Multi-view animal multi-modal data acquisition device

Abstract

The invention provides a multi-view animal multi-modal data acquisition device which is applied to the technical field of data acquisition. The device comprises an animal channel, a plurality of image acquisition devices, an image acquisition device support and a control device. The two sides of the animal channel are defined by two net-shaped fences, and the animal channel is a three-section type bevel route and comprises an inlet section, a collecting section and an outlet section which are sequentially connected. The plurality of image acquisition devices are used for acquiring multi-modal data of animals at different visual angles; the image acquisition equipment bracket is used for respectively arranging the plurality of image acquisition equipment at the front position, the rear position, the left position, the right position and the upper position of the acquisition section; and the control equipment is connected with the plurality of image acquisition equipment and is used for acquiring and storing part of data streams from the multi-modal data streams acquired by the plurality of image acquisition equipment under the condition that the trigger signal is received.

Description

Technical Field

[0001] The present invention relates to the technical field of data acquisition, and in particular to a multi-viewing angle animal multi-modal data acquisition device. Background Art

[0002] In modern animal breeding and animal selection, the accurate collection and analysis of animal shape, appearance and morphology information plays a pivotal role. This process provides a key basis for evaluating animal growth and development, selecting excellent varieties and optimizing breeding strategies.

[0003] Traditional methods of collecting key animal information mainly rely on close observation or close-up measurement. However, this mode of operation has many disadvantages. On the one hand, the efficiency of manual operation is extremely low, and it is difficult to meet the needs of fast and batch data acquisition in large-scale farming scenarios; on the other hand, close contact with animals can easily cause stress reactions in animals. When animals are under stress, the secretion of growth hormone in the body will be disordered, and the function of the immune system will also decline. This not only affects the healthy growth of animals, but may also have a negative impact on the overall farming benefits. Summary of the invention

[0004] The present invention provides a multi-viewing angle animal multi-modal data acquisition device, which is used to solve the problem that the traditional method for collecting key animal information is not only inefficient but also easily causes animal stress.

[0005] The present invention provides a multi-view animal multi-modal data acquisition device, comprising: an animal passage, a plurality of image acquisition devices, an image acquisition device bracket and a control device; the two sides of the animal passage are surrounded by two mesh fences, and the animal passage is a three-section angled route, including an inlet section, an acquisition section and an outlet section connected in sequence; the plurality of image acquisition devices are used to acquire multi-modal data of animals from different viewpoints; the image acquisition device bracket is used to respectively place the plurality of image acquisition devices at the front position, rear position, left position, right position and top position of the acquisition section; the control device is connected to the plurality of image acquisition devices, and is used to acquire and store part of the data stream from the multi-modal data stream acquired by the plurality of image acquisition devices when a trigger signal is received.

[0006] A multi-view animal multi-modal data acquisition device provided by the present invention also includes a synchronizer, which is communicatively connected to the multiple image acquisition devices and is used to perform time synchronization processing on the multi-modal data streams acquired by the multiple image acquisition devices.

[0007] According to a multi-view animal multimodal data acquisition device provided by the present invention, the synchronizer is specifically used to obtain the current image frame in the multimodal data stream; store the current image frames in storage containers of corresponding types according to the data types, and record the hardware timestamp of each frame; and filter out image frames that do not meet the synchronization requirements based on the hardware timestamp.

[0008] According to a multi-view animal multi-modal data acquisition device provided by the present invention, the width of the acquisition section is adapted to the body shape of the animal, and the opening end of the inlet section and the opening end of the outlet section are both larger than the width of the acquisition section.

[0009] According to a multi-view animal multi-modal data acquisition device provided by the present invention, the image acquisition equipment support includes a gantry and two upright frames, the gantry includes a front column, a rear column, a top beam, a front left column, a front right column, a rear left column, and a rear right column, the front left column and the front right column are located on both sides of the bottom of the front column, the rear left column and the rear right column are located on both sides of the bottom of the rear column, the gantry spans the middle of the acquisition section, and the two upright frames are respectively located on the front and rear sides of the acquisition section.

