TWI836453B - Animal growth identification method and system - Google Patents

Abstract

The invention relates to an animal growth identification method and system. At least one photographing unit is used for shooting an animal in a breeding area, and a captured stereoscopic image and a captured thermal image are transmitted to a control unit. A stereoscopic contour and a thermal image contour of the breeding animal are constructed in the control unit, respectively. Then the two contours are correspondingly superimposed to obtain the corrected contour of the animal, and then the volume of the corrected contour is calculated, and the conditional factors such as the breed and growth stage of the animal are identified to output the corresponding volume-to-weight conversion ratio. The obtained volume is then substituted into the volume-to-weight conversion ratio to obtain the weight of the animal. In this way, the weight of several animals in the breeding area can be summed up to accurately estimate the harvesting time of the animals in the breeding area, so as to achieve benefits such as effective savings in feed costs.

Description

Animal growth identification method and system

The present invention relates to an animal growth identification method and system, and more particularly to a method and system that can accurately estimate the harvesting time of farmed animals to effectively save feed costs.

According to the World Health Organization, farmed animals such as chickens, pigs, and cattle are the main sources of meat and protein in people's diet. They are not only high in nutritional value, but are also the raw materials for many processed foods.

The existing farms mainly rely on manual sampling and weighing of a small number of farmed animals, or use fixed physical scales to let the animals jump on to weigh themselves. However, whether it is manual sampling weighing or the animals weighing themselves, the measurement frequency and data are too little, which can easily cause inaccurate estimates of the weight distribution of the group. The feed cost of farmed animals generally accounts for more than 70% of the total operating cost. Therefore, the weight data of farmed animals is extremely important for farming operations. It can be used to calculate the optimal number of feed feeding days. In addition, weight distribution is also the main calculation basis for revenue. If the growth and harvesting time of animals can be accurately estimated, it will help optimize feed costs and increase the profits of farmers.

Therefore, the inventor, in view of the above-mentioned deficiencies in the existing methods for obtaining the weight of farmed animals, has developed the present invention with the assistance of his many years of manufacturing and design experience and knowledge in related fields and through various ingenuity.

The present invention relates to an animal growth identification method and system thereof, and its main purpose is to provide a method and system thereof that can accurately estimate the harvesting time of farmed animals to effectively save feed costs.

In order to achieve the above implementation objectives, the inventors have developed the following animal growth identification method, the implementation steps of which include:

A. Shooting images: using at least one photography unit to shoot animals within a preset range of a breeding area, and transmitting the three-dimensional visual images and thermal images of the animals to a control unit;

B. Constructing image outline: The control unit then constructs a stereoscopic visual outline and a thermal image outline of one of the animals in the stereoscopic visual image frame and the thermal image frame;

C. Obtain the corrected outline: The control unit then superimposes the stereoscopic visual outline and the thermal image outline of the animal to obtain the corrected outline of the animal;

D. Calculate the volume of the animal: continue to use the control unit to measure the edge of the correction contour to obtain the depth and bottom area of the edge of the correction contour, and then calculate the volume of the animal;

E. Obtain the volume-to-weight conversion ratio: The control unit will identify various condition factors of the animal and output the volume-to-weight conversion ratio that meets the condition factors;

F. Calculate the weight of the animal: The control unit then substitutes the obtained volume of the animal into the volume-to-weight conversion ratio to calculate the weight of the animal;

G. Calculate the total weight of the animals: The control unit then calculates the weights of the animals within the preset range of the breeding area photographed by the camera unit one by one, then adds up the weights of the animals and displays the weight statistics on a display unit.

The animal growth identification method as described above, wherein the animal growth identification method further includes an abnormality warning step, which enables the control unit to compare the thermal image of the animal with the normal body temperature of the animal to determine whether the animal Whether the animal is sick or dead, if it is determined that the animal is sick or dead, an abnormal warning message will be displayed on the display module.

The animal growth identification method as described above, wherein the condition factors of the animal identified by the control unit include at least one of the animal's species, growth stage and growth season.

