KR102297317B1 - Livestock weighing device and method through 3D image processing - Google Patents

Abstract

The present invention relates to a device and a method for measuring the weight of livestock through three-dimensional (3D) image processing, thereby enhancing measuring precision. In accordance with the present invention, the device for measuring the weight of livestock through 3D image processing comprises: a photographing unit for obtaining a 3D image of livestock; and a measuring module for processing the 3D image to measure the weight, wherein the measuring module includes: a pre-processing unit for extracting a central point and a central distance from the 3D image to remove a background and an outlier; a normalization unit for extracting and then removing a bottom plane from the pre-processed 3D image, grouping objects in accordance with density of points to extract a group of the livestock, and using a plane equation to correct a slope; a factor analysis unit for extracting a factor for one or more of length, height, circumference, an area, and a volume from the normalized 3D image; and a weight calculation unit using the extracted factor to calculate the weight.

Description

Livestock weighing device and method through 3D image processing

The present invention relates to an apparatus and method for measuring livestock weight through 3D image processing, and more particularly, by measuring the weight of livestock by processing a 3D image obtained through a photographing unit, and then analyzing a factor necessary for weight calculation, thereby measuring accuracy relates to an apparatus and method for weighing livestock through improved 3D image processing.

In the case of the livestock industry, regular weight management is required for individual feeding of domesticated animals.

In particular, in the case of pig farms, the standard for shipping is very important, and depending on the shipment of pigs that meet the standard, there is a very big difference in the income of the farm. Pigs are graded according to the quantitative criteria according to the weight and fat thickness and the qualitative criteria according to the fat distribution and meat color of pork. Generally, pigs weighing from 115 kg to 120 kg are called standard pigs.

It is very important to measure the weight accurately and select the pigs for shipment, because if the weight of the standard pig is satisfied, a higher grade can be obtained.

To this end, the need for periodic weight measurement or monitoring is required in the field.

Currently, the weight of pigs is measured by the chest measurement method and the pig breeding method.

The chest position measurement method is used to convert the weight by applying the value obtained by measuring the chest position of the money to the weight calculation formula.

In addition, the pig weighing machine directly measures the weight of pigs in an enclosed space after installing an accessory device on a scale that measures the weight of pigs. There is a problem of having to stop for a certain period of time within the sentence, and in this process, it takes about 10 minutes or more based on one worker to measure the weight of one animal. There were breakdowns and difficulties in maintenance.

In addition, there is a shortage of manpower due to a decrease in the farm household population and an aging population, so countermeasures are needed.

Therefore, it is possible to measure the weight of a pig simply and accurately to continuously manage the weight of the pig and reduce the labor force of the farmer, and there is a need for a technology for accurately predicting the weight of the pig at the time of shipment. As a prior art, Korean Patent No. 10-2031200 (Method and System for Measuring Livestock Weight) has been published.

Recently, a system or method for estimating the volume, weight, etc. of livestock through images by photographing livestock has been developed, but there is a problem in that measurement accuracy is deteriorated because noise due to the external environment such as background and floor is included in the image.

In order to solve the above problems, the present invention processes the 3D image obtained through the photographing unit, and then measures the weight of the livestock by analyzing the factors necessary for calculating the weight, thereby improving the measurement accuracy of the livestock through 3D image processing. An object of the present invention is to provide a measuring device and method.

In order to solve the above problems, the apparatus for measuring the weight of livestock through 3D image processing according to an embodiment of the present invention includes a photographing unit for acquiring a 3D image of livestock and a measuring module for measuring the weight by processing the 3D image An apparatus for measuring livestock weight through 3D image processing, the measurement module comprising: a preprocessing unit for removing a background and an outlier by extracting a center point and a center distance from the 3D image; a standardization unit that extracts and removes the floor plane from the preprocessed 3D image, classifies it into groups according to the density of points, extracts a group corresponding to livestock, and corrects the slope using a plane equation; It is possible to provide a livestock weight measurement device comprising a factor analysis unit for extracting factors for one or more of length, height, circumference, area, and volume from a standardized 3D image, and a weight calculation unit for calculating weight using the extracted factors have.

In addition, the pre-processing unit, a data conversion unit for converting the reverse image and scale of the 3D image; In the 3D image, a central point is extracted assuming that the livestock is at the center, and a central extraction unit that extracts a central distance through the extracted central point, and a point that removes a point considered as a background or an outlier using the central point and the central distance It may include a removal unit.

In addition, the normalization unit extracts and removes the floor plane through a plane equation using points located below the center point in the pre-processed 3D image, and is located within a certain distance from the floor plane, but the number of adjacent points is insufficient. a floor removal unit to remove; An object detection unit for classifying points in the 3D image into groups according to density, extracting a group corresponding to livestock, and a tilt correction unit for correcting the slope of the 3D image from which the group corresponding to the livestock is extracted using a plane equation can do.

In addition, the floor removal unit, when the size of the angle between the extracted floor plane and the photographing direction of the photographing unit is greater than a certain angle, it is determined that it is not the floor, and the floor plane is extracted again.

In addition, the normalization unit extracts the wall plane when the number of points in the extracted group is greater than or equal to a certain number, and removes the extracted wall plane when the extracted wall plane is at least a certain distance from the center point, and is located within a certain distance from the wall plane, but the number of adjacent points It may further include a wall removal unit for removing the insufficient point.

