Device and method for non-destructive testing of early hatching viability of fertilized eggs
By combining a robotic arm with an extraction mechanism, multi-view transmission images are acquired and a deep learning model is used to solve the problem of low detection efficiency during the hatching process of fertilized eggs, thus achieving efficient and accurate detection of egg viability.
Patent Information
- Application Number
- CN202511521344.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-23
- Publication Date
- 2026-01-30
- Estimated Expiration
- 2045-10-23
AI Technical Summary
In existing technologies, it is difficult to achieve efficient and automated multi-angle viability detection during the hatching process of fertilized eggs, resulting in low detection efficiency and unsuitability for small-scale breeding facilities.
The system employs a robotic arm in conjunction with a suction mechanism. A vacuum pump provides negative pressure, and a multi-finger gripper and conical suction cup are used to lift the eggs and arrange them in a stepped pattern. An image acquisition mechanism is used to obtain multi-view transmission images, and a deep learning model is used for viability detection.
It improves detection accuracy without increasing space and time costs, avoids incomplete detection and misjudgment caused by single-view imaging, and is suitable for small-scale breeding facilities.
Smart Images

Figure CN120992614B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of agricultural product testing, and in particular to a non-destructive testing device and method for detecting the early-stage viability of hatching eggs. Background Technology
[0002] In the poultry hatching industry, live embryo detection involves selecting and removing dead embryos and infertile eggs that have ceased development during incubation. This is a crucial process directly impacting hatching efficiency and economic benefits. Statistics show that the fertilization rate of hatching eggs under natural conditions is approximately 85%, but "dead embryos" can still occur during incubation. This phenomenon typically persists throughout the entire developmental cycle of the egg and is primarily caused by factors such as unhealthy mother eggs, microcracks in the eggshell, egg deformities, prolonged storage, or unsuitable incubation conditions. Statistics indicate that the incidence of dead embryos is approximately 18%. Dead embryos in high-temperature and high-humidity environments decompose rapidly, fostering the growth of microorganisms such as Pseudomonas bacteria. These microorganisms decompose proteins and fats within the egg, producing hydrogen sulfide. When hydrogen sulfide accumulates to a certain concentration, it can cause the eggs to burst, posing a serious safety hazard to surrounding normally developing fertilized eggs. Therefore, timely removal of dead embryos from the incubator is of great importance for conserving hatching resources and protecting the safe development of normal embryos.
[0003] Currently, in the poultry hatching industry, there are devices that utilize transmission imaging for inspection. However, when performing batch inspections, these devices often only allow for a fixed viewing angle, such as shooting from above or below, and cannot capture side views. For multi-angle inspections, a rolling mechanism is typically used, but this can only inspect one egg at a time, resulting in low efficiency. Furthermore, batch inspections usually rely on conveyor belt transport, which consumes a significant amount of space and is only suitable for large-scale factories, not small-scale farms, thus failing to meet the demands of modern, efficient, and automated poultry farming. Summary of the Invention
[0004] To address the problems existing in the prior art, this invention provides a non-destructive testing device and method for early-stage hatching viability detection of fertilized eggs.
[0005] This invention provides a non-destructive testing device for early-stage viability detection of hatching eggs, comprising: a testing device body and a processing module; the testing device body includes a vacuum pump, a detection dark chamber, a robotic arm disposed at the top of the detection dark chamber, a suction mechanism disposed at the end of the robotic arm, and an image acquisition mechanism disposed on the inner side of the detection dark chamber; the suction mechanism includes a multi-finger gripper, a hollow main rod, multiple hollow auxiliary rods perpendicular to the hollow main rod and parallel to each other, an electric push rod, a suction cup connecting rod, a conical suction cup, and a patch light source; one end of the multi-finger gripper is fixedly connected to the end of the robotic arm, and the other end is connected to the hollow main rod. The hollow auxiliary rods are connected by the electric push rod. When the electric push rod extends, each hollow auxiliary rod expands synchronously and uniformly to maintain a predetermined interval between the eggs. A row of suction cup connecting rods is provided below each hollow auxiliary rod. A conical suction cup is fixed below each suction cup connecting rod by a deep groove ball bearing. A patch light source is fixed on the inner side of each conical suction cup. The conical suction cup, suction cup connecting rods, hollow auxiliary rods, and hollow main rod are sequentially connected by air channels. The hollow main rod is connected to a vacuum pump through a gas delivery hose. The processing module is used to acquire the transmission side view image of the eggs through the image acquisition mechanism and determine the morphological detection results of each egg based on the transmission side view image.
