TWI833352B - Terminal device, system and method for image measurement of biological characteristics of livestock - Google Patents

Abstract

A method for measuring biological characteristics of livestock through images, including: an image acquisition step of obtaining an image from a photography device, wherein the photography device is used to photograph a pen housing livestock from above; and a recognition step of artificial intelligence image recognition on the image. to obtain a single image frame corresponding to the back of a domestic animal; the calculation steps of calculating the body length and width corresponding to the animal's back of the image frame based on the height and viewing angle of the installation of the photography device; and based on the calculation steps of the image frame The body length and width of the back of the domestic animal, find the comparison steps for the weight corresponding to the image frame in the biometric comparison table.

Description

Terminal device, system and method for image measurement of biological characteristics of livestock

This application relates to image processing, and specifically to the use of images to measure biometric characteristics of livestock.

Humans have been raising livestock such as pigs, sheep, horses, and cattle for a long time. Scientific and centralized breeding is the trend in the development of animal husbandry. In order to understand the growth of livestock, it usually takes a lot of manpower to manually measure the biological characteristics of livestock, such as weight, body length, body width, body height, etc. Then, personnel will judge whether it meets the standards for sale or slaughter.

Since the number of livestock raised in large-scale livestock farms may be tens of thousands, the above-mentioned measurement process is not only laborious, but also consumes a lot of time. Therefore, there is an urgent need for a system and method that can automatically measure the biological characteristics of livestock to help livestock farms solve the above problems.

According to an embodiment of the present application, a terminal device for image measurement of biological characteristics of livestock is provided, a photography device for photographing a barn housing livestock from above; and a host connected to the photography device for executing non-volatile Multiple instructions stored in the sexual memory to achieve the following: the image acquisition step of causing the photography device to obtain an image; the recognition step of artificial intelligence image recognition on the image to obtain a single image frame corresponding to the back of a domestic animal; Calculate the body length and width corresponding to the back of the livestock in the image frame based on the installation height and viewing angle of the photography device; and calculate the body length and width of the livestock back in the image frame in the biometric comparison table Find the comparison steps for the weight corresponding to the image frame.

Preferably, in order to obtain the average biological characteristics of multiple livestock in the pen, the host is further used to repeatedly implement the aforementioned image acquisition steps, identification steps, calculation steps and comparison steps to obtain Z image frames and their corresponding Biological characteristics, calculate one of the following average biological characteristics of the pen or any combination thereof: the average body length of the livestock's back; the average body width of the livestock's back; and the average weight of the livestock, where Z is a positive number greater than or equal to 2 integer.

Preferably, in order to more accurately calculate biometric characteristics through images, the body length and body width corresponding to the back of the livestock in the image frame are related to one of the following or any combination thereof: the distance between the photography device and the pen The height of the ground; and the viewing angle of the photographic device.

Preferably, in order to avoid large errors in measurement, the host computer further obtains a spine line segment when performing artificial intelligence image recognition on the image. When the angle between the spine line segment and the first axis of the image is greater than one When the threshold value is reached, the aforementioned image acquisition step is re-implemented, wherein the image is composed of a plurality of pixel arrays parallel to the first axis.

Preferably, in order to avoid measuring parts with large aberrations, the host is further used to perform artificial intelligence image recognition on a part of the image after obtaining the image from the photography device.

Preferably, in order to obtain better recognition results, the artificial intelligence image recognition system uses a neural network model for image recognition. The neural network model is trained using back images of domestic animals of the same species raised in the pen. Finish.

Preferably, in order to report the measured biological characteristics of the livestock, the terminal device further includes a network connection device connected to the network system, wherein the host is further used to cause the network connection device to transmit the following to the network system through the network system. One or any combination thereof is reported to the management device: the body length corresponding to the back of the livestock in the image frame; the body width corresponding to the back of the livestock in the image frame; and the weight corresponding to the image frame.

Preferably, in order to report the measured average livestock biometric characteristics of the stall, the terminal device further includes a network connection device connected to the network system, wherein the host is further used to enable the network connection device to pass through the network The system reports one of the following or any combination thereof to the management device: the average body length of the livestock's back; the average body width of the livestock's back; and the average weight of the livestock.

Preferably, in order to accurately use body length and body width to correspond to the weight of livestock of the same species, the livestock in the pen and the corresponding data in the biometric comparison table belong to the same species.

Preferably, in order to allow the management device to remotely configure the terminal device, the terminal device further includes a network connection device connected to the network system, wherein the host is further used to enable the network connection device to pass through the network The system self-management device receives one of the following or any combination thereof: the biometric comparison table; and the neural network model used in artificial intelligence image recognition.