In addition, the inclination correction unit obtains the tilted x-angle and z-angle through the extracted x-intercept and z-intercept of the floor plane, and corrects the inclination for the x-axis and y-axis of the 3D image according to the x-angle and z-angle characterized in that

In addition, the tilt correction unit extracts an arbitrary plane containing the most points using a plane equation, obtains coordinates of three points based on the origin, and obtains a tilted y angle using the coordinates of the three points, It is characterized in that the inclination of the 3D image with respect to the y-axis is corrected according to the y angle.

In addition, the factor analysis unit, as elements, back length, height (height), chest height, median height, posterior height, diagonal length, chest circumference, median circumference, posterior circumference, diagonal circumference, torso area, belly area, median belly It is characterized in that at least one of an area, a body volume, a belly volume, and a median belly volume is extracted.

In addition, the factor analysis unit may include: a presetting unit that extracts points corresponding to the lower abdomen from the standardized 3D image, and obtains forelimb points and hindlimb points by applying a quaternary function with the extracted points; a key point measurement unit for generating point groups of the chest, middle, and back points based on the forelimb points and hind limb points; Length measuring unit for measuring height and back length; a perimeter measuring unit that converts the chest, middle, and posterior point groups into functions, respectively, to obtain perimeters for the chest, middle and posterior; It may include a torso measurement unit that extracts the torso using the chest function and obtains the area and volume of the body, and a belly measurement unit that extracts the belly by using the chest function and the back function, and obtains the area and volume of the abdomen.

In addition, the method for measuring the weight of livestock through 3D image processing according to an embodiment of the present invention includes a photographing step in which a photographing unit acquires a 3D image from the livestock; a pre-processing step in which a measurement module extracts a center point and a center distance from the 3D image to remove a background and an outlier; a standardization step of extracting and removing the floor plane from the preprocessed 3D image, classifying it into groups according to the density of points, extracting a group corresponding to livestock, and correcting the slope using a plane equation; To provide a livestock weight measurement method comprising a factor analysis step of extracting factors for one or more of length, height, circumference, area, and volume from a standardized 3D image, and a weight calculation step of calculating weight using the extracted factors can

Here, the pre-processing step may include: a data conversion step in which the measurement module converts the reverse image and scale of the 3D image; A center point extraction step of extracting a center point assuming that the livestock is at the center from the 3D image, extracting a center distance through the extracted center point, and a point of removing a point considered as a background or an outlier using the center point and the center distance It may include a removal step.

In addition, in the normalization step, the measurement module extracts and removes a floor plane through a plane equation using points located below the center point in the preprocessed 3D image, and is located within a certain distance from the floor plane, but of adjacent points A floor removal step of removing the insufficient number of points; An object detection step of classifying points in the 3D image into groups according to density, extracting a group corresponding to livestock, and a tilt correction step of correcting the slope of the 3D image from which the group corresponding to the livestock is extracted using a plane equation may include.

In addition, the factor analysis step may include: a preset step in which the measurement module extracts points corresponding to the lower abdomen from the standardized 3D image, and applies a quaternary function with the extracted points to obtain a forelimb point and a hindlimb point; a key point measurement step of generating a point group of the chest, middle, and back points based on the forelimb points and hind limb points; Length measuring step of measuring the height and back length; a perimeter measurement step of converting the chest, median, and posterior point groups into functions, respectively, to obtain perimeters for the chest, median and posterior; The method may include extracting the torso using the chest function and measuring the torso to obtain the area and volume of the body, and extracting the belly by using the chest function and the back function, and measuring the belly to obtain the area and volume of the abdomen.

The apparatus for measuring the weight of livestock through 3D image processing according to the embodiment of the present invention as described above processes the 3D image obtained through the photographing unit, and then analyzes the factors necessary for weight calculation to measure the weight of the livestock, so that the measurement accuracy is improved can be further improved.

Therefore, there is no need for additional equipment to measure the weight of livestock, and it is possible to reduce the cost of breeding through feed control through continuous weight management of livestock, and it is possible to accurately predict the time of shipment, thereby increasing the profits of the farmer.

In addition, there is no hassle of inducing livestock to stop for a certain period of time to measure the weight, so it is possible to solve problems caused by shortage of manpower in farms, aging of manpower, and enlargement of scale.

In addition, as it can be applied not only to pigs but also to various livestock such as chickens and cattle, its utility is expected to expand.

1 is a manufacturing example diagram showing an apparatus for measuring livestock weight through 3D image processing according to an embodiment of the present invention.
Figure 2 is a block diagram showing the configuration of a livestock weight measurement device through 3D image processing according to an embodiment of the present invention.
3 is a block diagram showing the configuration of the measurement module of FIG.
FIG. 4 is a block diagram showing the configuration of the preprocessor of FIG. 3 .
FIG. 5 is a block diagram illustrating the configuration of a normalization unit of FIG. 3;
6 is a conceptual diagram illustrating a process of determining whether the floor plane extracted from the floor removal unit of FIG. 5 is the floor.
FIG. 7 is an exemplary diagram illustrating a state in which points are classified into groups according to density in the object detection unit of FIG. 5;
FIG. 8 is a conceptual diagram illustrating a process of obtaining a y-angle in order to correct the inclination with respect to the y-axis in the inclination correction unit of FIG. 5;
9 is a block diagram showing the configuration of the factor analysis unit of FIG.
10 (a) and (b) are exemplary views showing a state in which the forelimb point and the hind leg point are extracted from the preset unit of FIG. 9 and the head position is generalized.
11 is an exemplary view showing a state in which the chest, middle, and posterior point groups are generated based on the forelimb point and the hind leg point in the key point measurement unit of FIG. 9 .
12 is an exemplary view showing the torso extracted from the torso measurement unit of FIG. 9;
13 is an exemplary view showing a pear extracted from the pear measuring unit of FIG.
14 is a flowchart schematically illustrating a method for measuring livestock weight through 3D image processing according to an embodiment of the present invention.
15 is a flowchart sequentially illustrating steps S2 of FIG. 14;
16 is a flowchart sequentially illustrating steps S3 of FIG. 14;
17 is a flowchart schematically illustrating step S4 of FIG. 14 .