[0006] According to the present invention, a non-destructive testing device for early-stage hatching viability of hatching eggs is provided, wherein parallel pulley guide rails are fixed on both sides below the hollow auxiliary rod, and a pulley bracket for sliding on the pulley guide rails is provided on the hollow auxiliary rod. The hollow auxiliary rods are connected by electric push rods and reciprocate along the pulley guide rails.
[0007] According to the present invention, a non-destructive testing device for early-stage viability detection of hatching eggs is provided. The multi-finger clamp includes a mechanical arm end connector and a cross-shaped fixing bracket fixedly connected to the mechanical arm end connector. Each of the four branches of the cross-shaped fixing bracket is provided with a sliding groove. Each sliding groove is slidably connected to the top of a slider. The end of each slider is detachably connected to a clamping finger. The clamping finger is used to clamp and fix a hollow main rod and a hollow secondary rod located at the longitudinal center.
[0008] According to the present invention, a non-destructive testing device for early-stage hatching viability of hatching eggs is provided, wherein each hollow auxiliary rod is a detachable structure to enable simultaneous testing of hatching eggs in different rows.
[0009] According to the present invention, a non-destructive testing device for early-stage hatching viability of hatching eggs is provided, wherein the conical suction cup in the suction mechanism has a multi-layer flexible folded edge structure.
[0010] According to the present invention, a non-destructive testing device for early-stage hatching viability of fertilized eggs is provided, wherein the patch light source is a conical spiral light source.
[0011] According to the present invention, a non-destructive testing device for early-stage viability detection of hatching eggs is provided. The image acquisition mechanism includes a target position detector, a sliding guide rail module, and a camera. The target position detector is used to confirm that the suction mechanism has reached the predetermined imaging position. The camera moves on the sliding guide rail module.
[0012] The non-destructive testing device for early hatching viability of hatching eggs provided by the present invention further includes an interactive module for receiving detection parameters input by a user, and a display screen for displaying the hatching viability test results.
[0013] According to the present invention, a non-destructive detection device for early hatching viability of hatching eggs is provided. The processing module includes: an image acquisition unit for acquiring a transmissive side view image of the hatching egg through a camera; and a trait detection unit for inputting the transmissive side view image into a trained viability detection model and outputting the trait detection result of each hatching egg. The viability detection model is constructed based on the YOLOv11 model, and the Context Guided module replaces the CBS module to improve the C3K2 feature enhancement module, the Bi-FPN architecture replaces the PAFPN architecture, and the model is trained using EMASlideLoss as the loss function.
[0014] This invention also provides a non-destructive testing method for the early-stage viability of hatching eggs, implemented based on the aforementioned device. The method includes: upon receiving a work instruction, a robotic arm moves the extraction mechanism to a predetermined initial collection position on the egg to be tested according to a pre-set motion trajectory; a vacuum pump is activated, outputting a constant negative pressure to the spiral conical suction cup through a sealed gas supply pipeline; after initial adsorption, based on a multi-point detection path plan generated by the row and column geometric arrangement, the robotic arm drives the extraction mechanism to sequentially reach each target detection point along the planned trajectory; upon approaching the target position, an electric push rod extends, driving the hollow auxiliary rods on both sides to expand outward synchronously and uniformly; after the target position detector of the image acquisition mechanism confirms that the extraction mechanism is in place, the camera is controlled to acquire detection images in real time according to position parameters; the detection images are sent to a processing module, outputting the morphological detection results for each hatching egg.
[0015] The present invention provides a non-destructive testing device and method for early-stage hatching viability detection of hatching eggs. Through the close cooperation of a robotic arm and an extraction mechanism, the entire tray of hatching eggs is distributed in a stepped manner at the image acquisition position. The image acquisition mechanism can simultaneously acquire multiple side-view transmission images of hatching eggs, or transmission images at specific angles, and can also acquire multi-view transmission side-view images. Multi-view images can comprehensively reflect the characteristics of blood vessels and embryonic tissues within the hatching eggs. Viability-related features can be easily extracted from these images, improving detection accuracy without increasing additional space and time costs, and avoiding the problems of incomplete detection and misjudgment caused by single-view imaging. Attached Figure Description
[0016] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0017] Figure 1 This is one of the main views of the non-destructive testing device for early-stage hatching viability of fertilized eggs provided by the present invention;
[0018] Figure 2 This is the second front view of the non-destructive testing device for early-stage hatching activity of fertilized eggs provided by the present invention;
[0019] Figure 3 This is a perspective view of the overall structure of the non-destructive testing device for early-stage hatching viability of fertilized eggs provided by the present invention.