According to an embodiment of the present application, a method for measuring biological characteristics of livestock through images is provided, which includes: an image acquisition step of obtaining an image from a photography device, wherein the photography device is used to capture a barn housing livestock from above; The image is subjected to the recognition steps of artificial intelligence image recognition to obtain a single image frame corresponding to the back of a domestic animal; based on the height and viewing angle of the installation of the photography device, the corresponding body length and width of the livestock back of the image frame are calculated. The calculation step; and the comparison step of finding the weight corresponding to the image frame in the biometric comparison table based on the body length and body width of the livestock back of the image frame.

Preferably, in order to obtain the average biological characteristics of multiple livestock in the pen, the method for image measurement of biological characteristics of livestock further includes repeatedly implementing the aforementioned image acquisition steps, identification steps, calculation steps and comparison steps to obtain For Z image frames and their corresponding biological characteristics, calculate one of the following average biological characteristics of the pen or any combination thereof: the average body length of the livestock's back; the average body width of the livestock's back; and the average weight of the livestock, where Z is a positive integer greater than or equal to 2.

Preferably, in order to more accurately calculate biological characteristics through images, a method for measuring biological characteristics of livestock through images is used, wherein the body length and the body width corresponding to the back of the livestock in the image frame are related to one of the following or Any combination of: the height of the photography device from the ground of the pen; and the viewing angle of the photography device.

Preferably, in order to avoid large errors in measurement, when performing artificial intelligence image recognition on the image, a spine line segment is obtained. When the angle between the spine line segment and the first axis of the image is greater than a threshold value When, the aforementioned image acquisition step is re-implemented, wherein the image is composed of a plurality of pixel arrays parallel to the first axis.

Preferably, in order to avoid measuring parts with large aberrations, the method for measuring biological characteristics of livestock using images further includes performing artificial intelligence image recognition on a part of the image after obtaining the image from the photography device. .

Preferably, in order to obtain better recognition results, the artificial intelligence image recognition system uses a neural network model for image recognition. The neural network model is trained using back images of domestic animals of the same species raised in the pen. Finish.

Preferably, in order to report the measured biometric characteristics of livestock, the method for image measuring biometric characteristics of livestock further includes causing the network connection device to report one of the following or any combination thereof to the management device through the network system: The body length corresponding to the back of the livestock in the image frame; the body width corresponding to the back of the livestock in the image frame; and the weight corresponding to the image frame.

Preferably, in order to measure the average livestock biometric characteristics of the stall, the method for image measuring livestock biometric characteristics further includes causing the network connection device to report one of the following or any combination thereof to the network system through the network system. Management device: the average body length of the livestock's back; the average body width of the livestock's back; and the average body weight of the livestock.

Preferably, in order to accurately use body length and body width to correspond to the weight of livestock of the same species, the livestock in the pen and the corresponding data in the biometric comparison table belong to the same species.

Preferably, in order to allow the management device to remotely configure the terminal device, the method for image measurement of livestock biometrics further includes causing the network connection device to receive one or more of the following from the management device through the network system. Any combination: the biometric comparison table; and the neural network model used in artificial intelligence image recognition.

A system for image measurement of biological characteristics of livestock includes a terminal device and a management device as described above.

In summary, the present invention provides a system and method that can automatically measure the biological characteristics of livestock to help livestock farms save labor costs and time in measuring the biological characteristics of livestock, avoid human operation errors, and also This avoids stress to livestock during measurement and reduces the possibility of livestock being injured during measurement. When large-scale livestock farms raise livestock of different ages and varieties in multiple stalls at the same time, the automated measurement and management system can centralize the data obtained from multiple terminal devices into the management device to facilitate management and statistics.

In order to make the purpose, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described in detail below. Obviously, the described embodiments are only part of the embodiments of the present application and are not comprehensive embodiments. Based on the embodiments in this application, all other implementations obtained by those of ordinary skill in the art without any creative work fall within the scope of protection of this application.

The terms "first", "second", "third", etc. (if present) in the description, patent scope and drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. . It is to be understood that the items described are interchangeable under appropriate circumstances. In the description of this application, "plural" means two or more than two, unless otherwise expressly and specifically limited. Furthermore, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. Some of the block diagrams shown in the Figures are functional entities and do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software form, or implemented in one or more hardware circuits or integrated circuits, or in different networks and/or processor devices and/or microcontroller devices. Implement these functional entities.

In the description of this application, it should be noted that, unless otherwise clearly stated and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense. For example, it can be a fixed connection or a detachable connection. Connection, or integral connection; it can be mechanical connection, electrical connection or mutual communication; it can be direct connection, or indirect connection through an intermediary, it can be internal connection of two components or mutual communication between two components functional relationship. For those of ordinary skill in the art, the specific meanings of the foregoing terms in this application can be understood according to specific circumstances.

In order to make the purpose, features and advantages of the present application more obvious and easy to understand, the present application will be described in further detail below in conjunction with the drawings and specific implementation modes.