Hereinafter, the description of the present invention with reference to the drawings is not limited to specific embodiments, and various modifications may be made and various embodiments may be provided. In addition, it should be understood that the contents described below include all conversions, equivalents, and substitutes included in the spirit and scope of the present invention.

In the following description, terms such as first and second are terms used to describe various components, meanings are not limited thereto, and are used only for the purpose of distinguishing one component from other components.

Like reference numbers used throughout this specification refer to like elements.

As used herein, the singular expression includes the plural expression unless the context clearly dictates otherwise. In addition, terms such as "comprises", "comprising" or "have" described below are intended to designate the existence of features, numbers, steps, operations, components, parts, or combinations thereof described in the specification. It should be construed as not precluding the possibility of addition or existence of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and should not be interpreted in an ideal or excessively formal meaning unless explicitly defined in the present application. does not

In addition, in the description with reference to the accompanying drawings, the same components are given the same reference numerals regardless of the reference numerals, and the overlapping description thereof will be omitted. In describing the present invention, if it is determined that a detailed description of a related known technology may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted.

In the present invention, the measurement module is preferably provided in a mobile communication terminal including a smart phone communicating through a wireless communication network as an application, but may also be provided in a computer environment communicating through a wired communication network. Hereinafter, the mobile communication terminal will be mainly described.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the accompanying drawings 1 to 17.

1 is a manufacturing example diagram showing an apparatus for measuring livestock weight through 3D image processing according to an embodiment of the present invention, and FIG. 2 is a block showing the configuration of an apparatus for measuring livestock weight through 3D image processing according to an embodiment of the present invention 3 is a block diagram showing the configuration of the measurement module of FIG. 2 , FIG. 4 is a block diagram showing the configuration of the preprocessor of FIG. 3 , and FIG. 5 is a block diagram showing the configuration of the normalization unit of FIG. 6 is a conceptual diagram illustrating a process of determining whether the floor plane extracted by the floor removal unit of FIG. 5 is the floor, and FIG. 7 is an example showing a state in which the points are classified into groups according to the density in the object detection unit of FIG. 5 FIG. 8 is a conceptual diagram illustrating a process of obtaining a y-angle to correct the inclination on the y-axis in the inclination correction unit of FIG. 5, and FIG. 9 is a block diagram illustrating the configuration of the factor analysis unit of FIG. 10 (a) and (b) are exemplary views showing a state in which the forelimb points and hind leg points are extracted from the preset unit of FIG. 9 and the head position is generalized, and FIG. 11 is a forelimb point and It is an exemplary view showing the generation of chest, middle, and rear point groups based on the hind leg points, FIG. 12 is an exemplary view showing the torso extracted from the torso measurement unit of FIG. 9, and FIG. 13 is the belly measurement unit of FIG. It is an exemplary diagram showing a pear extracted from .

1 and 2, the livestock weight measurement device (1, hereinafter referred to as 'livestock weight measurement device') through 3D image processing according to an embodiment of the present invention is implemented in the form of a mobile terminal, so that mobility and convenience are improved It allows the user to easily measure the weight of the livestock according to its ease.

In addition, the livestock weight measurement apparatus 1 may include a photographing unit 10 , a measurement module 20 , and a display 30 .

The photographing unit 10 is capable of acquiring a 3D image by photographing livestock, and may be implemented with various sensors such as a temperature-based lidar capable of acquiring a 3D image.

The photographing unit 10 may start photographing as the livestock weight measurement device 1 operates to detect livestock, and may provide an image of the livestock photographed on the display 30 . In addition, as the livestock is detected, the photographing button may be activated on the display 30 .

Accordingly, as the photographing button is pressed by the user, a photographing signal may be received from the display 30 and the capture process may be performed to obtain a 3D image.

That is, when the user presses the photographing button and a photographing signal is input, the photographing button may be deactivated on the display 30 and one or more 3D images may be acquired through capture.

It is preferable that the capture consists of continuous capture so that a plurality of 3D images can be extracted, and continuous capture can be performed according to the value of the number of consecutive captures. The continuous capture count value may be 10 to 12, but is not limited thereto.

The 3D image obtained here may include two-dimensional information and depth information, and may further include temperature information.

By extracting the 3D image through capture in this way, performance degradation can be minimized.

The measurement module 20 is a software type, and may be an application (or mobile app), where the application refers to a general application based on Android and iOS.

This measurement module 20 may measure the weight of livestock by processing the 3D image.

Referring to FIG. 3 , the measurement module 20 may include a database 21 , a preprocessor 22 , a standardization unit 23 , a factor analysis unit 24 , and a weight calculation unit 25 .

The database 21 may store the collected 3D image, the measured weight of livestock, and the like, and may store all information generated in measuring the livestock weight.

