a value larger than the gray value is set to 255 based on the gray average value for the entire image.

Subsequently, by performing the same process as described above, the image of the fish image and the background image as a whole can be clearly distinguished.

Meanwhile, as illustrated in FIG. 8, the side image recognition process S320 may include a body width recognition process S322 for recognizing a body width by recognizing an upper highest point and a lower lowest point of the target fish.

15A and 15B illustrate a body width recognition process in a side image recognition process.

In this case, FIG. 15A is a diagram illustrating recognition of the uppermost point, and FIG. 15B is a diagram illustrating recognition of the lowermost point.

In the body width recognition process (S322), as shown in FIG. 15A, an uppermost peak may be recognized for body width recognition.

Specifically, in the body width recognition process (S320), first, the values of the y coordinates are stored while scanning from S1 to S2, and the coordinates whose y values are significantly changed compared to the previous y values are removed.

Then, the maximum value max_y and the minimum value min_y and the average value aver_y of the two values are obtained.

At this time, if the absolute value of the value of (max_y-min_y) is greater than 10 and the difference is more than 60, select max_y.If the absolute value of the value of (max_y-min_y) is greater than 10 and the difference is less than 60, select min_y. , If the absolute value of (max_y-min_y) is less than 10, select aver_y.

Then, the highest point T1 is determined based on the above three conditions.

On the other hand, the body width recognition process (S322), as shown in Figure 15b, can recognize the lowermost point for the body width recognition.

Specifically, in the body width recognition process (S322), first, the values of the y coordinates are stored while scanning from S1 to S2, and the coordinates where the y values are significantly changed compared to the previous y values are removed.

Then, after recognizing the average value of the stored y values as the lowest point (T2), the body width is obtained as the value T2-T1.

The marker recognition process S330 is a process of calculating an actual length value per pixel for calculating the size of the target fish.

16A and 16B are diagrams illustrating a marker recognition process in a measurement trait measurement step in the fish measurement trait measurement automation method according to an embodiment of the present invention.

16A is a view showing a side marker, and FIG. 16B is a view showing a front marker.

In order to calculate the actual size of the fish, the actual length value per pixel must be calculated.

In the marker recognition process (S330), as shown in FIGS. 16A and 16B, a marker may be manufactured to obtain an image of the marker.

Specifically, as shown in FIG. 16A, the number of pixels corresponding to the height fdh of the side marker is counted, and as shown in FIG. 16B, the width (fdh) and vertical (fdh) of the front marker are shown. Count the number of pixels

At this time, the actual height (10 cm) / fdh of the side marker is the length per pixel for the height in the side image, the width (10 cm) / fdh of the front marker is the length per pixel for the width in the front image, the front marker The vertical width (3 cm) of f dh is the length per pixel for the vertical width in the front image.

The obesity measurement step (S400) is a step of measuring the degree of obesity of the target fish by the weight of the target fish.

Specifically, in the obesity measurement step (S400), as shown in Figure 1, the measurement trait measuring unit 300 measures the degree of obesity of the target fish by the weight of the target fish.

The measurement information storing step (S500) is a step of storing and managing image information, measurement trait information, weight information, and obesity information of the target fish in real time.

Specifically, in the measurement information storage step (S500), the measurement trait measuring unit 300 real-time the image information, measurement trait information, weight information and obesity information of the target fish to the measurement trait storage unit (not shown) Manage by storing it.

As described above with reference to the drawings illustrating a fish measurement and trait measurement automation system and method according to the present invention, the present invention is not limited by the embodiments and drawings disclosed herein, the scope of the technical spirit of the present invention Of course, various modifications can be made by those skilled in the art.

1 is a block diagram of an automated fish measurement trait measurement system according to an embodiment of the present invention.

Figure 2a is a front view of the conveyor imaging unit of the fish measurement trait measurement automation system according to an embodiment of the present invention.

Figure 2b is a plan view of the conveyor imaging unit of the fish measurement trait measurement automation system according to an embodiment of the present invention.

Figure 3a is a front view of the conveyor weight measuring unit of the fish measurement trait measurement automation system according to an embodiment of the present invention.

Figure 3b is a plan view of the conveyor weight measuring unit of the fish measurement trait measurement automation system according to an embodiment of the present invention.

Figure 4 is a program screen configuration of the measurement trait measurement unit of the fish measurement trait measurement automation system according to an embodiment of the present invention.

5 is a block diagram of a method for automating fish measurement trait measurement according to an embodiment of the present invention.

Figure 6 is a block diagram of the measurement trait measurement step of the method for automatic measurement of fish measurement trait measurement according to an embodiment of the present invention

Figure 7 is a block diagram of a front image recognition process of the measurement trait measurement step in the fish measurement trait measurement automation method according to an embodiment of the present invention.

Figure 8 is a block diagram of the side image recognition process of the measurement trait measurement step in the automated fish measurement trait measurement method according to an embodiment of the present invention.

9A and 9B are diagrams illustrating a preprocessing process in a front image recognition process.

10A and 10B illustrate a full length recognition process of a front image recognition process;

11A and 11B illustrate a height recognition process in a front image recognition process;

12 is a diagram illustrating a two-sheet recognition process of the front image recognition process.

FIG. 13 is a diagram illustrating an unsigned disease recognition process in a front image recognition process; FIG.