Please refer to FIG. 1 , which is a schematic block diagram of a system 100 for imaging biological characteristics of livestock according to an embodiment of the present application. The system 100 may include a network system 120, a management device 110 and at least one terminal device 130. The network system 120 may be a wired or wireless network, such as a local area network that complies with the IEEE 802.3 and IEEE 802.11 series standards, or it may be a wireless telecommunications network system that complies with the third generation (3G) and later wireless communication network standards. . Those of ordinary skill in the art can understand that any network system that can transmit information between the management device 110 and the terminal device 130 can be used as the network system 120 in this embodiment.

The terminal device 130 that uses images to measure biological characteristics of livestock can be placed in a breeding place, for example, installed above a pen. Each pen can be equipped with one or more terminal devices 130 . Since the animal husbandry industry usually raises livestock of a single breed born at the same time in the same pen, each terminal device 130 can correspond to a certain breed of livestock. In one example, livestock housed in the same pen are of the same sex.

The livestock mentioned in this application may include pigs, sheep, horses, cows and other mammals, but this application does not limit the types of livestock to the above-mentioned animals. For example, this application may be applicable to capybaras. In one embodiment, different terminal devices 130 may respectively correspond to different types of livestock.

The management device 110 is connected to the above-mentioned terminal device 130 through the network system 120. In one embodiment, the management device 110 can be more than one computer, which may have more computing resources (computing power and storage space), and is installed in the office of the livestock farm. In another embodiment, the management device 110 may be a mobile device, such as a laptop, a tablet, a smart phone, a personal digital assistant, etc.

Livestock farm personnel can view images, videos, real-time images and/or measured livestock biometric information captured by the terminal device 130 through the management device 110 . The management device 110 can store and replay one or any combination of the above video videos, real-time images and measured livestock biometric information, and perform statistics and comparison on the measured livestock biometric information to make analysis. Report.

As mentioned above, when changing the livestock housed in a certain pen, the manager can set the type of livestock housed in the pen through the management device 110, so that the terminal device 130 installed in the pen can be used correctly. Imaging measures biometric characteristics of livestock.

In one embodiment, the system 100 may also include one or more mobility management devices 140 connected to the network system 120 . As the name suggests, the mobile management device 140 can be a mobile device, such as a laptop computer, a tablet computer, a smart phone, a personal digital assistant, etc. The functions of the mobility management device 140 may be the same as the aforementioned management device 110 , or may be reduced to a part of the management device 110 . The mobile management device 140 can directly view the images, videos, real-time images and/or measured livestock biometric information captured by the terminal device 130, and can also replay the images, videos, real-time images and measurements stored by the management device 110. One or any combination of the measured livestock biometric information, as well as statistics, comparison and analysis reports produced by the management device 110.

Please refer to FIG. 2 , which is a block diagram of a management device 110 according to an embodiment of the present application. The management device 110 may include a host 210, an optional output device 220, an optional input device 230, a network connection device 240 and a storage device 250. The host 210 may include one or more central processing units (CPUs), a memory hierarchy required for computing, and interfaces for connecting to peripheral devices. The central processing unit may include one or more computing cores for performing logical and arithmetic operations. Generally speaking, the host 210 can execute an operating system (OS, Operating System) for controlling the management device 110 . For example, the central processor can execute x86, x64, ARM, and RISC-V series instruction sets. The operating system can include Windows system, Android system, iOS system, MacOS system, Linux series and other operating systems. The interface connected to the peripheral device may include one of the following SCSI, PCI, PCI-Express, AHB, AXI, USB and other interfaces or any combination thereof. This application does not limit the form of the host 210, operating system and interface.

Optional output devices 220 may include screens, touch screens, speakers, printers, etc. Optional input devices 230 may include keyboards, mice, touch screens, microphones, etc. The storage device 250 may include a solid-state electronic memory module composed of non-volatile memory, a storage medium such as a hard disk, an optical disk, and a magnetic tape, and a reading device, and is used to store the above-mentioned operating system and the software operating in the operating system environment. One or any combination of applications, images, videos, real-time images and measured livestock biometric information, as well as statistics, comparison and analysis reports produced by the management device 110 . The network connection device 240 is used to connect to the above-mentioned network system 120, and may include a wired or wireless network interface.

The embodiment of the management device 110 shown in FIG. 2 may also be applied to the aforementioned mobility management device 140. Simply add a battery to power each of the above devices while on the move.

Please refer to FIG. 3 , which is a block diagram of a terminal device 130 according to an embodiment of the present application. The terminal device 130 may include a host 310, a photography device 320, a storage device 330 and a network connection device 340. The description of the host 310 may be applicable to the description of the aforementioned host 210 . The storage device 330 may include a solid-state electronic memory module composed of non-volatile memory, a storage medium such as a hard disk, an optical disk, and a magnetic tape, and a reading device, and is used to store the above-mentioned operating system and the software operating in the operating system environment. One or any combination of applications, images, videos, live images and measured livestock biometric information. In some embodiments, the storage device 330 may also include a removable memory card, memory chip, etc. The network connection device 340 is used to connect to the above-mentioned network system 120, and may include a wired or wireless network interface.