The pre-processing unit 22 may change the obtained 3D image into a form suitable for processing, extract the center point and the center distance, and remove the background and outliers.

Referring to FIG. 4 , the preprocessor 22 may include a data conversion unit 220 , a center extraction unit 221 , and a point removal unit 222 .

The data conversion unit 220 may convert an inverse image and a scale of the 3D image, and may convert it using a transform matrix. The reverse phase can be transformed by changing the direction of the axis to be transformed to a negative number through a transform matrix. At this time, the 3D image is converted and corrected by changing the number of points and a transform matrix according to the type of the photographing unit 10 .

In addition, the data conversion unit 220 may convert the 3D image into a 3D coordinate data form by reading the scale of the 3D image and extracting points.

The center extraction unit 221 may extract a center point from the 3D image and extract a center distance through the extracted center point.

Here, the center point refers to a point that becomes the center of the 3D image, and the center distance may be the distance between the livestock weight measurement device 1 , that is, the photographing unit 10 and the livestock.

The center extraction unit 221 may extract a center point on the assumption that the livestock is at the center of the 3D image, and since the center distance is the distance between the photographing unit 10 and the livestock, the origin corresponding to the center point of the photographing unit 10 with the center point Judging by the distance from (o), the z-coordinate value of the center point can be set as the center distance.

To describe the extraction of the center point in more detail, the center extraction unit 221 may extract the first recognized coordinate as the center point while gradually increasing the range from 0 to the maximum value of the x-coordinate and y-coordinate. This is because the coordinates (0,0,z) become the central point, that is, since it is impossible to shoot in line with the center of the livestock when shooting, since the point may not exist at (0,0,z), it is assumed in the same way as above. to extract the central point.

The point removing unit 222 may remove a point regarded as a background or an outlier by using the center point and the center distance.

First, the point removing unit 222 may remove points located outside the center distance from the center point as a background.

In addition, the point removal unit 222 may remove a point in which the number of adjacent points is less than the set number of adjacent points among points located within a predetermined distance from the center point as an outlier (outlier point). Here, the adjacent point means a point adjacent to the corresponding point within a predetermined interval.

The normalization unit 23 may extract and remove the floor plane from the preprocessed 3D image, classify it into groups according to the density of points, extract a group corresponding to livestock, and correct the inclination using a plane equation.

Referring to FIG. 5 , the standardization unit 23 may include a floor removal unit 230 , an object detection unit 231 , a wall removal unit 232 , and a tilt correction unit 233 .

The floor removal unit 230 may extract the floor plane through a plane equation using points located below the center point in the pre-processed 3D image. Here, the plane equation corresponding to the floor can be obtained using the RANSAC algorithm.

More specifically, the floor removal unit 230 may add an arbitrary value to the smallest y-coordinate value among the points, select a point based on the value, and extract the floor plane. That is, by extracting points having a y-coordinate value smaller than that value, the floor plane can be extracted from the extracted points.

At this time, since the extracted floor plane may not be the floor, the floor removal unit 230 determines the size of the angle θ between the extracted floor plane and the photographing direction of the photographing unit 10 as shown in FIG. 6 . If the angle (θ) formed is equal to or greater than a certain angle, it is determined that it is not a floor, and the floor plane can be extracted again.

Here, the predetermined angle is preferably 50°, but is not limited thereto.

Accordingly, the floor removal unit 230 may remove a point corresponding to the floor plane when a floor plane in which the angle θ formed with the photographing direction of the photographing unit 10 is less than a predetermined angle is obtained.

In addition, the floor removal unit 230 is located within a predetermined distance from the floor plane, the number of adjacent points is less than the set number of adjacent points is determined as an abnormal point can be removed.

The object detector 231 may classify points in a 3D image into groups according to densities by using the DBSCAN algorithm as shown in FIG. 7 . The DBSCAN algorithm can create multiple groups by tying adjacent points together by calculating the density with the surrounding distance and the minimum neighbor data points.

Thereafter, the object detection unit 231 may extract the group including the central point as a group corresponding to the livestock. That is, points on the same level as the center point are extracted as points corresponding to livestock.

At this time, when the livestock is attached to the wall, the viscosity corresponding to the wall may be extracted as a group, and the wall may be additionally removed through the wall removal unit 232 .

When the number of points in the group extracted by the object detection unit 231 is greater than or equal to a certain number, the wall removal unit 232 may extract the wall plane by using the RANSAC algorithm to obtain a planar equation corresponding to the wall. Here, when the number of points is 100 or more, the wall may be removed, but the present invention is not limited thereto.

In this case, the wall removal unit 232 may extract the wall plane using points located outside a predetermined reference range with respect to the central point.

Also, the wall removal unit 232 may obtain a distance between the extracted wall plane and the center point, and remove the wall plane when the obtained distance is greater than or equal to a predetermined distance.

In addition, the wall removal unit 232 may be located within a predetermined distance from the extracted wall plane, and may be removed by determining that the number of adjacent points is insufficient as an abnormal point.

The tilt correction unit 233 may correct the tilt of the 3D image from which the group corresponding to the livestock is extracted using a plane equation. The inclination of the floor and the side can be corrected using the floor plane and any plane, which will be described in more detail below.

First, the inclination correction unit 233 obtains the tilted x-angle and z-angle through the extracted x-intercept and z-intercept of the floor plane in order to correct the floor inclination, and the x-axis of the 3D image according to the x-angle and z-angle and the inclination of the y-axis can be corrected.