[0020] Figure 4 This is a front view of the image acquisition process of the non-destructive testing device for early-stage viability detection of hatching eggs provided by this invention;
[0021] Figure 5 This is a side cross-sectional schematic diagram of the non-destructive testing device for early-stage hatching viability of hatching eggs provided by the present invention;
[0022] Figure 6 This is a front view of the extraction mechanism of the non-destructive testing device for early-stage viability detection of hatching eggs provided by the present invention;
[0023] Figure 7 This is a side view of the extraction mechanism of the non-destructive testing device for early-stage viability detection of hatching eggs provided by the present invention;
[0024] Figure 8 This is a perspective view of the extraction mechanism of the non-destructive testing device for early-stage viability detection of hatching eggs provided by the present invention;
[0025] Figure 9 This is a perspective view of the multi-finger clamp of the non-destructive testing device for early-stage viability detection of hatching eggs provided by the present invention;
[0026] Explanation of reference numerals in the attached diagram: 1-Detection dark box; 2-Display screen and control switch; 3-Device support frame; 4-Vacuum pump; 5-Detection dark box cabinet door; 6-Gas supply hose; 7-Robotic arm; 8-Suction mechanism; 9-Robotic arm base; 10-Robotic arm main arm; 11-Robotic arm upper arm; 12-Robotic arm forearm; 13-Image acquisition mechanism; 14-Target position detector; 15-Slide table guide rail module; 16-Camera; 17-Multi-finger gripper; 18-Pulley bracket; 19-Hollow auxiliary rod; 20-Suction cup connecting rod; 21-Conical suction cup; 22-Egg to be inspected; 23-Hollow main rod; 24-Sealing cover; 25-Electric push rod; 26-Pulley guide rail; 27-Deep groove ball bearing; 28-Robotic arm end connector; 29-Slider; 30-Finger gripper. Detailed Implementation
[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.
[0028] Achieving large-scale, efficient, and non-destructive viability testing of multi-breed hatching eggs in the early stages of incubation is a key research direction for the future of the poultry hatching industry. Building upon existing testing processes and incorporating advanced deep learning technology, a multi-breed online viability testing system for early-stage hatching eggs is being developed. This system can transform the traditional manual candling-based viability assessment method into an intelligent and automated identification mode. Therefore, employing accurate and efficient non-destructive testing methods to assess the viability of hatching eggs before incubation and in the early stages of incubation is of great significance and can effectively improve the economic benefits of the poultry hatching industry.
[0029] The following is combined with Figures 1-9 This invention describes a non-destructive testing device and method for early-stage hatching viability detection of fertilized eggs. Figure 1 This is one of the main views of the non-destructive testing device for early-stage hatching viability of fertilized eggs provided by the present invention. Figure 2 To close the main view of the inspection cabinet door, Figure 3 It is an oblique view. Figure 4 This is the main view of the image acquisition process. Figure 5 This is a side view sectional diagram. Figure 6 This is the main view of the suction mechanism. Figure 7 This is a side view of the suction mechanism. Figure 8 This is a perspective view of the suction and extraction mechanism. (Example) Figures 1-8 As shown, the present invention provides a non-destructive testing device for the early incubation activity of hatching eggs, comprising:
[0030] The detection device body and processing module; the detection device body includes a vacuum pump 4, a detection dark box 1, a robotic arm 7 set at the top inside the detection dark box 1, a suction mechanism 8 set at the end of the robotic arm 7, and an image acquisition mechanism 13 set on the inner side of the detection dark box 1.
[0031] The suction and lifting mechanism 8 includes a multi-finger clamp 17, a hollow main rod 23, multiple hollow auxiliary rods 19 that are perpendicular to the hollow main rod 23 and parallel to each other, an electric push rod 25, a suction cup connecting rod 20, a conical suction cup 21, and a patch light source.