The above-mentioned photography device 320 may include a lens assembly, an image sensing module, and an interface module. In some examples, lighting modules can also be included. In some embodiments, the lens group may have a fixed field of view (FOV, Field of View). In other embodiments, the angle of view of the lens group can be adjusted, and can be adapted to various stall heights after adjustment.

The image sensing module can sense one of infrared, visible light and/or ultraviolet bands or any combination thereof. The shorter the wavelength, the higher the resolution. Image sensing modules of different bands can be used according to the environment of various pens. For example, when taking night photography, you can use infrared band sensing modules or filter lenses.

The interface module is used to connect to the above-mentioned host 310, and may include one of the following SCSI, PCI, PCI-Express, AHB, AXI, USB, I2C and other interfaces or any combination thereof. The lighting module may include one or more light sources for timely illuminating the area where the lens group intends to capture images. The image sensing module can control the lighting module to obtain appropriate illumination according to the sensed image.

Please refer to FIG. 4 , which is a schematic side view of a pen according to an embodiment of the present invention. There is at least one photography device 320 above the pen, and the lens group illuminates downward toward the ground of the pen. The livestock 440 is standing on the ground of the stall and can be photographed by the photographing device 320 . As shown in the figure, the head of the domestic animal 440 faces to the left, the tail faces to the right, the forelimbs are thinner, and the hindlimbs are thicker.

The photography device 320 is installed at a height 430 of the pen. When the angle of view of the lens group is fixed, the height 430 determines the area of the captured pen. As shown in the figure, the length of the captured range on the first axis is 410. Those of ordinary skill in the art can understand that the lens assembly has aberration, that is, objects at the edge of the image will be deformed. Lower cost lens sets have more pronounced aberrations. In order to save costs, in one embodiment, the edges of the image that are too deformed can be cut off, leaving only the middle part. For example, the image portion with the length 420 of the first axis is the image for measuring the biological characteristics of livestock.

Please refer to FIG. 5 , which is a top view of a pen according to an embodiment of the present invention. The above-mentioned photographing device 320 is located above the pen, and its shooting range covers a larger part. Due to aberration, the surrounding image edges can be cut off, leaving the middle part. However, those of ordinary skill in the art can understand that when the aberration of the lens group used does not affect the measurement, there is no need to cut off the edge part and the obtained image can be directly used for processing.

When looking down at the livestock 440 from above, artificial intelligence image recognition technology can be used to obtain the image frame 510 of the livestock's back from the image. The above-mentioned artificial intelligence image recognition technology has been widely used in daily life. The most obvious example is the facial recognition function, which has been applied to personal computers and mobile phones to identify users. In one photo, multiple faces can be listed and recognized. In another example, license plate recognition can be performed on vehicles entering and exiting a parking lot.

Those of ordinary skill in the art can understand that in some embodiments, the above-mentioned artificial intelligence image recognition technology of livestock backs can prepare multiple photos with marked livestock back image frames and input them into a neural network model for training. . When the recognition success of this type of neural network model reaches a certain level, this type of neural network model can be used to identify the backs of domestic animals. The above-mentioned photos can be taken at a height of 430 degrees from the ground of the pen. The viewing angles are the same as those shown in Figure 4 or Figure 5.

The back of the livestock mentioned in this application may refer to the back of the livestock from the forelimbs to the hind limbs. As shown in FIG. 5 , the forelimbs and hindlimbs of the livestock can be identified from the top-view image, and the image frame 510 circles the back of the livestock. The image frame 510 may be a rectangle, and two adjacent sides of the rectangle are respectively parallel to the first axis 550 and the second axis 560 of the image. The first axis 550 is perpendicular to the second axis 560 .

Those of ordinary skill in the art can understand that in some embodiments, the above-mentioned artificial intelligence image recognition technology can also be used to identify line segments of the back vertebrae of livestock. When measuring the biological characteristics of livestock, if the direction the livestock faces can be parallel to the first axis or the second axis of the image, the measurement accuracy can be improved. In the embodiment shown in Figure 5, the spinal cord segment 520 of the domestic animal 440 can be identified.

After finding the image frame 510 corresponding to the livestock's back, the first axis length 530, which is the body length of the livestock's back, can be found, and the second axis length 540, which is the body width of the livestock's back, can also be found. In some embodiments, since the body length of the back of the domestic animal will be longer than the body width, the larger of the first axis length 530 and the second axis length 540 will be regarded as the body length, and the smaller one will be regarded as the body width. .