The inventor further developed the following animal growth identification system, which mainly includes a fixed unit, which is installed in a breeding area, and at least one photography unit is installed on the fixed unit, and the photography unit is inside It is equipped with at least two stereoscopic imaging lenses and at least one thermal imaging lens. It also includes a control unit for signal connection with the photography unit, and an animal growth recognition artificial intelligence program is built in the control unit. The animal growth recognition artificial intelligence program The smart program includes a three-dimensional image contour construction module, a thermal image contour construction module, a contour correction module, a volume calculation module, a volume-to-weight conversion ratio module, a weight calculation module and a weight statistics module. The group is operated by the control unit connected to each module of the animal growth recognition artificial intelligence program. Among them, the photography unit shoots the farmed animals within a preset range of the breeding area and combines the captured three-dimensional visual images. and the thermal image frame is transmitted to the control unit, the three-dimensional image contour construction module constructs a three-dimensional visual contour of one of the animals in the stereoscopic vision image frame, and the thermal image contour construction module constructs a three-dimensional visual contour of the animal in the thermal image frame The one animal constructs its thermal image outline, and the outline correction module superimposes the animal's stereoscopic vision outline and the thermal image outline correspondingly to obtain the corrected outline of the one animal, and the volume calculation module is Measure the edge of the correction outline to obtain the base area and depth of the edge of the correction outline, and then calculate the volume of the animal. The volume-to-weight conversion ratio module will identify various condition factors of the animal to output The volume and weight conversion ratio that meets the condition factors, and the weight calculation module will substitute the volume of the animal obtained from the volume calculation module into the volume and weight conversion ratio provided by the volume and weight conversion ratio module to calculate The weight of the animal is obtained, and the weight statistics module adds up statistics of the weight of a number of animals within a preset range captured by the photography unit, and connects the control unit to a display unit to calculate the weight of the animal. The weight statistics results are displayed on the display unit.

In the animal growth identification system as described above, the animal growth identification artificial intelligence program further includes an abnormal warning module, which identifies the body temperature of the animal in the thermal image and compares it with the normal body temperature of the animal to determine whether the animal is sick or dead. If the animal is determined to be sick or dead, the control unit is driven to display an abnormal warning message on the display module.

In the animal growth identification system as described above, the animal condition factor identified by the volume-to-weight conversion ratio module includes at least one of the animal species, growth stage, and growth season.

In this way, when the present invention is implemented, at least one photography unit takes pictures of animals in a breeding area, and transmits the captured stereoscopic vision images and thermal images to a control unit, so that the control unit can respectively Construct the stereoscopic visual outline and thermal image outline of the farmed animal, and then superimpose the two outlines to obtain the corrected outline of the animal. Then calculate the volume of the corrected outline, and then identify the species, growth stage and other conditions of the animal. The factor is to output the corresponding volume-to-weight conversion ratio, and then the obtained volume is substituted into the volume-to-weight conversion ratio to obtain the weight of the animal. This way, the weight of the many animals photographed in the breeding area can be added up. Statistics can be used to accurately estimate the harvesting time of animals in the breeding area, thereby effectively saving feed costs and other benefits.

In order to make the technical means of the present invention and the effects that can be achieved more completely and clearly disclosed, the following is a detailed description, and please refer to the disclosed drawings and figure numbers:

First, please refer to the first and third figures, which are the animal growth identification method and system of the present invention, which mainly include a fixed unit (1), which is a bracket mounted above the breeding area, and at least one camera unit (2) is assembled on the fixed unit (1), and the camera unit (2) is equipped with at least two stereoscopic image lenses (21) and at least one thermal image lens (22), and also includes a control unit (3), which can be a desktop computer, a laptop computer or a tablet computer, etc., connected to the camera unit (2), and built in the control unit (3). An animal growth recognition artificial intelligence program (4) is provided. The animal growth recognition artificial intelligence program (4) includes a three-dimensional image contour construction module (41), a thermal image contour construction module (42), a contour correction module (43), a volume calculation module (44), a volume weight conversion ratio module (45), a weight calculation module (46), a weight statistics module (47) and an abnormal warning module (48). The control unit (3) is connected to the operation of each module of the animal growth recognition artificial intelligence program (4), and the control unit (3) is connected to a display unit (5) such as a screen.

Accordingly, when the present invention is used and implemented, please refer to the second figure, and the implementation steps include:

A. Shooting images: The photography unit (2) is installed on the fixed unit (1) and located above a breeding area, and then the photography unit (2) is allowed to shoot a preset range of the breeding area. Chickens, pigs, cows and other farmed animals (6) are stored in the farm, and the three-dimensional visual images and thermal images of the captured animals (6) are transmitted to the control unit (3);

B. Constructing image contours: The three-dimensional image contour construction module (41) established by the animal growth recognition artificial intelligence program (4) built in the control unit (3) uses 3D point cloud technology to construct the stereoscopic visual image. One of the animals (6) constructs its stereoscopic visual outline (61), and the thermal image outline construction module (42) uses 3D point cloud technology to construct its thermal image of the one of the animals (6) in the thermal image. Image outline(62):