[0032] One end of the multi-finger clamp 17 is fixedly connected to the end of the robotic arm 7, and the other end is connected to the hollow main rod 23 and the hollow secondary rod located in the longitudinal center. The hollow secondary rods 19 are connected by an electric push rod 25. When the electric push rod 25 extends, each hollow secondary rod 19 extends synchronously and uniformly to maintain the predetermined interval between the hatching eggs.
[0033] Below each hollow auxiliary rod 19 is a row of suction cup connecting rods 20. Below each suction cup connecting rod 20 is a conical suction cup 21 fixed by a deep groove ball bearing 27. A patch light source is fixed on the inner side of each conical suction cup 21. The conical suction cup 21, suction cup connecting rods 20, hollow auxiliary rods 19 and hollow main rod 23 are connected in an airtight manner in sequence. The hollow main rod 23 is connected to the vacuum pump 4 through a gas supply hose 6.
[0034] The processing module is used to acquire a transmissive side view image of the hatching egg through the image acquisition mechanism 13, and determine the morphological detection result of each hatching egg based on the transmissive side view image.
[0035] The top of the multi-finger gripper 17 is connected to the end of the robotic arm 7 by bolts, and the bottom of the multi-finger gripper 17 is connected to the hollow main rod 23 and the hollow secondary rod 19 located at the longitudinal center. The hollow main rod 23, the hollow secondary rod 19 of the same specification, and the suction cup connecting rod 20 together constitute the frame and pneumatic passage of the suction mechanism 8. The hollow main rod 23 and the hollow secondary rod 19 are connected by air channels, and each hollow secondary rod 19 is connected by an electric push rod 25. The vacuum pump 4 can be set in the exhaust chamber at the top outside the detection dark box 1.
[0036] The hollow main rod 23 serves as the backbone support component of the entire suction mechanism. One end has an open air passage running through it, connected to the vacuum pump 4 via a flexible air supply hose 6. This allows it to continuously provide high-intensity negative pressure to the conical suction cup 21 (which can be a spiral conical suction cup) and the patch light source, ensuring the stability and reliability of the adsorption and bonding process. The hollow auxiliary rods 19 located on both sides of the hollow main rod 23 maintain the same dimensions as the hollow auxiliary rod 19 located at the longitudinal center. They are used to connect the conical suction cup 21 and the patch light source, while also carrying and distributing the negative pressure airflow from the main rod.
[0037] In addition to being connected through the main air passage, the hollow auxiliary rods 19 in different rows are also equipped with electric push rods 25 and pulley rails 26 on the outside. After the controller sends a command, the electric push rods 25 extend and drive the hollow auxiliary rods 19 to extend to both sides, expanding the subsequent image acquisition space and ensuring that the eggs to be inspected in adjacent rows will not overlap.
[0038] Embedding the lighting source inside the conical suction cup 21 reduces the space requirement for external light source placement, making the overall device more compact and advantageous for use in small or complex spaces (such as darkrooms).
[0039] The cone-shaped suction cup 21 can be made of liquid silicone, which is soft and highly elastic, able to conform to eggs of various shapes, and has good temperature resistance (150~200℃). The outer surface of the cone-shaped suction cup 21 is coated with black tactile paint to prevent light leakage. The inner surface is smoothed to help reflect and guide light. Anti-slip stripes are added to the lower edge of the suction surface to further improve the friction of the contact surface.
[0040] Upon receiving a controller command, the robotic arm 7 drives the suction mechanism 8 to the starting point of the hatching eggs 22 to be inspected. Simultaneously, the vacuum pump 4 starts, providing negative pressure to the conical suction cups 21. The suction mechanism 8 can then lift multiple rows of hatching eggs 22. The initial arrangement of the conical suction cups 21 can be consistent with the arrangement of the egg trays. Based on the number of rows of hatching eggs, the target detection point is planned. The robotic arm 7 drives the suction mechanism 8 to the target detection point, and the electric push rod 25 extends, causing the hollow auxiliary rod 19 to extend to both sides, expanding the subsequent image acquisition space and ensuring that adjacent rows of hatching eggs 22 do not overlap. The required angle images can be acquired through the movement of the robotic arm 7. Furthermore, multi-angle image acquisition can be performed when the robotic arm 7 moves to different positions.