[0010] A multi-view animal multi-modal data acquisition device provided by the present invention also includes an animal identification identifier and an infrared grating trigger. The animal identification identifier is arranged on the front column of the gantry and is used to cyclically read the animal ear tag; the infrared grating trigger is arranged on the mesh fence and is used to sense the animal position and send the trigger signal to the control device.

[0011] According to a multi-view animal multi-modal data acquisition device provided by the present invention, the multiple image acquisition devices include multiple multi-modal cameras, and the multiple multi-modal cameras are respectively located on the front pillar, the rear pillar, the top beam, the front left pillar, the front right pillar, the rear left pillar, the rear right pillar and the two stands, and the multiple multi-modal cameras are used to capture three-dimensional morphological information of animals.

[0012] According to a multi-view animal multi-modal data acquisition device provided by the present invention, the data categories collected by the multi-modal camera include: RGB images, depth maps, point cloud data, infrared images, depth maps aligned with RGB, and depth grayscale maps aligned with RGB.

[0013] According to a multi-view animal multi-modal data acquisition device provided by the present invention, the multiple image acquisition devices include two color cameras, one color camera is located between the front left column and the front right column, and the other color camera is located between the rear left column and the rear right column. The two color cameras are used to photograph the abdomen of the animal.

[0014] According to a multi-view animal multi-modal data acquisition device provided by the present invention, there is an acquisition interval between the acquisition times of the multiple image acquisition devices.

[0015] The multi-view animal multi-modal data acquisition device provided by the present invention includes an animal passage, multiple image acquisition devices, an image acquisition device bracket and a control device; since the animal passage is a three-section angled route and is surrounded by two mesh fences, the image acquisition device bracket can enable multiple image acquisition devices to be distributed in all directions of the acquisition section without line of sight obstruction, thereby achieving all-round shooting field of view coverage of the animal without blind spots, and further achieving a comprehensive, efficient, accurate and high-throughput contactless animal data acquisition effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 This is one of the structural schematic diagrams of the multi-view animal multi-modal data acquisition device provided by the present invention; Figure 2 This is the second structural schematic diagram of the multi-view animal multi-modal data acquisition device provided by the present invention; Figure 3 It is a schematic diagram of the data acquisition triggering flow of the multi-view animal multi-modal data acquisition device provided by the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0019] It should be noted that, in the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0020] It should be noted that, in this article, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0021] The embodiments of the present application describe some exemplary embodiments for the purpose of explanation. It should be understood that the present application can be implemented in other ways that are not specifically shown in the drawings.

[0022] An embodiment of the present application provides a multi-view animal multi-modal data acquisition device, including: an animal channel, multiple image acquisition devices, an image acquisition device bracket and a control device.

[0023] like Figure 1 As shown, the two sides of the animal passage 10 are surrounded by two mesh fences, and the animal passage is a three-section angled route, including an inlet section, a collection section and an outlet section connected in sequence.

[0024] Optionally, the mesh fence can be welded with 3-5 mm stainless steel rods and installed on the ground using ground nails, expansion screws, welding, etc. Such a design strength can withstand the brute force collision of animals and prevent the passage from being damaged.

[0025] It should be noted that the grid size of the mesh fence can be optimized according to the size of the animal's head. On the one hand, the gap and structure of the mesh can avoid the animal's head from being stuck and reduce the animal's stress response. On the other hand, compared with traditional railings, it can greatly reduce the obstruction of the image acquisition equipment's line of sight, providing the image acquisition equipment with a clearer and more coherent shooting field of view, thereby ensuring that high-quality data can be collected.

[0026] Optionally, the width of the collecting section is adapted to the body size of the animal, and the opening end of the inlet section and the opening end of the outlet section are both larger than the width of the collecting section.