In addition, since the body length should correspond to the direction of the spine line segment 520, in some embodiments, it can be checked whether the two correspond. Please refer to FIG. 6 , which is a schematic diagram of the angle between a spine line segment and two axes according to an embodiment of the present application. When the first axis length 530 is longer than the second axis length 540, the present application can calculate the angle 610 between the spine line segment 520 and the first axis 550. When the angle 610 exceeds a certain threshold, it means that the orientation of the livestock is not parallel to the first axis. That is to say, the first axis length 530 cannot accurately represent the body length of the livestock's back, and the second axis length 540 cannot accurately represent the body width of the livestock's back.

In one embodiment, when the first axis length 530 is shorter than the second axis length 540, the present application can calculate the angle between the spine line segment 520 and the second axis 560. When the angle exceeds a certain threshold, it means that the orientation of the livestock is not parallel to the second axis. That is to say, the second axis length 540 cannot accurately represent the body length of the livestock's back, and the first axis length 530 cannot accurately represent the body width of the livestock's back.

After accurately measuring the body length and width, the corresponding weight can be found based on the biological characteristics comparison table of the livestock of that breed and/or gender. For example, the biometric comparison table may include three fields, namely body length, body width and weight. Using body length and body width, you can look up the table to find the corresponding weight, and weight is usually an important indicator in the animal husbandry industry to judge the growth of livestock.

The biometric comparison table can be obtained by self-measurement and statistics in the livestock farm in advance. Since domestic animals of different breeds and/or different genders may have different biological growth curves, the storage device 330 may store a biological characteristic comparison table of the livestock corresponding to the terminal device 130 . When different terminal devices 130 photograph different livestock in different stalls, the biometric comparison tables stored in these terminal devices 130 are different.

Please refer to FIG. 7 , which is a schematic flowchart of a method 700 for image measuring livestock biometrics according to an embodiment of the present application. The image measurement method 700 for livestock biometrics can be implemented by the above-mentioned terminal device 130 . In one embodiment, the host 310 can execute a program composed of a plurality of instructions stored in a non-volatile memory to implement the image measurement method 700 for livestock biometrics. The image measurement method 700 for livestock biometrics may start from step 710 .

Step 710: Obtain the image. The host 310 can cause the photography device 320 to provide a frame of images. Next, the flow may proceed to step 715 or 720.

Optional step 715: Capture the central image. Since the central image has smaller aberrations, the edges of the original image can be removed and the central image can be captured. Flow may proceed to step 720.

Step 720: Use the previously trained artificial intelligence model to perform artificial intelligence image recognition on the obtained image or central image to try to obtain an image frame of the back of the livestock. In order to minimize the error when estimating body weight, the requirements of the livestock back image frame must be strictly limited in order to accurately obtain the length and width of a single livestock back image frame. In one embodiment, this step can also obtain a spine line segment corresponding to the above-mentioned livestock back image frame. Flow may proceed to step 730.

Step 730: Determine whether a single livestock back image frame has been obtained? As mentioned above, in order to minimize the error in estimating body weight, the requirements of the livestock back image frame must be strictly limited in order to accurately obtain the length and width of a single livestock back image frame. When the artificial intelligence image in step 720 identifies a single livestock back image frame, the process may proceed to step 740. Otherwise, when the artificial intelligence image in step 720 cannot identify a single livestock back image frame, the process can return to step 710.

Step 740: According to the image frame obtained in step 720, obtain the body length and width information of the livestock back corresponding to the image frame. Since the height and angle of view of the photography device 320 are known, the corresponding body length and width of the back of the livestock can be calculated based on the lengths of two adjacent sides of the image frame. Then, the process can go to step 750 or step 760.

Optional step 750: Determine whether the spine line segment identified in step 720 is aligned with one of the two axes of the image. As mentioned before, when the angle between the spine segment and a certain axis is less than a certain threshold, the spine segment can be considered to be aligned with one of the two axes. In addition, when the vertebral line segment corresponds to an axis parallel to the body length of the livestock's back, the vertebral line segment can be considered to be aligned with the axis. When the captured spinal line segment of the livestock is aligned with one of the two axes of the image, the process may proceed to step 760 or optional step 770 . Otherwise, the flow may return to step 710.

In one embodiment, the order of step 740 and step 750 may be reversed. That is, after step 730 is completed, step 750 can be executed. When the judgment result in step 750 is yes, the process proceeds to step 740. After step 740 is completed, step 760, step 765 or optional step 770 is executed.

Step 760: Use the body length and width information on the back of the livestock to find the weight information of the livestock in the corresponding biometric comparison table. Next, the process may proceed to optional step 765, step 770, or step 780.

Optional step 765: Determine whether sufficient data has been obtained? In one embodiment, when the data of Z livestock animals is obtained, the process can proceed to step 770 to calculate the average value of the biometric characteristics. Otherwise, when the data of Z livestocks has not been obtained, the process can be returned to step 710 to obtain new images. Where Z is a positive integer greater than or equal to 2.

Optional step 770: Calculate the average of the biometric characteristics. In this step, the average body length and length of the livestock in this pen can be calculated based on the number of image frames captured on the back of the livestock, the measured body length and width information, and the corresponding weight information. One or any combination of average body width and average weight. Flow then proceeds to optional step 780.