After changing the tilted x-angle and z-angle to radian degrees, you can create a transformation matrix "R (rotation)" and rotate the 3D image to correct the tilt for the floor.

In addition, the inclination correction unit 233 extracts an arbitrary plane including the most points by using a plane equation to correct the side inclination, obtains coordinates of three points based on the origin (o), and You can use the coordinates to find the tilted y-angle.

For example, as shown in Fig. 8, the origin (o) and the origin (o) in the extracted arbitrary plane are obtained, respectively, and _{the points (P 1} , P _{2 ) corresponding to the points (o') spaced apart in the x direction are obtained,} The _{point (o") extending from P 1} , P ₂ and meeting vertically can be obtained. Then _{, using P 1} , P ₂ , o" forming a right triangle, the y angle (α) can be obtained with the cosine (cos). can

After changing the y angle obtained through this to radian angles, a transformation matrix "R (rotation)" is created and the 3D image is rotated to correct the inclination to the side.

The factor analysis unit 24 may extract factors for one or more of length, height, circumference, area, and volume from the standardized 3D image.

More specifically, the factor analysis unit 24 includes elements such as back length, height (height), chest height, median height, back height, diagonal length, chest circumference, median circumference, back circumference, diagonal circumference, torso area, It is characterized in that at least one of the belly area, the median belly area, the body volume, the belly volume, and the median belly volume is extracted.

Referring to FIG. 9 , the factor analysis unit 24 includes a preset unit 240 , a key point measurement unit 241 , a length measurement unit 242 , a circumference measurement unit 243 , a body measurement unit 244 , and a belly measurement unit. (245).

The preset unit 240 may extract points corresponding to the lower abdomen from the standardized 3D image, and apply a quaternary function with the extracted points to obtain a forelimb point and a hindlimb point. This will be described in more detail below with reference to FIG. 10 .

Prior to that, since the residual outliers may remain, the presetting unit 240 may remove the points once more by applying the RANSAC algorithm to the 3D image.

In this case, the preset unit 240 may obtain a plane equation of an arbitrary plane through the RANSAC algorithm, and may obtain a rotation angle rotated by the x-axis, y-axis, and z-axis of an arbitrary plane by using the plane equation. If the obtained rotation angle does not exceed an arbitrary value, points included in the plane are regarded as fitting points (points corresponding to livestock), and points opposite to this are regarded as outliers and deleted. Such a process may be repeated according to a specified number of repetitions.

Then, the preset unit 240 may extract points corresponding to the lower abdomen from the standardized 3D image, and apply a quaternary function with the extracted points. Accordingly, two coordinates having the minimum y value can be extracted from the quaternary function. Accordingly, as shown in (a) of FIG. 10 , points P _R and P _L may be extracted using the x and y values of the two coordinates. In this case, the point on the right is called the right point (P _R ), and the point on the left is called the _{left point (P L ).}

Then, the preset unit 240 may generalize the position of the head so that the position of the head is unified to the left or the right in the 3D image.

As such, the preset unit 240 may use a right point (P _R ) and a left point (P _L ) in order to generalize the position of the head. _L ) and the right point (P _R ) can be determined as the head direction by calculating the distance difference along the x direction.

For example, the pre-set unit 240, obtain the distance difference along the x direction between the left point which is located on right and left both ends in the 3D image and each point (P _L), right point (P _R), right point (P _R ) and the point located at the right end is determined to be greater, the direction of the 3D image may be changed so that the position of the head is to the left, as shown in FIG.

At this time, the left point (P _L) and the right point (P _R) and finally it is possible to extract the hind point (P _H) and the front legs point (P _F).

The key point measurement unit 241 may generate a chest, middle, and rear point group based on the _{extracted forelimb point P F} and hind leg point P _{H .}

Referring to FIG. 11 , the key point measurement unit 241 uses the forelimb points (P _F ) and the hind leg points (P _H ) to the forelimb minimum and maximum points ( _PF, P _F '), the hind limb minimum and maximum points (P _H , P _H ') can be obtained. Here, the hindlimb minimum and maximum points (P _H , P _H ') may be generated as a post-point group.

In addition, the key point measurement unit 241 virtually obtains the back inclination line according to the back inclination from the _{forelimb maximum point (P F '), and the point where the forelimb minimum point (P F} ) and the back inclination line meet perpendicularly to the shoulder point (P) _S ) can be extracted. The front legs of the minimum point (P _F) and the shoulder point (P _S) can be generated as a chest circumference point group.

This is a configuration to correct the inclination of the chest, so that the circumference of the chest and the height of the chest can be measured more accurately.

Also located in the middle of the key point measuring section 241 is the front legs minimum point (P _F) and the rear legs minimum point (P _H) x-coordinate value difference to obtain the front legs minimum point (P _F) and the rear legs minimum point (P _H) of The fold minimum (P _M ) can be extracted. In addition, it is possible to extract a shoulder point (P _S) and the rear legs up point (P _H ') (the times up to point P _M) via the home, such as the use.

Accordingly, the double minimum point (P _M ) and the double maximum point (P _M ') can be created as a median point group.

The length measurement unit 242 may measure the height and the back length.

First, the length measurement unit 242 may measure a height (height), a chest height, a median height, and a back height. The height (height) can be obtained from the difference between the maximum and minimum values of y at the points.