[0041] The non-destructive testing device for early hatching viability of hatching eggs of the present invention, through the close cooperation of the robotic arm 7 and the suction mechanism 8, makes the entire tray of hatching eggs distributed in a stepped manner at the image acquisition position. The image acquisition mechanism 13 can simultaneously acquire multiple side transmission images of hatching eggs, or transmission images at specific angles, and can also acquire multi-view transmission side images. Multi-view images can comprehensively reflect the characteristics of blood vessels and embryonic tissues in hatching eggs. Viability-related features can be easily extracted through these images, improving detection accuracy without increasing additional space and time costs, and avoiding the problems of incomplete detection and misjudgment caused by single-view imaging.
[0042] In some embodiments, the device also includes a display screen and a control switch 2 disposed on the outer side of the detection dark box 1. In addition, the detection device can be mounted on the device support frame 3, the detection dark box can be provided with a detection dark box door 5, and the robotic arm 7 can be disposed on the robotic arm base 9 on the top of the detection dark box 1. The robotic arm 7 includes a main robotic arm 10, a large robotic arm 11, and a small robotic arm 12.
[0043] In some embodiments, parallel pulley guide rails 26 are fixed on both sides below the hollow auxiliary rod 19, and a pulley bracket 18 for sliding on the pulley guide rail 26 is provided on the hollow auxiliary rod 19. The hollow auxiliary rods 19 are connected by electric push rods 25 and reciprocate along the pulley guide rail 26.
[0044] Specifically, a pulley bracket 18 is provided on the hollow auxiliary rod 19 to ensure that the hollow auxiliary rod 19 can move along the pulley guide rail 26 and the hollow main rod 23. Below the pulley guide rail 26, a suction cup connecting rod 20 is provided on the hollow auxiliary rod 19, and a conical suction cup 21 is installed below the suction cup connecting rod 20. The patch light source is installed on the inner side of the conical suction cup 21. The egg to be inspected 22 is in close contact with the conical suction cup 21, ensuring that the suction mechanism can suck the egg to be inspected 22 to the target detection point. The reciprocating movement of each hollow auxiliary rod along the pulley guide rail 26 can improve the stability during the movement.
[0045] In some embodiments, the multi-finger gripper 17 includes a robotic arm end connector 28 and a cross-shaped fixing bracket fixedly connected to the robotic arm end connector 28. Each of the four branches of the cross-shaped fixing bracket is provided with a sliding groove, each sliding groove being slidably connected to the top of a slider 29. The end of each slider 29 is detachably connected to a gripping finger 30, which is used to grip and fix the hollow main rod and the hollow secondary rod 19 located at the longitudinal center.
[0046] like Figure 9 As shown, the slider 29 can slide along the groove of the cross-shaped fixing bracket below the connector 28 at the end of the robotic arm, causing the gripping fingers 30 below the multi-finger gripper 17 to move axially along the hollow main rod 23 and the hollow secondary rod 19 located at the longitudinal center, to provide different magnitudes of torque to ensure the gripping stability of the hollow main rod 23 and the hollow secondary rod 19 located at the longitudinal center. The hollow main rod 23 and the hollow secondary rod 19 located at the longitudinal center cooperate with the multi-finger gripper 17 and are the main support components of the entire suction and lifting mechanism 8. Each gripping finger 30 of the multi-finger gripper 17 can be freely disassembled, and the collar-type connector gripping fingers can be replaced with different shapes of adapter connectors according to the appearance and surface material characteristics of the gripped target.
[0047] In some embodiments, each hollow auxiliary rod 19 is a detachable structure to enable simultaneous testing of hatching eggs in different rows. The hollow auxiliary rod 19 can be freely assembled and disassembled according to the actual application scenario to achieve single-row, three-row, or five-row testing requirements. After disassembly, it can be sealed using the sealing cap 24.
[0048] In some embodiments, the conical suction cup 21 in the suction mechanism 8 has a multi-layered flexible folded edge structure. This facilitates the formation of a stable negative pressure contact with the surface of the hatching egg, improves adsorption stability, reduces mechanical damage to the eggshell, and is suitable for processing hatching eggs of various sizes.
[0049] In some embodiments, the patch light source is a conical spiral light source. The conical spiral light source enables multi-angle, multi-directional illumination, providing uniform transmitted illumination to each hatching egg, effectively reducing blind spots, and is suitable for visual inspection of hatching egg activity. The conical structure has natural light-guiding properties, helping to concentrate light at the center of the egg and improving light transmission efficiency. Furthermore, this structural design significantly increases the scattering area, resulting in low heat concentration and facilitating long-term stable use.