C. Obtaining a corrected contour: Please refer to the fourth figure, and then the contour correction module (43) superimposes the constructed three-dimensional visual contour (61) and thermal image contour (62) of the one animal (6) to obtain a corrected contour of the animal (6) after removing feathers or hair, etc.;

D. Calculate the volume of the animal: The volume calculation module (44) measures the edge of the corrected contour to obtain the average depth of the corrected contour edge and the depth and bottom area of each edge pixel of the corrected contour, and then substitutes them into A volume solution formula is used, wherein p is the average depth of the edge, di is the depth of the i-th pixel, and Ai is the bottom area of the i-th pixel. Accordingly, the volume of the animal (6) is calculated through the corrected outline of the animal (6);

E. Obtain the volume-to-weight conversion ratio: Then the volume-to-weight conversion ratio module (45) will identify at least one of the condition factors such as the species, growth stage, and growth season of the animal (6) to output an output that matches these conditions. The volume-to-weight conversion ratio of the conditional factors is obtained by calculating the actual weight of general animals that conforms to the species, growth stage, growing season and other conditional factors through the neural network;

F. Calculating the weight of the animal: The weight calculation module (46) then substitutes the volume of the animal (6) obtained from the volume calculation module (44) into the volume-to-weight conversion ratio provided by the volume-to-weight conversion ratio module (45) that is consistent with the species, growth stage and growth season of the animal (6), and then calculates the weight of the animal (6);

G. Calculate the total weight of the animals: After calculating the weights of the plurality of animals (6) within the preset range of the breeding area photographed by the photographic unit (2), the weight statistical module (47) adds up the weights of the plurality of animals (6) and displays the weight statistical results on the display unit (5) connected to the control unit (3).

Thus, the breeder can accurately know the group weight distribution status of a large number of animals (6) in the breeding area, and then accurately estimate the best harvesting time and feed feeding days of the animals (6) in the breeding area, so as to achieve practical benefits such as effectively saving feed costs and increasing the profit of the breeder. It is also worth mentioning that the animal growth identification artificial intelligence program (4) of the present invention is provided with an abnormal warning module (48), and further includes an abnormal warning step. When the camera unit (2) transmits the thermal image of the animal (6) photographed to the control unit (3), the abnormal warning module (48) identifies the body temperature of the animal (6) in the thermal image to compare it with the normal body temperature of the animal (6) to determine whether the animal (6) is sick or dead. If it is determined that the animal (6) is sick or dead, the control unit (3) is driven to display an abnormal warning message on the display unit (5), thereby facilitating the breeder to dispose of the sick or dead animal in time to avoid greater losses such as the spread of the disease.

The aforementioned embodiments or drawings do not limit the animal growth identification method and system implementation aspects of the present invention. Any appropriate changes or modifications made by a person having ordinary knowledge in the relevant technical field should be deemed to be within the patent scope of the present invention.

It can be seen from the above structure and implementation that the present invention has the following advantages:

1. The animal growth identification method and its system of the present invention introduces the technology of visual calculation of the weight of farmed animals, which can monitor a number of farmed animals within a range of the breeding area at one time, thereby accurately obtaining the information of the number of animals in the breeding area. Group weight distribution status, and then predict the best harvest time and feed feeding days of these farmed animals, in order to achieve effective savings in feed costs and other benefits.

2. The animal growth identification method and system of the present invention captures thermal images of animals to identify the animal's body temperature and further judge whether the animal is sick or dead, so as to display a warning message when an abnormality occurs, thereby facilitating timely disposal of the sick or dead animal to avoid greater losses such as the spread of the disease.

3. The animal growth identification method and system of the present invention can identify the species, growth stage and season of different animals to obtain the corresponding volume-to-weight conversion ratio. Therefore, even if various animals are mixed in a farm, the mixed animals can be conveniently calculated separately to accurately estimate the growth and harvesting time of various animals.

To sum up, the embodiments of the present invention can indeed achieve the expected effects, and the specific structure disclosed has not only not been seen in similar products, but has also not been disclosed before the application. It has fully complied with the provisions of the patent law and If you apply for an invention patent in accordance with the law, please review it and grant a patent, which will be very convenient.

1: Fixed unit

2: Photography unit

21: Stereoscopic lens

22: Thermal imaging lens

3: Control unit

4: Animal growth identification artificial intelligence program

41: Stereoscopic image outline construction module

42: Thermal image contour construction module

43: Contour correction module

44: Volume calculation module

45: Volume to weight conversion ratio module

46: Weight calculation module

47: Weight Statistics Module

48: Abnormal warning module

5: Display unit

6: Animals

61: Stereoscopic Outline

62: Thermal Image Contour

Figure 1: System architecture diagram of the present invention

Second figure: flow chart of the present invention

Figure 3: Usage status diagram of the present invention

Figure 4: The superimposed state diagram of the three-dimensional visual contour and the thermal image contour of the present invention.