In addition, the length measurement unit 242 measures the chest height, median height, and posterior height in the 3D image of the minimum value of the y value and the shoulder point (P _S ), the belly maximum point (P _M '), and the hind leg maximum point (P _H '). Each can be obtained by finding the y-value difference.

In addition, the length measuring unit 242 by specifying to the x value of the shoulder point (P _S) the rear legs up point (P _H ') from the value of x in that the measuring range, and the value of x, which is included in the point measurement range corresponding A coordinate corresponding to the largest y value among y values may be generated, and points included in the generated coordinates may be extracted as equal points.

Accordingly, the length measuring unit 242 may create a cubic function using the extracted points, and obtain the equal length using the polynomial of the created cubic function.

In addition, the length measuring unit 242 is the front legs minimum point (P _F) and the rear legs up point by using the (P _H '), there is also available a length (diagonal_length) of the diagonals, to obtain the length of the straight line, regardless of the surface area of cattle As a result, the length of a straight line between the minimum point of the forelimbs (P _F ) and the maximum point of the hind legs (P _{H ') can be obtained.}

The perimeter measurement unit 243 makes the chest point group, the midpoint group, and the back point group into functions, respectively, and obtains the circumferences for the bust, median, and back respectively through the polynomials of the chest function, median function, and post function. .

In the case of a linear function, the perimeter can be obtained by finding the distance between two points, and in the case of an nth-order function, the perimeter can be obtained by calculating the distance between two points and summing them. That is, in the case of an nth-order function, the perimeter can be obtained using a polynomial. Here, n may be 2 or more.

Also circumference measurement unit 243 there is also available circumference (diagonal_girth) of the diagonal with the front legs minimum point (P _F) and the rear legs up point (P _H '), taking into account the surface area of the cattle the length of the curved portion to obtain, by deriving a function in accordance with the front legs minimum point (P _F) and the rear legs up point (P _H '), can be obtained through the periphery of the diagonal of the derived polynomial function.

The torso measurement unit 244 may extract the torso as shown in FIG. 12 using the chest function. At this time, if the slope of the chest function is positive, only points located on the right side of the chest function reference are extracted, and if the slope is negative, only points located on the left side are extracted.

In addition, the body measuring unit 244 can obtain the body area and body volume using the extracted body, and can be converted into a triangle mesh form by connecting three points within a given radius through the ball_pivoting algorithm. This is to transform data made up of points into data made up of triangular faces.

In this case, the body measuring unit 244 may further simplify the surface formed through the smooth simple. This is to reduce the number of faces.

Accordingly, the body measuring unit 244 may obtain the body area by the sum of the triangular areas.

In addition, the torso measurement unit 244 may obtain the body volume using the extracted torso, but may be obtained through a convex hull algorithm, but is not limited thereto. Meanwhile, even in this case, the volume can be obtained after applying the ball_pivoting algorithm and smooth simple.

The belly measurement unit 245 may extract the belly as shown in FIG. 13 using the chest function and the back function.

Similarly, the belly measurement unit 245 may calculate the area and volume in the same way as the body measurement unit 244 after applying the ball_pivoting algorithm and the smooth simple before calculating the belly area and the belly volume.

In addition, the belly measurement unit 245 may additionally separate only the central portion from the extracted belly to extract the central belly. In this case, only a portion included in a certain range based on the median function may be extracted as a median fold.

In addition, the belly measurement unit 245 can be obtained in the same way as the area and volume of the central vessel extracted as described above.

The weight calculation unit 25 may calculate the weight by using the factors extracted as described above. In this case, the weight may be calculated using a formula extracted through regression analysis using the above factors. In addition, the calculated weight may be provided to the user through the display 30 .

The display 30 may output an image of livestock photographed from the photographing unit 10 , and when the photographing unit 10 detects the livestock, a photographing button may be activated. In addition, a captured 3D image, a measured weight, etc. may be provided to the user.

Hereinafter, a method for measuring the weight of a livestock made through the apparatus for measuring the weight of a livestock as described above will be described in detail.

14 is a flowchart schematically illustrating a method for measuring livestock weight through 3D image processing according to an embodiment of the present invention, FIG. 15 is a flowchart sequentially illustrating step S2 of FIG. 14, and FIG. 16 is step S3 of FIG. It is a flowchart sequentially shown, and FIG. 17 is a flowchart schematically illustrating step S4 of FIG. 14 .

14 , the livestock weight measurement method through 3D image processing according to an embodiment of the present invention includes a photographing step (S1), a pre-processing step (S2), a standardization step (S3), a factor analysis step (S4), and a weight calculation It may include step S5.

First, in the photographing step (S1), the photographing unit 10 of the livestock weight measurement device 1 may acquire a 3D image from the livestock. As the livestock weight measurement device 1 is executed, the photographing unit 10 may photograph the livestock and capture it according to a photographing signal to obtain one or more 3D images.

In the pre-processing step (S2), the measurement module of the livestock weight measurement device 1 extracts the center point and the center distance from the 3D image to remove the background and outliers.

15, the pre-processing step (S2) is a data conversion step (S20) in which the measurement module 20 converts the reverse image and scale of the 3D image, assuming that the livestock is in the center of the 3D image, extracting the central point, and extracting It may include a center extraction step (S21) of extracting a center distance through the center point and a point removal step (S22) of removing a point regarded as a background or an outlier using the center point and the center distance. A description thereof will be omitted since it has been described in detail in the above device.