[0027] Specifically, on the one hand, the appropriate width can prevent animals from feeling crowded or restricted when passing through the collection section, and reduce the stress response of animals. If the collection section is too narrow, the animals may feel uneasy due to the small space and bump around, which will not only affect the quality of the collected data, but also may cause damage to the animals themselves and the equipment; if the collection section is too wide, it cannot be guaranteed that the camera can comprehensively and accurately collect the body shape, appearance and morphological data of the animals from all angles. Therefore, the width of the collection section is adapted to the body shape of the animals, which can effectively ensure the smooth progress of the collection process. On the other hand, the larger opening end of the entrance section can allow the animals to have a wide field of vision when entering the channel, reduce their vigilance and resistance, and allow them to enter the collection area more naturally. When animals walk from a relatively spacious breeding area to the collection channel, they will not have a stress response due to suddenly entering a narrow space. The larger opening end of the exit section allows animals to leave quickly after completing data collection, avoiding their stay in the channel and causing blockage, thereby improving the collection efficiency. In addition, the larger entrance and exit openings also make it easier for breeders to guide animals in and out of the channel, and there is more sufficient operating space when manual intervention is required.

[0028] like Figure 1 As shown, the image acquisition equipment support includes a gantry and two vertical frames. The gantry spans the middle of the acquisition section of the animal channel and includes a front column 31, a rear column 32, a top beam 33, a front left column 34, a front right column 35, a rear left column 36, and a rear right column 37. The front left column 34 and the front right column 35 are located on both sides of the bottom of the front column 31, and the rear left column 36 and the rear right column 37 are located on both sides of the bottom of the rear column 32. The two vertical frames include a vertical frame 38 and a vertical frame 39, which are respectively located on the front and rear sides of the acquisition section.

[0029] Optionally, the image acquisition device bracket may be made of high-strength aluminum alloy.

[0030] Optionally, the image acquisition device bracket is used to place multiple image acquisition devices at the front, rear, left, right and top positions of the acquisition section respectively.

[0031] It is understandable that the image acquisition equipment bracket spans the middle of the animal channel. This structure provides a stable and reasonable support framework for the layout of the image acquisition equipment. The camera mounting columns of the image acquisition equipment bracket are designed and installed according to different parts of the animal's body, which can achieve full coverage of the animal's shooting field of view without blind spots, ensuring that no details of the appearance and morphology are missed.

[0032] The above-mentioned multiple image acquisition devices are used to collect multimodal data of animals from different perspectives, including multiple multimodal cameras and two color cameras. The multiple multimodal cameras are used to capture the three-dimensional morphological information of the animals, and the two color cameras are used to photograph the abdomen of the animals.

[0033] Optionally, the plurality of multimodal cameras are respectively located on the front column, the rear column, the top beam, the front left column, the front right column, the rear left column, the rear right column and the two stands. Of the two color cameras, one color camera is located between the front left column and the front right column, and the other color camera is located between the rear left column and the rear right column. The two color cameras are used to photograph the abdomen of the animal.

[0034] like Figure 1 As shown, the plurality of multimodal cameras include a multimodal camera 21, a multimodal camera 22, a multimodal camera 23, a multimodal camera 24, a multimodal camera 25, a multimodal camera 26, and a multimodal camera 27, and the two color cameras include a color camera 28 and a color camera 29. The multimodal camera 21 is arranged on a stand 38, the multimodal camera 22 is arranged on a stand 39, the multimodal camera 23 is arranged on a front right column 35, the multimodal camera 24 is arranged on a front left column 34, the multimodal camera 25 is arranged on a rear right column 37, the multimodal camera 26 is arranged on a rear left column 36, and the multimodal camera 27 is arranged on a top beam 33. The color camera 28 is arranged between the front left column 34 and the front right column 35, and the color camera 29 is arranged between the rear left column 36 and the rear right column 37.

[0035] Specifically, the multimodal camera 27 can be embedded in the top beam 33 to photograph the back of the animal, obtain the three-dimensional shape of the back, skin and hair texture depth information, etc., and provide key data for the assessment of the animal's body shape, appearance, morphology, and health status.

[0036] The multimodal camera 23 is set on the front right column 35, the multimodal camera 24 is set on the front left column 34, the multimodal camera 25 is set on the rear right column 37, and the multimodal camera 26 is set on the rear left column 36, which can form a large-angle shooting array to comprehensively capture the animal's side profile, limbs, hooves and other data from both sides.