2: Photographic unit 21: Stereoscopic image lens 22: Thermal image lens 3: Control unit 4: Animal growth recognition artificial intelligence program 41: Stereoscopic image contour construction module 42: Thermal image contour construction module 43: Contour correction module 44: Volume calculation module 45: Volume weight conversion ratio module 46: Weight calculation module 47: Weight statistics module 48: Abnormal warning module 5: Display unit

Claims (6)

An animal growth identification method, the implementation steps of which include: A. Shooting images: using at least one photography unit to shoot animals within a preset range of a breeding area, and transmitting the three-dimensional visual images and thermal images of the animals to a control unit; B. Construct image outline: Then let the control unit construct its stereoscopic visual outline and thermal image outline for one of the animals in the stereoscopic visual image frame and thermal image frame; C. Obtain the corrected outline: The control unit then superimposes the stereoscopic vision outline and the thermal image outline of the animal to obtain the corrected outline of the animal; D. Calculate the volume of the animal: continue to use the control unit to measure the edge of the correction contour to obtain the depth and bottom area of the edge of the correction contour, and then calculate the volume of the animal; E. Obtain the volume-to-weight conversion ratio: The control unit will identify various condition factors of the animal and output the volume-to-weight conversion ratio that meets the condition factors; F. Calculate the weight of the animal: The control unit then substitutes the obtained volume of the animal into the volume-to-weight conversion ratio to calculate the weight of the animal; G. Statistics of the total weight of animals: The control unit then calculates the weights of the many animals within the preset range of the breeding area photographed by the photography unit one by one, and then adds up the weights of the many animals. And the weight statistical result is displayed on a display unit. The animal growth identification method as described in claim 1, wherein the animal growth identification method further includes an abnormality warning step, which enables the control unit to compare the thermal image of the animal with the animal's normal body temperature to determine Determine whether the animal is sick or dead. If it is judged that the animal is sick or dead, an abnormal warning message is displayed on the display module. In the animal growth identification method as described in claim 1, the condition factor of the animal identified by the control unit includes at least one of the animal's species, growth stage and growth season. An animal growth identification system mainly includes a fixed unit, which is set up in a breeding area, and at least one camera unit is arranged on the fixed unit. The camera unit has at least two stereoscopic image lenses and at least one thermal image lens. It also includes a control unit to connect with the camera unit signal, and an animal growth identification artificial intelligence program is built in the control unit. The animal growth identification artificial intelligence program includes a stereoscopic image contour construction module, A thermal image contour construction module, a contour correction module, a volume calculation module, a volume weight conversion ratio module, a weight calculation module and a weight statistics module are connected to the operation of each module of the animal growth recognition artificial intelligence program by the control unit, wherein the photography unit is to shoot the farmed animals within a preset range of the breeding area, and transmit the photographed three-dimensional visual image picture and thermal image picture to the control unit, and the three-dimensional image contour construction module is to The thermal image contour construction module constructs a thermal image contour of the animal in the thermal image frame, and the contour correction module overlaps the stereoscopic visual contour and the thermal image contour of the animal to obtain the corrected contour of the animal. The volume calculation module measures the edge of the corrected contour to obtain the bottom area and depth of the edge of the corrected contour, and then calculates the volume of the animal. The volume weight conversion ratio module identifies the weight of the animal. Each condition factor is used to output a volume-to-weight conversion ratio that meets the condition factor. The weight calculation module substitutes the animal volume obtained from the volume calculation module into the volume-to-weight conversion ratio provided by the volume-to-weight conversion ratio module to calculate the weight of the animal. The weight statistics module adds up the weights of a plurality of animals within a preset range photographed by the photography unit. The control unit is connected to a display unit to display the weight statistics result on the display unit. An animal growth identification system as described in claim 4, wherein the animal growth identification artificial intelligence program further includes an abnormal warning module, which identifies the body temperature of the animal in the thermal image and compares it with the normal body temperature of the animal to determine whether the animal is sick or dead. If the animal is determined to be sick or dead, the control unit is driven to display an abnormal warning message on the display module. An animal growth identification system as described in claim 4, wherein the animal condition factor identified by the volume-to-weight conversion ratio module includes at least one of the animal's species, growth stage, and growth season.

Images