[0050] In some embodiments, the image acquisition mechanism 13 includes a target position detector 14, a slide rail module 15, and a camera 16; the target position detector 14 is used to confirm that the suction mechanism 8 has reached the predetermined image capture position; the camera 16 moves on the slide rail module.
[0051] The image acquisition mechanism 13 is located on the inner side of the detection dark box 1, and a fixed target position detector 14, a sliding table guide module 15, and a movable camera 16 are mounted on it. The camera 16 can move freely along the sliding table guide module 15 to ensure that the optimal acquisition point can be found. The target position detector 14 is fixed above the image acquisition mechanism to ensure accurate detection of the suction and extraction mechanism 8 reaching the target detection point.
[0052] The target position detector 14 is responsible for monitoring the spatial position of the suction mechanism 8 in real time. The target position detector 14 and the positioning algorithm determine whether the suction mechanism 8 has reached the preset detection point. When the target position detector 14 detects that the suction mechanism 8 has reached the target detection point, the target position detector 14 sends the location of the best shooting point to the camera 16. The camera 16 moves to the corresponding position and begins to acquire images.
[0053] In some embodiments, the system further includes an interaction module for receiving detection parameters input by a user, and a display screen for displaying the results of the egg viability test.
[0054] The interaction module receives user-inputted detection parameters and methods, while the display screen shows the detected images of the hatching eggs and the detection report on egg viability. The interaction module can be implemented via buttons or a touch-screen display. Users can select communication methods, detection parameters, detection results, and detection methods (offline / online) through the interaction module. In offline detection mode, this invention can process local images or videos; in online detection mode, it can process images acquired in real time. The interaction module can transmit the input images or videos to the backend viability detection model, which performs inference calculations on the local images or videos and displays the detection results on the display screen. Furthermore, the file path of the tested images, as well as the number, category, detection confidence score, and detection frame coordinates of each target object, are displayed in rows in the detection report window at the lower right of the display screen.
[0055] In some embodiments, a control module is also included for controlling the start and stop of the vacuum pump 4, the movement of the robotic arm 7, the movement of the suction mechanism 8, and the start and stop of the image acquisition mechanism 13. Upon receiving the operation command from the upper controller, the robotic arm 7 moves the suction mechanism 8 to the predetermined starting collection position of the first egg to be inspected 22 with high precision and high speed, following a pre-set motion trajectory. Simultaneously, the vacuum pump 4 is started synchronously, continuously outputting a constant and high-intensity negative pressure to the conical suction cup 21 through a sealed gas supply pipeline, ensuring that the suction mechanism 8 can reliably adsorb and lift multiple rows of eggs to be inspected 22 in a single operation.
[0056] After initial adsorption, the system, based on the multi-point detection path plan generated earlier based on row and column geometry, drives the robotic arm 7 to move the suction mechanism 8 along the planned trajectory to each target detection point sequentially. Upon approaching each target position, the electric push rod 25 automatically extends, driving the hollow auxiliary rods 19 on both sides to expand outwards synchronously and at a uniform speed, further expanding the visual acquisition workspace below. This ensures that adjacent eggs maintain a predetermined safe distance, effectively avoiding image overlap or distortion caused by egg bodies occluding each other. Once the target position detector 14 confirms that the suction mechanism 8 is precisely in place via a high-precision sensor, it immediately transmits the optimal shooting position parameters back to the camera 16 in real time. Upon receiving the feedback, the camera 16 quickly performs fine-tuning and positioning, and activates the high-speed shooting camera module to capture high-resolution, non-overlapping images of the egg surface, providing reliable visual data for subsequent non-destructive testing algorithms based on machine vision and image processing.
[0057] In some embodiments, the processing module includes: an image acquisition unit for acquiring a transmissive side view image of the hatching egg via a camera; and a trait detection unit for inputting the transmissive side view image into a trained survival detection model and outputting the trait detection result for each hatching egg; wherein the survival detection model is built based on the YOLOv11 model, and the Context Guided module replaces the CBS module to improve the C3K2 feature enhancement module, the Bi-FPN architecture replaces the PAPPN architecture, and it is trained using EMASlideLoss as the loss function.