Optional step 780: report to the management device 110. The host 330 connects to the management device 110 through the network system 120 through the network connection device 340, and reports all or part of the aforementioned obtained information to the management device 110.

According to an embodiment of the present application, there is provided a terminal device for image measurement of biological characteristics of livestock, a photography device for photographing a barn housing livestock from above; and a host connected to the photography device for performing non-volatile Multiple instructions stored in the sexual memory to achieve the following: the image acquisition step of causing the photography device to obtain an image; the recognition step of artificial intelligence image recognition on the image to obtain a single image frame corresponding to the back of a domestic animal; Calculate the body length and width corresponding to the back of the livestock in the image frame based on the installation height and viewing angle of the photography device; and calculate the body length and width of the livestock back in the image frame in the biometric comparison table Find the comparison steps for the weight corresponding to the image frame.

Preferably, in order to obtain the average biological characteristics of multiple livestock in the pen, the host is further used to repeatedly implement the aforementioned image acquisition steps, identification steps, calculation steps and comparison steps to obtain Z image frames and their corresponding Biological characteristics, calculate one of the following average biological characteristics of the pen or any combination thereof: the average body length of the livestock's back; the average body width of the livestock's back; and the average weight of the livestock, where Z is a positive number greater than or equal to 2 integer.

Preferably, in order to more accurately calculate biometric characteristics through images, the body length and body width corresponding to the back of the livestock in the image frame are related to one of the following or any combination thereof: the distance between the photography device and the pen The height of the ground; and the viewing angle of the photographic device.

Preferably, in order to avoid large errors in measurement, the host computer further obtains a spine line segment when performing artificial intelligence image recognition on the image. When the angle between the spine line segment and the first axis of the image is greater than one When the threshold value is reached, the aforementioned image acquisition step is re-implemented, wherein the image is composed of a plurality of pixel arrays parallel to the first axis.

Preferably, in order to avoid measuring parts with large aberrations, the host is further used to perform artificial intelligence image recognition on a part of the image after obtaining the image from the photography device.

Preferably, in order to obtain better recognition results, the artificial intelligence image recognition system uses a neural network model for image recognition. The neural network model is trained using back images of domestic animals of the same species raised in the pen. Finish.

Preferably, in order to report the measured biological characteristics of the livestock, the terminal device further includes a network connection device connected to the network system, wherein the host is further used to cause the network connection device to transmit the following to the network system through the network system. One or any combination thereof is reported to the management device: the body length corresponding to the back of the livestock in the image frame; the body width corresponding to the back of the livestock in the image frame; and the weight corresponding to the image frame.

Preferably, in order to report the measured average livestock biometric characteristics of the stall, the terminal device further includes a network connection device connected to the network system, wherein the host is further used to enable the network connection device to pass through the network The system reports one of the following or any combination thereof to the management device: the average body length of the livestock's back; the average body width of the livestock's back; and the average weight of the livestock.

Preferably, in order to accurately use body length and body width to correspond to the weight of livestock of the same species, the livestock in the pen and the corresponding data in the biometric comparison table belong to the same species.

Preferably, in order to allow the management device to remotely configure the terminal device, the terminal device further includes a network connection device connected to the network system, wherein the host is further used to enable the network connection device to pass through the network The system self-management device receives one of the following or any combination thereof: the biometric comparison table; and the neural network model used in artificial intelligence image recognition.

According to an embodiment of the present application, a method for measuring biological characteristics of livestock through images is provided, which includes: an image acquisition step of obtaining an image from a photography device, wherein the photography device is used to capture a barn housing livestock from above; The image is subjected to the recognition steps of artificial intelligence image recognition to obtain a single image frame corresponding to the back of a domestic animal; based on the height and viewing angle of the installation of the photography device, the corresponding body length and width of the livestock back of the image frame are calculated. The calculation step; and the comparison step of finding the weight corresponding to the image frame in the biometric comparison table based on the body length and body width of the livestock back of the image frame.

Preferably, in order to obtain the average biological characteristics of multiple livestock in the pen, the method for image measurement of biological characteristics of livestock further includes repeatedly implementing the aforementioned image acquisition steps, identification steps, calculation steps and comparison steps to obtain For Z image frames and their corresponding biological characteristics, calculate one of the following average biological characteristics of the pen or any combination thereof: the average body length of the livestock's back; the average body width of the livestock's back; and the average weight of the livestock, where Z is a positive integer greater than or equal to 2.

Preferably, in order to more accurately calculate biological characteristics through images, a method for measuring biological characteristics of livestock through images is used, wherein the body length and the body width corresponding to the back of the livestock in the image frame are related to one of the following or Any combination of: the height of the photography device from the ground of the pen; and the viewing angle of the photography device.