The standardization step (S3) extracts and removes the floor plane from the 3D image preprocessed in step S2, classifies it into groups according to the density of points, extracts the group corresponding to the livestock, and uses the plane equation to correct the slope. have.

Referring to FIG. 16 , step S3 may include a floor removal step S30 , an object detection step S31 , a wall removal step S32 , and a tilt correction step S33 .

In the floor removal step (S30), the measurement module 20 extracts and removes the floor plane through a planar equation using points located below the center point in the pre-processed 3D image, and is located within a certain distance from the floor plane but adjacent points It is possible to eliminate the lack of the number of .

In the object detection step S31, points in the 3D image may be classified into groups according to density, and a group corresponding to livestock may be extracted.

The wall removal step (S32) is performed only when the number of points in the extracted group is greater than or equal to a certain number, and when the number of points in the extracted group is greater than or equal to a certain number, a wall plane is extracted, and the extracted wall plane is a certain distance from the center point If it is abnormal, it is removed, and it is possible to remove a point located within a certain distance from the wall plane but lacking the number of adjacent points.

The inclination correction step (S33) may correct the inclination of the 3D image from which the group corresponding to the livestock is extracted using a plane equation.

Since a more detailed description of this has been described in detail in the above device, it will be omitted.

The factor analysis step S4 may extract factors for one or more of length, height, circumference, area, and volume from the standardized 3D image.

Referring to Figure 17, step S4 is a preset step (S40), a key point measurement step (S41), a length measurement step (S42), a circumference measurement step (S43), a body measurement step (S44), and a belly measurement step (S45) may include.

In the preset step (S40), the measurement module 20 extracts points corresponding to the lower abdomen from the standardized 3D image, and applies a quaternary function with the extracted points to obtain a forelimb point and a hindlimb point.

In the key point measurement step (S41), the chest, middle, and rear point groups may be generated based on the forelimb point and the hind leg point.

In the length measurement step (S42), the height and the back length may be measured.

In the perimeter measurement step (S43), the perimeter for the chest, median, and posterior can be obtained by converting the chest, median, and posterior point groups into functions, respectively.

In the torso measurement step ( S44 ), the torso is extracted using the chest function, and the torso area and volume can be obtained.

In the belly measurement step ( S45 ), the belly may be extracted using the chest function and the back function, and the area and volume of the belly may be obtained.

Since a more detailed description of this has been described in detail in the above device, it will be omitted.

In the weight calculation step (S5), the weight may be calculated using the factors extracted in the step S4. A detailed description thereof will be omitted.

As described above, the livestock weight measurement apparatus through 3D image processing according to an embodiment of the present invention processes the 3D image acquired through the photographing unit, and then analyzes the factors necessary for weight calculation to measure the weight of the livestock, Measurement accuracy can be further improved.

Therefore, there is no need for additional equipment to measure the weight of livestock, and it is possible to reduce the cost of breeding through feed control through continuous weight management of livestock, and it is possible to accurately predict the time of shipment, thereby increasing the profits of the farmer.

In addition, there is no hassle of inducing livestock to stop for a certain period of time to measure the weight, so it is possible to solve problems caused by shortage of manpower in farms, aging of manpower, and enlargement of scale.

In addition, as it can be applied not only to pigs but also to various livestock such as chickens and cattle, its utility is expected to expand.

Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improved forms of the present invention are also provided by those skilled in the art using the basic concept of the present invention as defined in the following claims. is within the scope of the right.

1: Livestock Weighing Device
10: Cinematography
20: measurement module
21: Database
22: preprocessor
220: data conversion unit
221: center extraction unit
222: point removal unit
23: standardization unit
230: floor removal unit
231: object detection unit
232: wall removal unit
233: tilt correction unit
24: Factor Analysis Department
240: preset unit
241: key point measurement unit
242: length measurement unit
243: circumference measurement unit
244: body measurement unit
245: stomach measurement unit
25: weight calculation unit
30: display

Claims (13)