[0037] The multimodal camera 21 is set on the stand 38, and the multimodal camera 22 is set on the stand 39, and is used to photograph the face and buttocks of the animal, focusing on the head features, the fatness of the buttocks, the presence or absence of external injuries, the front limb hooves and the hind limb hooves, etc.

[0038] Color camera 28 and color camera 29 support angle adjustment and front-to-back distance adjustment functions. Color camera 28 and color camera 29 are designed for photographing the abdomen of animals. Since the abdomen of animals is located at a low position and has an irregular shape, the data quality collected by multimodal cameras is poor. Therefore, with the help of the characteristics of RGB color cameras, detailed information such as the color of the animal's abdominal skin, the presence of lumps, the condition of the nipples, and the limbs and hooves can be supplemented.

[0039] It is understandable that multimodal cameras are used simultaneously to capture the three-dimensional morphological information of animals, and have advantages in measuring the body shape and posture of animals. Multiple multimodal cameras are installed at various key angles to ensure that accurate depth data can be obtained from different perspectives; RGB cameras focus on supplementing image information at tricky angles, especially RGB color cameras with adjustable angles and distances, which can flexibly shoot the abdomen of animals and make up for the deficiencies of multimodal cameras in acquiring data at tricky angles and presenting data quality. The two complement each other and can jointly build a complete animal body shape, appearance and morphology image information library.

[0040] Optionally, there is an acquisition interval between the acquisition times of the multiple image acquisition devices.

[0041] Specifically, if multiple image acquisition devices are synchronously acquiring images of the same target, the lasers emitted by the multiple image acquisition devices may interfere with each other. To prevent the lasers from interfering with each other, the camera exposure time offset can be set through the JSON configuration file (for example, the offset can be 160μs or more) to ensure that the laser activity time of each camera is staggered. For example, when the laser of one camera is turned on, the lasers of other cameras are idle.

[0042] Optionally, data categories collected by the multimodal camera include: RGB images, depth maps, point cloud data, infrared images, depth maps aligned with RGB, and depth grayscale images aligned with RGB.

[0043] It is understandable that, since the animal passage is a three-section angled route, the stand 38 and the stand 39 can shoot the animal from a unique angle. This design of the passage changes the animal's walking path, so that when the animal passes through the collection section, its face and buttocks can be more clearly presented within the shooting range of the image acquisition device supported by the two stands. In conjunction with the image acquisition devices at other locations, it is possible to achieve all-round coverage of the animal's shooting field of view without blind spots, ensuring that complete and comprehensive animal body shape, appearance and morphological data are collected.

[0044] Optionally, the multi-view animal multi-modal data acquisition device provided in the embodiment of the present application may also include a synchronizer, which is communicatively connected to the multiple image acquisition devices and is used to perform time synchronization processing on the multi-modal data streams acquired by the multiple image acquisition devices.

[0045] Optionally, the synchronizer is specifically used to obtain the current image frame in the multimodal data stream; store the current image frames in storage containers of corresponding types according to the data types, and record the hardware timestamp of each frame; and filter out image frames that do not meet the synchronization requirements based on the hardware timestamp.

[0046] Specifically, Figure 2As shown in the figure, the multi-view animal multi-modal data acquisition device uses seven multi-modal cameras and completes the multi-machine synchronous data acquisition task through efficient hardware and software cooperation. The whole device includes seven multi-modal cameras, control equipment, synchronizer, USB data cable, network cable, and power supply module. Each camera is connected to the control device through a USB3.0 interface to achieve high-speed data transmission. At the same time, all cameras are connected to the synchronizer through a network cable to ensure accurate hardware-level synchronization. The synchronization accuracy between cameras is effectively controlled by the hardware timestamp mechanism of the camera, but when the bandwidth is insufficient and the signal fluctuates, there will be a phenomenon of asynchrony. Therefore, a hardware timestamp judgment mechanism is added to ensure the accurate synchronization of data frames. The system architecture ensures efficient and stable multi-machine synchronous acquisition, which can meet the strict requirements for acquisition accuracy and timeliness in dynamic environments, and provide reliable data support for subsequent data analysis and processing.