Preferably, in order to avoid large errors in measurement, when performing artificial intelligence image recognition on the image, a spine line segment is obtained. When the angle between the spine line segment and the first axis of the image is greater than a threshold value When, the aforementioned image acquisition step is re-implemented, wherein the image is composed of a plurality of pixel arrays parallel to the first axis.

Preferably, in order to avoid measuring parts with large aberrations, the method for measuring biological characteristics of livestock using images further includes performing artificial intelligence image recognition on a part of the image after obtaining the image from the photography device. .

Preferably, in order to obtain better recognition results, the artificial intelligence image recognition system uses a neural network model for image recognition. The neural network model is trained using back images of domestic animals of the same species raised in the pen. Finish.

Preferably, in order to report the measured biometric characteristics of livestock, the method for image measuring biometric characteristics of livestock further includes causing the network connection device to report one of the following or any combination thereof to the management device through the network system: The body length corresponding to the back of the livestock in the image frame; the body width corresponding to the back of the livestock in the image frame; and the weight corresponding to the image frame.

Preferably, in order to measure the average livestock biometric characteristics of the stall, the method for image measuring livestock biometric characteristics further includes causing the network connection device to report one of the following or any combination thereof to the network system through the network system. Management device: the average body length of the livestock's back; the average body width of the livestock's back; and the average body weight of the livestock.

Preferably, in order to accurately use body length and body width to correspond to the weight of livestock of the same species, the livestock in the pen and the corresponding data in the biometric comparison table belong to the same species.

Preferably, in order to allow the management device to remotely configure the terminal device, the method for image measuring livestock biometrics further includes causing the network connection device to receive one or more of the following from the management device through the network system. Any combination: the biometric comparison table; and the neural network model used in artificial intelligence image recognition.

According to an embodiment of the present application, a system for imaging biological characteristics of livestock is provided, including a terminal device and a management device as described above.

In summary, the present invention provides a system and method that can automatically measure the biological characteristics of livestock to help livestock farms save labor costs and time in measuring the biological characteristics of livestock, avoid human operation errors, and also This avoids stress to livestock during measurement and reduces the possibility of livestock being injured during measurement. When large-scale livestock farms raise livestock of different ages and varieties in multiple stalls at the same time, the automated measurement and management system can centralize the data obtained from multiple terminal devices into the management device to facilitate management and statistics.

This article uses specific examples to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the technical solutions and core ideas of the present application; those of ordinary skill in the art should understand that: they still The technical solutions described in the foregoing embodiments may be modified, or some of the technical features may be equivalently replaced; and such modifications or substitutions will not cause the essence of the corresponding technical solutions to depart from the scope of the technical solutions of the embodiments of the present application. .

100: Image measurement system for livestock biometrics 110:Manage device 120:Network system 130:Terminal device 140:Mobile management device 210:Host 220:Output device 230:Input device 240:Network connection device 250:Storage device 310:Host 320:Photography installation 330:Storage device 340:Network connection device 410: length of first axis 420:The length of the first axis 430:Height 440:Livestock 510: Image frame on the back of livestock 520: Spine line segment 530: First axis length 540: Second axis length 550:First axis 560:Second axis 610:Angle 710~780: steps

FIG. 1 is a schematic block diagram of a system 100 for imaging biological characteristics of livestock according to an embodiment of the present application. Figure 2 is a block diagram of a management device 110 according to an embodiment of the present application. FIG. 3 is a block diagram of a terminal device 130 according to an embodiment of the present application. Figure 4 is a schematic side view of a pen according to an embodiment of the present invention. Figure 5 is a schematic top view of a pen according to an embodiment of the present invention. Figure 6 is a schematic diagram of the angle between a spine line segment and two axes according to an embodiment of the present application. FIG. 7 is a schematic flowchart of a method 700 for measuring livestock biometrics through images according to an embodiment of the present application.

710~780: steps

Claims (19)