In an apparatus for measuring livestock weight through 3D image processing, comprising: a photographing unit for acquiring a 3D image of livestock; and a measuring module for measuring the weight by processing the 3D image,
The measurement module is
a preprocessor for removing a background and an outlier by extracting a center point and a center distance from the 3D image;
a standardization unit that extracts and removes the floor plane from the preprocessed 3D image, classifies it into groups according to the density of points, extracts a group corresponding to livestock, and corrects the slope using a plane equation;
A factor analysis unit that extracts factors for one or more of length, height, circumference, area, and volume from a standardized 3D image, and
It includes a weight calculation unit that calculates the weight using the extracted factors,
The standardization unit,
a floor removal unit that extracts and removes a floor plane through a plane equation using points located below the center point in the pre-processed 3D image, and removes points located within a certain distance from the floor plane but lacking the number of adjacent points;
an object detection unit for classifying points in the 3D image into groups according to density and extracting a group corresponding to livestock; and
Livestock weight measuring device including a tilt correction unit for correcting the tilt of the 3D image extracted from the group corresponding to the livestock using a plane equation.
According to claim 1,
The preprocessor is
a data conversion unit for converting an inverse image and a scale of the 3D image;
A center extraction unit for extracting a center point assuming that the livestock is at the center from the 3D image, and extracting a center distance through the extracted center point; and
Livestock weight measuring device including a point removing unit for removing a point considered as a background or outlier using the center point and the center distance.
delete According to claim 1,
The floor removal unit,
When the angle between the extracted floor plane and the photographing direction of the photographing unit is equal to or greater than a certain angle, it is determined that it is not the floor, and the floor plane is extracted again.
According to claim 1,
The standardization unit,
When the number of points in the extracted group is more than a certain number, the wall plane is extracted, and when the extracted wall plane is more than a certain distance from the center point, it is removed, and a point located within a certain distance from the wall plane but lacking the number of adjacent points is removed Livestock weight measurement device further comprising a wall removal unit.
According to claim 1,
The tilt correction unit,
Livestock, characterized in that the inclination of the x-axis and the y-axis of the 3D image is corrected according to the x-angle and the z-angle obtained through the extracted x-intercept and the z-intercept of the floor plane weighing device.
According to claim 1,
The tilt correction unit,
Extracting an arbitrary plane containing the most points using the plane equation, obtaining the coordinates of three points based on the origin, obtaining the inclined y angle using the coordinates of the three points, Livestock weighing device, characterized in that for correcting the inclination on the y-axis of the 3D image.
According to claim 1,
The factor analysis unit,
Factors such as back length, height (height), chest height, median height, posterior height, diagonal length, chest girth, median girth, posterior girth, diagonal girth, torso area, belly area, mid-belly area, torso volume, belly Livestock weighing device, characterized in that for extracting at least one of the volume and the median belly volume.
In an apparatus for measuring livestock weight through 3D image processing, comprising: a photographing unit for acquiring a 3D image of livestock; and a measuring module for measuring the weight by processing the 3D image,
The measurement module is
a preprocessor for removing a background and an outlier by extracting a center point and a center distance from the 3D image;
a standardization unit that extracts and removes the floor plane from the preprocessed 3D image, classifies it into groups according to the density of points, extracts a group corresponding to livestock, and corrects the slope using a plane equation;
A factor analysis unit that extracts factors for one or more of length, height, circumference, area, and volume from a standardized 3D image, and
It includes a weight calculation unit that calculates the weight using the extracted factors,
The factor analysis unit,
Factors such as back length, height (height), chest height, median height, posterior height, diagonal length, chest girth, median girth, posterior girth, diagonal girth, torso area, belly area, mid-belly area, torso volume, belly extracting at least one of volume and median volume,
a presetting unit that extracts points corresponding to the lower abdomen from the standardized 3D image, and obtains forelimb points and hindlimb points by applying a quaternary function with the extracted points;
a key point measurement unit for generating point groups of the chest, middle, and back points based on the forelimb points and hind limb points;
Length measuring unit for measuring height and back length;
a perimeter measuring unit that converts the chest, middle, and posterior point groups into functions, respectively, to obtain perimeters for the chest, middle and posterior;
A torso measurement unit that extracts the body using the chest function and calculates the area and volume of the body;
A livestock weight measuring device including a belly measurement unit for extracting a belly using a chest function and a back function, and calculating a belly area and volume.
In a livestock weight measurement method through 3D image processing,
A photographing step in which the photographing unit acquires a 3D image from the livestock;
a pre-processing step in which a measurement module extracts a center point and a center distance from the 3D image to remove a background and an outlier;
a standardization step of extracting and removing the floor plane from the preprocessed 3D image, classifying it into groups according to the density of points, extracting a group corresponding to livestock, and correcting the slope using a plane equation;
A factor analysis step of extracting factors for one or more of length, height, circumference, area, and volume from a standardized 3D image, and
including a weight calculation step of calculating the weight using the extracted factors,
The standardization step is
The measurement module extracts and removes a floor plane through a plane equation using points located below the center point in the preprocessed 3D image, and removes a point located within a certain distance from the floor plane but lacking the number of adjacent points a floor removal step;
An object detection step of classifying points in the 3D image into groups according to density, and extracting a group corresponding to livestock; and
and a tilt correction step of correcting the tilt of the 3D image from which the group corresponding to the livestock is extracted using a plane equation.
11. The method of claim 10,
The pre-processing step is
a data conversion step in which the measurement module converts the reverse image and scale of the 3D image;
A center extraction step of extracting a center point assuming that the livestock is in the center from the 3D image, and extracting a center distance through the extracted center point; and
and a point removing step of removing a point regarded as a background or an outlier using the center point and the center distance.
delete In a livestock weight measurement method through 3D image processing,
A photographing step in which the photographing unit acquires a 3D image from the livestock;
a pre-processing step in which a measurement module extracts a center point and a center distance from the 3D image to remove a background and an outlier;
a standardization step of extracting and removing the floor plane from the preprocessed 3D image, classifying it into groups according to the density of points, extracting a group corresponding to livestock, and correcting the slope using a plane equation;
A factor analysis step of extracting factors for one or more of length, height, circumference, area, and volume from a standardized 3D image, and
including a weight calculation step of calculating the weight using the extracted factors,
The factor analysis step is
a preset step in which the measurement module extracts points corresponding to the lower abdomen from the standardized 3D image, and applies a quaternary function with the extracted points to obtain a forelimb point and a hindlimb point;
a key point measurement step of generating a point group of the chest, middle, and back points based on the forelimb points and hind limb points;
Length measuring step of measuring the height and back length;
a perimeter measurement step of converting the chest, median, and posterior point groups into functions, respectively, to obtain perimeters for the chest, median and posterior;
The torso measurement step of extracting the torso using the chest function and obtaining the body area and volume; and
A method for measuring the weight of livestock, comprising extracting a belly using a chest function and a back function, and measuring a belly to obtain an area and volume of the belly.

Images