[0047] The multi-view animal multi-modal data acquisition device can achieve multi-modal camera data synchronization through the collaborative work of hardware and software. The hardware provides basic support, and the software is responsible for specific control and data processing. The specific process includes: In the camera configuration stage: When the system starts, the camera must be configured first. If you choose to use the previous configuration, you can directly enter the relevant instructions to enter the open stream program; if you need to reconfigure, load the json configuration file. In this file, the camera's synchronization mode, trigger signal input delay, trigger signal output delay and other key parameters will be set.

[0048] In the open stream program stage: After completing the configuration, you can enter the open stream program, which implements a full-featured data stream startup function to support the synchronous acquisition of depth stream, color stream and infrared stream. The program starts the stream through the pipeline object (pipeline) and configures the frame synchronization mode to ensure that RGB, depth and infrared data can be collected in the same frame. In the specific implementation process, the Config object is used to configure the resolution, frame rate and data format of the RGB stream, depth stream and infrared stream. The specific format (Y16 format) is preferred. If no matching configuration is found, it falls back to the default setting. At the same time, the D2C alignment mode is configured in the program. In the stream startup stage, data acquisition is started, and the callback function is registered to process the real-time stream data. According to different frame types (such as color frames, depth frames or infrared frames), special processing functions are called to process and save the corresponding data. For depth stream data, point cloud data can also be saved and the camera internal parameters of the corresponding device can be updated.

[0049] In the data acquisition and storage stage: after the stream is opened, the color stream data is pushed into the colorFrames container, and the depth stream and infrared stream data are pushed into the depthFrames and irFrames containers respectively. These containers are designed based on the std map of the device index (deviceIndex). Each container can only store the latest frame data of each device, and the new frame will overwrite the previously stored frame when it arrives. After all cameras are opened, the user triggers data acquisition by pressing a button. The program grabs the current frame from the raw frame data container of all cameras, and stores the current frame in the same type of storage container as the raw frame data container according to the data type, and records the hardware timestamp of each frame. By judging the hardware timestamp, the frames that do not meet the synchronization requirements are filtered out. After the synchronization verification is passed, the program saves these frame data to the corresponding folders on the disk. The supported data types include RGB images, depth maps, infrared images, point cloud data, grayscale images, and aligned depth maps. The file name contains the device number and timestamp. The storage path is dynamically generated, and the data is organized in groups for subsequent analysis.

[0050] Optionally, the multi-view animal multi-modal data acquisition device provided in an embodiment of the present application also includes an animal identification identifier and an infrared grating trigger, wherein the animal identification identifier is arranged on the front column of the gantry and is used to cyclically read the animal ear tag; the infrared grating trigger is arranged on the mesh fence and is used to sense the animal position and send the trigger signal to the control device.

[0051] Specifically, Figure 1 As shown, the infrared grating trigger 41 is installed on the side of the front column 31 of the gantry to form a light curtain. When the animal enters the animal channel 10 and contacts the light curtain, the infrared grating trigger 41 immediately generates a trigger signal. On the one hand, the trigger signal is transmitted to the camera control system, and after a short and precise delay setting (the delay time is adjusted and determined according to factors such as the channel length and the average walking speed of the animal), all cameras are started to collect images to ensure that the camera captures the picture of the animal in the collection section; on the other hand, the trigger signal synchronously activates the animal identification identifier, which is installed in the touch box 42. The RFID antenna 43 transmits a high-frequency radio frequency signal to communicate with the RFID chip in the animal ear tag, quickly read the animal's identity information, such as breed, date of birth, immunization record, etc., and associate this information with the collected image data in real time to achieve accurate data matching.

[0052] like Figure 3As shown in the figure, the data acquisition trigger process of the multi-view animal multimodal data acquisition device is shown, which is mainly used to collect data of animals wearing electronic ear tags. There are two trigger conditions: successful reading of the ear tag and successful reading of the grating code. In order to optimize the system performance, the grating code reading operation is placed in the child thread to avoid blocking the main thread. The main thread determines whether the trigger condition is met by cyclically checking two flag variables (F1 represents the ear tag reading status, and F2 represents the grating code reading status). Only when F1 and F2 are both "true", the multimodal data is stored locally.