A terminal device for image measurement of biological characteristics of livestock, including: a photography device for photographing a pen in which livestock is kept from above; and a host connected to the photography device for executing multiple processes stored in a non-volatile memory Instructions to achieve the following: the image acquisition step of causing the photography device to obtain an image; the recognition step of artificial intelligence image recognition on the image to obtain a single image frame and a spine line segment corresponding to the back of a domestic animal. When the strip When the angle between the spine line segment and the first axis of the image is greater than a threshold value, the aforementioned image acquisition step is re-implemented, wherein the image is composed of multiple pixel arrays parallel to the first axis; according to the camera installed on the Calculate the height and angle of view of the corresponding body length and width of the livestock back of the image frame; and based on the body length and body width of the livestock back of the image frame, find the weight corresponding to the image frame in the biometric comparison table comparison steps. The terminal device for image measurement of biological characteristics of livestock as described in claim 1, wherein the host is further used to repeatedly implement the aforementioned image acquisition steps, identification steps, calculation steps and comparison steps to obtain Z image frames and their For the corresponding biological characteristics, calculate one of the following average biological characteristics of the pen or any combination thereof: the average body length of the livestock's back; the average body width of the livestock's back; and the average weight of the livestock, where Z is greater than or equal to 2 positive integer. The terminal device for image measurement of biological characteristics of livestock as described in claim 1, wherein the body length and body width corresponding to the back of the livestock in the image frame are related to one of the following or any combination thereof: the photography device The height from the floor of the pen; and the viewing angle of the photographic device. As claimed in claim 1, the terminal device for image measurement of biological characteristics of livestock, wherein the host is further used to perform artificial intelligence image recognition on a part of the image after obtaining the image from the photography device. The terminal device for image measurement of biological characteristics of livestock as described in claim 1, wherein the artificial intelligence image recognition system uses a neural network-like model to perform image recognition, and the neural network model uses the same animals raised in the pen. Training is completed on back images of various breeds of domestic animals. The terminal device for image measurement of livestock biometrics as described in claim 1 further includes a network connection device connected to a network system, wherein the host is further used to cause the network connection device to transmit data to the network system through the network system. One of the following or any combination thereof is reported to the management device: the body length corresponding to the back of the livestock in the image frame; the body width corresponding to the back of the livestock in the image frame; and the weight corresponding to the image frame. The terminal device for image measurement of livestock biometrics as described in claim 2 further includes a network connection device connected to a network system, wherein the host is further used to cause the network connection device to transmit data to the network system through the network system. One of the following or any combination thereof is reported to the management device: the average body length of the livestock's back; The average body width of the livestock at the back; and the average weight of the livestock. The terminal device for image measurement of biological characteristics of livestock as described in claim 1, wherein the livestock in the pen and the corresponding data in the biological characteristics comparison table belong to the same species. The terminal device for image measurement of livestock biometrics as described in claim 1 further includes a network connection device connected to a network system, wherein the host is further used to enable the network connection device to automatically use the network system. The management device receives one of the following or any combination thereof: the biometric comparison table; and the neural network model used in artificial intelligence image recognition. A method for measuring biological characteristics of livestock through images, including: an image acquisition step of obtaining an image from a photography device, wherein the photography device is used to photograph a pen housing livestock from above; and a recognition step of artificial intelligence image recognition on the image. , to obtain a single image frame and a spine line segment corresponding to the back of a domestic animal. When the angle between the spine line segment and the first axis of the image is greater than a threshold value, re-implement the aforementioned image acquisition steps, where the image It is composed of a plurality of pixel arrays parallel to the first axis; the calculation steps of calculating the corresponding body length and body width of the livestock back of the image frame according to the height and viewing angle of the installation of the photography device; and according to the calculation steps of the image frame The body length and width of the back of the domestic animal, find the comparison steps for the weight corresponding to the image frame in the biometric comparison table. The method for measuring biological characteristics of livestock via images as described in claim 10 further includes repeatedly implementing the aforementioned image acquisition steps, identification steps, calculation steps and comparison steps to obtain Z of the The image frame and its corresponding biological characteristics are used to calculate one or any combination of the following average biological characteristics of the pen: the average body length of the livestock's back; the average body width of the livestock's back; and the average weight of the livestock, where Z is greater than Or a positive integer equal to 2. The method for image measuring biological characteristics of livestock as described in claim 10, wherein the body length and body width corresponding to the back of the livestock in the image frame are related to one of the following or any combination thereof: the distance of the photography device The height of the pen floor; and the viewing angle of the photographic device. The method for measuring biological characteristics of domestic animals from images as described in claim 10 further includes performing artificial intelligence image recognition on a part of the image after obtaining the image from the photography device. The method for image measuring biological characteristics of livestock as described in claim 10, wherein the artificial intelligence image recognition system uses a neural network-like model to perform image recognition, and the neural network model uses the same species raised in the pen. Dorsal images of domestic animals were trained to complete. The method for image measuring biological characteristics of livestock as described in claim 10 further includes causing the network connection device to report one of the following or any combination thereof to the management device through the network system: the corresponding image frame on the back of the livestock The body length of the image frame; the corresponding body width of the livestock back of the image frame; and the corresponding weight of the image frame. The method for image measuring biological characteristics of livestock as described in claim 11 further includes causing the network connection device to report one of the following or any combination thereof to the management device through the network system: the average body length of the livestock's back; The average body width of the livestock at the back; and the average weight of the livestock. The method for image measuring biological characteristics of livestock as described in claim 10, wherein the livestock in the pen and the corresponding data in the biological characteristics comparison table belong to the same species. The method for image measuring biometric characteristics of livestock as described in claim 10 further includes causing the network connection device to receive one of the following or any combination thereof from the self-management device through the network system: the biometric comparison table; and artificial intelligence Neural network model used in smart image recognition. A system for image measurement of biological characteristics of livestock, including a terminal device and a management device as described in one of claim 6, claim 7, and claim 9.