[0058] High-resolution images of multiple rows of hatching eggs lifted simultaneously below the extraction mechanism 8 are acquired. These image data are then fed into a deep learning detection model pre-trained and optimized using a large-scale labeled dataset. The model analyzes each hatching egg independently, outputs a corresponding viability determination result, and automatically classifies the eggs into "live eggs" or "dead embryo eggs" (abnormal eggs) based on the determination score, thereby achieving fully automated, non-destructive, and highly accurate early-stage hatching viability screening.
[0059] Multi-image fusion data enhancement and network structure optimization enable the model to be both efficient and accurate under edge computing conditions. This invention uses a Context Guided downsampling module to enhance global information perception, improves the C3K2 feature enhancement module to reduce the number of parameters and enhance expressive power, replaces the traditional PAFPN with a bidirectional feature pyramid (BiFPN) to improve multi-scale fusion efficiency, optimizes the detection head structure and introduces the EMASlideLoss loss function to enhance the recognition ability of a small number of difficult samples. Without increasing additional computing power and deployment costs, it achieves real-time detection with high accuracy and high efficiency.
[0060] Based on the above embodiments, in order to further describe the solution of the present invention, the following example is used for illustration:
[0061] The experiment selected 30 duck eggs from three common breeds: Cherry Valley, Muscovy, and Basilica Gorilla, with 10 eggs from each breed. The eggs were grouped by breed and numbered at the bottom. Before incubation, the surface of the eggs was wiped clean with a 75(±5)% alcohol solution and then air-dried. After 6 days of incubation in an intelligent incubator, the eggs were retrieved in batches to test the performance of the designed device. On the 10th day of incubation, experienced workers determined the viability of each egg using candling. A total of 20 viable eggs, 6 infertile eggs, and 4 dead embryos were found. The test results showed that the overall detection accuracy of the device was 90.90%, the average detection accuracy for viable eggs was 92.01%, and the average detection accuracy for abnormal eggs was 89.80%. The experimental results demonstrate that the device has certain application prospects and provides technical support for the subsequent development of multi-breed egg viability detection devices in the early stages of incubation.
[0062] The following describes the non-destructive testing method for early hatching viability of fertilized eggs provided by the present invention. The non-destructive testing method for early hatching viability of fertilized eggs described below can be referred to in correspondence with the non-destructive testing device for early hatching viability of fertilized eggs described above.
[0063] This invention also provides a non-destructive testing method for the early-stage viability of hatching eggs, implemented based on the aforementioned device. The method includes: upon receiving a work instruction, a robotic arm moves the extraction mechanism to a predetermined initial collection position on the egg to be tested according to a pre-set motion trajectory; a vacuum pump is activated, outputting a constant negative pressure to the spiral conical suction cup through a sealed gas supply pipeline; after initial adsorption, based on a multi-point detection path plan generated by the row and column geometric arrangement, the robotic arm drives the extraction mechanism to sequentially reach each target detection point along the planned trajectory; upon approaching the target position, an electric push rod extends, driving the hollow auxiliary rods on both sides to expand outward synchronously and uniformly; after the target position detector of the image acquisition mechanism confirms that the extraction mechanism is in place, the camera is controlled to acquire detection images in real time according to position parameters; the detection images are sent to a processing module, outputting the morphological detection results for each hatching egg.
[0064] The method embodiments provided in this invention are implemented through the above-described device embodiments. For specific processes and details, please refer to the above-described device embodiments, which will not be repeated here.
[0065] The method for non-destructive testing of early-stage hatching viability of fertilized eggs provided in this embodiment of the invention has the same implementation principle and technical effect as the aforementioned non-destructive testing device embodiment for early-stage hatching viability of fertilized eggs. For the sake of brevity, any parts not mentioned in the embodiment of the method for non-destructive testing of early-stage hatching viability of fertilized eggs can be referred to the corresponding content in the aforementioned non-destructive testing device embodiment for early-stage hatching viability of fertilized eggs.
[0066] In summary, the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0067] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.