[0053] After the system is started, the main thread and the child thread are executed in parallel. The main thread starts to read the electronic ear tag code in a loop and determines whether the electronic ear tag code is successfully read. If not, continue to try to read in the current loop; if so, enter the next step, perform a header check on the read ear tag code, and determine whether the check is successful. If not, set the ear tag read flag variable F1 to "false" and restart the loop to read the ear tag code; if so, enter the next step, perform a data check on the ear tag code, and determine whether the check is successful. If not, set the ear tag read flag variable F1 to "false" and restart the loop to read the ear tag code. If yes, set the ear tag read flag variable F1 to "true".

[0054] At the same time, the child thread starts to read the grating code in a loop, waiting for the animal to trigger the grating, and judging whether the grating code is read successfully. If not, it continues to try to read in the current loop; if yes, it enters the next step of verification, performs a header check on the read grating code, and judges whether the check is successful. If not, the read grating flag variable F2 is set to "false", and the grating code is read again in a loop; if yes, it enters the next step, performs a data check on the grating code, and judges whether the check is successful. If not, the read grating flag variable F2 is set to "false", and the grating code is read again in a loop. If yes, the read grating flag variable F2 is set to "true".

[0055] The main thread continues to determine whether the F1 and F2 flag variables are all "true". If not, it continues to loop to read the ear tag code and grating code. If so, the collected multimodal data is stored locally.

[0056] For example, an animal wearing an electronic ear tag starts to pass through the animal channel. When it enters the vicinity of the electronic ear tag reading antenna, the main thread starts trying to read the electronic ear tag code. If the ear tag code is successfully read and the header check and data check are both successful, F1 is set to "true". Then, the animal continues to move forward and passes the gantry, triggering the grating, and the child thread starts trying to read the grating code. If the grating code is successfully read and the header check and data check are also successful, F2 is set to "true", and multi-machine synchronous multi-modal data is collected at the same time. When the main thread detects that both F1 and F2 are "true", the previously collected multi-modal data is stored locally to complete this data collection operation.

[0057] It is understandable that it is equipped with an infrared grating trigger and an animal identification identifier, and utilizes the action of the animal touching the grating to instantly start all cameras to collect images, while quickly reading the animal ear tag information. This not only realizes the automation of the collection process, but also accurately binds the body shape, appearance and morphology data to the individual identity, thereby improving the efficiency and relevance of data collection.

[0058] Optionally, the control device is used to acquire and store a portion of the data stream from the multimodal data stream acquired by the multiple image acquisition devices when a trigger signal is received.

[0059] Optionally, in terms of data storage, the RGB image is saved in PNG format, the point cloud data is saved in PLY format, and the depth map and IR image are saved as image files.

[0060] Optionally, the multi-view animal multi-modal data acquisition device provided in the embodiment of the present application may also include a 15.6-inch industrial-grade touch screen for subsequent visual display of the animal's body shape, appearance and morphological system, so that the operator can easily control camera parameter settings, image preview, data storage path selection, equipment status monitoring and other functions through touch operations, thereby facilitating on-site adjustment of acquisition strategies.

[0061] The host is equipped with a high-performance processor, large-capacity memory and high-speed hard disk to meet the needs of fast storage, processing and transmission of image data. At the same time, the host supports WIFI and 4G communication modules, which facilitates the remote transmission of data to the farm management center or cloud server, and facilitates subsequent data analysis and sharing. In addition to the host, a cooling system is installed in the control box to ensure that the temperature in the control box is stable within an appropriate range and to ensure the normal operation of electronic components. The operating status indicator light is installed in a conspicuous position on the touch box panel, and the operation status of the equipment is intuitively displayed with different colors of light, which is convenient for operators to quickly determine the fault point. The design of external USB interface, hard disk / mouse tray further enhances the expandability and operating comfort of the equipment, facilitates the backup of data on external storage devices, connects debugging tools to troubleshoot faults, and provides comfortable operating support for operators. The knob-type start switch is convenient and easy to use, and the multiple protection designs such as short circuit, leakage and external grounding ensure the safe, stable and efficient operation of the equipment in all directions.