14 is a diagram illustrating a preprocessing process in a side image recognition process.

15A and 15B are diagrams illustrating a body width recognition process of a side image recognition process;

Figures 16a and 16b is a view showing a marker recognition process of the measurement trait measurement step in the automated fish measurement trait measurement method according to an embodiment of the present invention.

<Description of Symbols for Main Parts of Drawings>

100: conveyor video photographing unit 110,210: base part

120: conveyor belt 130: front image photographing part

140: side image photographing unit 150: limit switch

200: conveyor weight measuring unit 220: incoming conveyor belt

230: weighing conveyor belt 240: weight measuring unit

300: measurement trait measuring unit

S100: video information transmission step S200: weight information transmission step

S300: measurement trait measurement step S310: front image recognition process

S311, S321: Preprocessing S312: Field Recognition

S313: Height Recognition Process S314: Head Recognition Process

S315: Disease Recognition Process S316: Disease Recognition Process

S320: Side image recognition process S322: Body width recognition process

S330: Marker recognition process S400: Obesity measurement step

S500: step of storing measurement information

Claims (19)

A conveyor image photographing unit for photographing and transmitting a front image and a side image of a target fish placed on a conveyor belt; A conveyor weight measuring unit configured to measure and transmit a weight of the target fish moved from the conveyor image photographing unit; The measurement trait of the target fish is measured by the front image and the side image of the target fish, the obesity degree of the target fish is measured by the weight of the target fish, the image information, the measurement trait information, the weight information and the obesity degree of the target fish. Fish measurement trait measurement automation system comprising a measurement trait measuring unit for storing and managing information in real time. The method of claim 1, The conveyor imaging unit, A base portion comprising a plurality of support members; A conveyor belt installed on the base to move the target fish to the conveyor weight measuring unit; A front image photographing unit installed on the conveyor belt to photograph a front image of the target fish; And And a side image photographing unit installed at a side of the conveyor belt to capture a side image of the target fish. 3. The method of claim 2, The conveyor imaging unit, And a limit switch for driving the conveyor belt to move the target fish to the conveyor weighing unit. 3. The method of claim 2, The front image capturing unit and the side image capturing unit, The fish measurement trait measuring system, characterized in that for transmitting the front image and the side image of the target fish to the measurement trait measuring unit using the IEEE 1394 port, respectively. The method of claim 1, The conveyor weight measuring unit, A base portion comprising a plurality of support members; An inlet conveyor belt installed on the base to introduce the target fish; A metering conveyor belt installed on the base part and continuously installed on the inlet conveyor belt to move the target fish to a metering position; And And a weight measuring unit for measuring a weight of the target fish on the weighing conveyor belt. The method of claim 5, The weight measuring unit, Fish measuring trait measurement automation system, characterized in that for transmitting the weight of the target fish to the measurement trait measuring unit using RS232 communication. The method of claim 1, The measurement quality measuring unit, And an image analysis module for measuring the measurement trait of the target fish by using the front image and the side image of the target fish. The method of claim 7, wherein The image analysis module, Measuring the full length, body length, head length, body height, head length and tail length of the target fish using the front image, Fish measuring trait measurement automation system characterized in that for measuring the body width of the target fish by the side image. The method of claim 7, wherein The image analysis module is a fish measurement trait measurement automation system, characterized in that consisting of Python PIL or Numpy module. The method of claim 1, And a measurement trait storage unit for storing image information, measurement trait information, weight information, and obesity information of the target fish. Image information transmission step of photographing and transmitting the front image and the side image of the target fish; A weight information transmission step of measuring and transmitting the weight of the target fish; A measurement trait measurement step of measuring the measurement trait of the target fish by using the front image and the side image of the target fish; An obesity degree measuring step of measuring an obesity degree of the target fish by the weight of the target fish; And And a measurement information storage step of storing and managing image information, measurement trait information, weight information, and obesity information of the target fish in real time. The method of claim 11, The measurement trait measurement step, A front image recognition process of analyzing the front image and recognizing a full length, a height, two heads, a diseased bottle and a bottled fish of a target fish; A side image recognition process of recognizing a body width of a target fish by analyzing the side image; And And a marker recognition process of calculating an actual length value per pixel for calculating the size of the target fish. The method of claim 12, The front image recognition process and the side image recognition process fish measurement trait measurement automation method characterized in that it comprises a pre-processing process for distinguishing the image and the background portion of the target fish. The method of claim 12, The front image recognition process, And a full length recognition process for recognizing the full length head and the full length tail of the target fish. The method of claim 12, The front image recognition process, And a fish height recognition process for recognizing a fish height upper fin boundary and a fish height lower fin boundary of the target fish. The method of claim 12, The front image recognition process, A fish measurement and trait measurement automation method comprising a two-head recognition process for recognizing the gills of the target fish. The method of claim 12, The front image recognition process, An automatic measurement method for measuring fish traits, characterized in that it comprises a non-lesight recognition process for recognizing the unsigned diseases of the target fish. 15. The method of claim 14, The front image recognition process, A method for automating fish measurement and trait measurement, characterized in that it comprises a diseased bottle recognition process for recognizing the diseased bottle of the target fish. The method of claim 12, The side image recognition process, And a body width recognition process of recognizing a body width by recognizing an upper highest point and a lower lowest point of the target fish.

Images