[0068] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A non-destructive testing device for the early incubation activity of hatching eggs, characterized in that, The utility model relates to a kind of detection device and processing module comprising: The detection device body comprises a vacuum pump, a detection dark box, a mechanical arm arranged at the top of the detection dark box, a suction mechanism arranged at the end of the mechanical arm, and an image acquisition mechanism arranged on the side of the detection dark box. The suction mechanism comprises a multi-fingered gripper, a hollow main rod, a plurality of hollow auxiliary rods perpendicular to the hollow main rod and parallel to each other, an electric push rod, a suction disc connecting rod, a conical suction disc, and a patch light source. The multi-fingered gripper is fixedly connected to the end of the mechanical arm at one end and connected to the hollow main rod at the other end. The electric push rod is connected between the hollow auxiliary rods. When the electric push rod is extended, each hollow auxiliary rod expands synchronously and uniformly to expand the subsequent image acquisition space, thereby maintaining a predetermined interval between the eggs and avoiding image overlap or distortion caused by egg body occlusion. Each hollow auxiliary rod is provided with a row of suction disc connecting rods below.
2. The non-destructive testing device for early stage viability of hatching eggs according to claim 1, wherein Each suction disc connecting rod is fixed with a conical suction disc below through a deep groove ball bearing.
3. The non-destructive testing device for early stage viability of hatching eggs according to claim 1, wherein The inner side of each conical suction disc is fixed with a patch light source.
4. The non-destructive testing apparatus for early stage viability of incubated eggs according to claim 1, wherein The conical suction disc, suction disc connecting rod, hollow auxiliary rod, and hollow main rod are sequentially connected by air passages.
5. The non-destructive testing device for early stage viability of hatching eggs according to claim 1, wherein The hollow main rod is connected to the vacuum pump by a gas delivery hose.
6. The non-destructive testing device for early stage viability of incubated eggs according to claim 1, wherein The processing module is used to acquire the transmission side image of the eggs by the image acquisition mechanism and determine the trait detection result of each egg according to the transmission side image.
7. The non-destructive testing device for early stage viability of hatching eggs according to claim 1, wherein The hollow auxiliary rods are fixed with parallel pulley guide rails on both sides below. The hollow auxiliary rods are provided with pulley supports for sliding on the pulley guide rails. The electric push rod connects between the hollow auxiliary rods and reciprocates along the pulley guide rails.
8. The non-destructive testing device for early stage viability of incubated eggs according to claim 1, wherein The multi-fingered gripper comprises a mechanical arm end connector, a cross fixed support fixedly connected to the mechanical arm end connector, and a sliding groove on each branch of the cross fixed support.
9. The non-destructive testing device for early stage viability of hatching eggs according to claim 1, wherein, The end of each sliding block is detachably connected to a finger. Each hollow auxiliary rod is a detachable structure to simultaneously detect eggs of different rows. The conical suction disc in the suction mechanism is a multi-layer flexible folded edge structure. The patch light source is a conical spiral light source. The image acquisition mechanism comprises a target position detector, a sliding table guide rail module, and a camera. The target position detector is used to confirm that the suction mechanism has reached the predetermined shooting position. The camera moves on the sliding table guide rail module. The utility model further comprises an interaction module for receiving detection parameters input by a user and a display screen for displaying the survival rate detection result of the eggs. The processing module comprises: An image acquisition unit for acquiring the transmission side image of the eggs by the camera; A trait detection unit for inputting the transmission side image into a trained survival detection model and outputting the trait detection result of each egg. The survival detection model is constructed based on a YOLOv11 model, and a Context Guided module is used to replace a CBS module to improve a C3K2 feature enhancement module, a Bi-FPN architecture is used to replace a PAFPN architecture, and an EMASlideLoss is used as a loss function for training.
10. A method for non-destructive testing of the viability of an egg in the early stages of incubation based on the device for non-destructive testing of the viability of an egg in the early stages of incubation according to any one of claims 1 to 9, characterized in that, The method comprises the following steps: After receiving the job instruction, the robot arm moves the suction mechanism to the predetermined starting collection position of the eggs to be detected according to the pre-set motion trajectory; The vacuum pump is started, and constant negative pressure is output to the spiral conical suction disc through the sealed gas conveying pipeline; After completing the initial adsorption, based on the multi-point detection path planning generated based on the row-column geometric arrangement, the robot arm drives the suction mechanism to reach each target detection point along the planned trajectory in turn; When approaching the target position, the electric push rod is extended, driving the two hollow auxiliary rods to expand outward at a uniform speed; After the target position detector of the image acquisition mechanism confirms that the suction mechanism has been positioned, the camera acquires detection images in real time according to the position parameters; The detection images are sent to the processing module, and the trait detection results of each egg are output.
Citation Information
Patent Citations
Egg online nondestructive detection system
CN110161121A
Abnormal egg detection device
JP2019158837A