[0062] In the embodiment of the present application, it includes an animal passage, multiple image acquisition devices, an image acquisition device bracket and a control device; because the animal passage is a three-section angle route and is surrounded by two mesh fences, the image acquisition device bracket can enable multiple image acquisition devices to be distributed in all directions of the acquisition section without line of sight obstruction, thereby achieving all-round shooting field of view coverage of the animal without blind spots, and further achieving comprehensive, efficient, accurate and high-throughput contactless animal data acquisition effects.

[0063] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0064] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0065] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-view animal multi-modal data acquisition device, characterized in that: include: Animal access, multiple image acquisition devices, image acquisition device stands, and control equipment; The two sides of the animal passage are surrounded by two mesh fences, and the animal passage is a three-section angled route, including an inlet section, a collection section and an outlet section connected in sequence; The multiple image acquisition devices are used to acquire multimodal data of animals from different perspectives; The image acquisition device bracket is used to place the multiple image acquisition devices at the front position, rear position, left position, right position and top position of the acquisition section respectively; The control device is connected to the multiple image acquisition devices and is used to obtain and store a portion of the data stream from the multimodal data stream acquired by the multiple image acquisition devices when a trigger signal is received.

2. The multi-view animal multi-modal data acquisition device according to claim 1, characterized in that: It also includes a synchronizer, which is communicatively connected to the multiple image acquisition devices and is used to perform time synchronization processing on the multimodal data streams acquired by the multiple image acquisition devices.

3. The multi-view animal multi-modal data acquisition device according to claim 2, characterized in that: The synchronizer is specifically used to obtain the current image frame in the multimodal data stream; store the current image frame in a storage container of a corresponding type according to the data type, and record the hardware timestamp of each frame; Image frames that do not meet synchronization requirements are filtered out based on the hardware timestamp.

4. The multi-view animal multi-modal data acquisition device according to claim 1, characterized in that: The width of the collecting section is adapted to the body size of the animal, and the opening end of the inlet section and the opening end of the outlet section are both larger than the width of the collecting section.

5. The multi-view animal multi-modal data acquisition device according to claim 1, characterized in that: The image acquisition device bracket includes a gantry and two vertical frames. The gantry includes a front column, a rear column, a top beam, a front left column, a front right column, a rear left column, and a rear right column. The front left column and the front right column are located on both sides of the bottom of the front column, and the rear left column and the rear right column are located on both sides of the bottom of the rear column. The gantry spans the middle of the acquisition section, and the two vertical frames are respectively located on the front and rear sides of the acquisition section.

6. The multi-view animal multi-modal data acquisition device according to claim 5, characterized in that: It also includes an animal identification identifier and an infrared grating trigger. The animal identification identifier is arranged on the front column of the gantry and is used to cyclically read the animal ear tag; the infrared grating trigger is arranged on the mesh fence and is used to sense the animal position and send the trigger signal to the control device.

7. The multi-view animal multi-modal data acquisition device according to claim 5, characterized in that: The multiple image acquisition devices include multiple multimodal cameras, which are respectively located on the front pillar, the rear pillar, the top beam, the front left pillar, the front right pillar, the rear left pillar, the rear right pillar and the two stands, and the multiple multimodal cameras are used to capture three-dimensional morphological information of animals.

8. The multi-view animal multi-modal data acquisition device according to claim 7, characterized in that: The data categories collected by the multimodal camera include: RGB images, depth maps, point cloud data, infrared images, depth maps aligned with RGB, and depth grayscale images aligned with RGB.

9. The multi-view animal multi-modal data acquisition device according to claim 5, characterized in that: The multiple image acquisition devices include two color cameras, one color camera is located between the front left column and the front right column, and the other color camera is located between the rear left column and the rear right column. The two color cameras are used to photograph the abdomen of the animal.

10. The multi-view animal multi-modal data acquisition device according to any one of claims 1 to 9, characterized in that: There is an acquisition interval between the acquisition times of the multiple image acquisition devices.

Images