Broiler weight estimation system based on two-view image
The broiler weight estimation system based on two-view images utilizes top-view and side-rear-view cameras combined with radio frequency identification technology to achieve efficient, accurate, and automated measurement of broiler weight. This solves the difficulties in broiler weight measurement in existing technologies, adapts to the activity characteristics of broilers, and improves data acquisition efficiency and accuracy.
Patent Information
- Application Number
- CN202520610923.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Utility models(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-02
- Publication Date
- 2026-02-10
- Estimated Expiration
- 2035-04-02
AI Technical Summary
Existing technologies for measuring broiler weight suffer from problems such as time-consuming and labor-intensive manual operation, stress response, inconsistent measurement data, and large errors. Furthermore, existing machine vision technology is not ideal for broilers due to their small size and frequent activity.
A broiler weight estimation system based on two-view images is adopted. The system uses a top-view camera and a side-rear-view camera to acquire the top view and side-rear view of the broiler, respectively. The image acquisition is triggered by RFID tags and sensors, and the weight is automatically estimated by a weight estimation device. Consumer-grade infrared cameras and supplementary lights are used to improve image quality.
It achieves efficient and accurate broiler weight estimation, reduces system power consumption, improves data acquisition efficiency and accuracy, adapts to the activity characteristics of broilers, reduces stress response, and is suitable for automated management of large-scale farms.
Smart Images
Figure CN223896885U_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The utility model belongs to computer vision technical field, especially relate to a broiler weight estimation system based on two view images. BACKGROUND
[0002] With the continuous development of poultry breeding industry, weight as an important indicator of broiler growth condition, occupies the vital position in production management and genetic breeding. Weight data not only can reflect the health status and nutrition level of broiler, but also provides a scientific basis for individual selection in genetic breeding, at the same time, plays a key role in the production management of breeding farm, helps to determine the best market time, avoids the economic loss caused by overfeeding. Therefore, how to accurately and quickly obtain the broiler weight information has important significance for the efficient management and breeding selection of breeding farm.
[0003] In the prior art, the measurement of broiler weight mainly depends on two types of methods, namely contact type and non-contact type. Contact type measurement usually adopts electronic scale, measuring ruler and other equipment, and the weight or body size data is collected by manual operation. Although this method has certain measurement accuracy, it has many limitations in practical application. First, manual operation is time-consuming and laborious, which is difficult to meet the production needs of large-scale breeding farm. Secondly, contact type measurement is easy to cause stress reaction of animals, which affects their normal growth and health status. In addition, the accuracy of measurement data is easily affected by the experience and subjective judgment of the operator, and the data consistency is poor, which affects the decision accuracy and breeding accuracy. Finally, the weighing equipment may have measurement error under the interference of animal behavior, which further affects the reliability of data.
[0004] In order to solve the above problems, in recent years, with the rapid development of machine vision technology, non-contact measurement method based on 2D image is gradually applied to the field of livestock weight estimation. Compared with contact type measurement, non-contact measurement has multiple advantages, including reducing the disturbance to animals, reducing stress reaction, meeting the requirements of animal welfare. At the same time, non-contact measurement technology can realize automatic operation, reduce manual intervention, and improve data acquisition efficiency and accuracy. In addition, non-contact measurement can estimate the weight of multiple animals in a short time, which greatly improves the production efficiency.
[0005] At present, a large number of weight estimation techniques based on machine vision have been applied to pigs, cattle, sheep and other large livestock. These techniques usually rely on 2D or 3D cameras to capture the side or top images of animals, use image processing algorithms to extract animal body features, and combine regression algorithms to estimate the weight, which has achieved good results in large livestock. However, for broilers, due to their small size and frequent activity, the existing technology is not ideal when applied to broiler weight estimation.
[0006] Therefore, it is urgent to develop a broiler weight estimation system based on two-view images, which is optimized for broiler characteristics, uses consumer-level camera equipment to obtain broiler images, and combines existing image processing and deep learning algorithms to achieve efficient and stable weight estimation. Utility model content
[0007] The utility model aims at providing a broiler weight estimation system based on two-view images, characterized in that it comprises a broiler image acquisition device and a weight estimation device.
[0008] The broiler image acquisition device comprises a cage, a top view camera, a rear view camera and an acquisition trigger device; one side of the cage is provided with a cage opening, and the broiler to be measured enters the inside of the cage through the cage opening; a partition is arranged in the inside of the cage, which divides the inside of the cage into a feed storage area and a broiler standing area; a feed container is arranged in the feed storage area; the broiler standing area can accommodate a single broiler to be measured to stand; a feeding hole is arranged on the partition; the feeding hole limits the broiler to be measured to only pass the head through the feeding hole to complete feeding; the top view camera is fixed to the top of the cage and is used for acquiring the top view image of the broiler to be measured when feeding; the rear view camera is fixed to the outside of the cage and is close to the side of the cage opening, and is used for acquiring the rear side view image of the broiler to be measured when feeding; the acquisition trigger device is used for triggering the acquisition of pictures and weight estimation;
[0009] The weight estimation device is used for estimating the weight value of the broiler to be measured by using two-view images.
[0010] The two-view images comprise the top view image and the rear side view image of the broiler to be measured when feeding.
[0011] The acquisition trigger device comprises a radio frequency identification tag worn on the broiler to be measured and a radio frequency identification sensor; when the broiler to be measured wearing the radio frequency identification tag enters the cage to feed, the radio frequency identification sensor triggers the acquisition of pictures and weight estimation.
[0012] The top of the cage is provided with a fill light.
[0013] The weight estimation device further comprises a result display device, which is used for real-time display of the weight estimation result.
[0014] The top view camera and the rear view camera are consumer-level infrared cameras.
[0015] The utility model has the advantages of:
[0016] The utility model discloses a kind of broiler weight estimation systems based on two view images, in multiplexing existing breeding equipment, 2 consumer-grade infrared cameras are installed in the side and above of breeding cage respectively, collect two view images under the two visual angles of natural feeding state of single broiler. It can collect the complete broiler body image of different ages in the whole growth cycle of broiler, better adapt to the characteristics of broiler active by comprehensive two visual angle picture information. Utilize the feeding instinct of broiler, after the characteristics that broiler enters shooting area and probably initiatively feeds, after 1-3 seconds of perception broiler enters collection area, collect multiple two view images, and automatically set camera into standby state, greatly reduce the power consumption of camera. BRIEF DESCRIPTION OF DRAWINGS
[0017] Figure 1 It is the structure overhead view schematic diagram of the utility model one kind broiler image collection device;
[0018] Figure 2 It is the structure rear view schematic diagram of the utility model one kind broiler image collection device;
[0019] Figure 3 It is the structure schematic diagram of the utility model one kind broiler weight estimation systems based on two view images;
[0020] Among them, broiler image collection device-100, weight estimation device-200, cage-101, baffle-102, feed storage area-103, broiler standing area-104, feed container-105, feeding hole-106, overhead view camera-107, post-view camera-108, cage opening-109. DETAILED DESCRIPTION
[0021] The utility model provides a kind of broiler weight estimation systems based on two view images, the following further detailed description of the utility model is combined with drawing.
[0022] As Figure 1 The utility model embodiment shown in the figure discloses a kind of broiler weight estimation systems based on two view images, including: broiler image collection device 100 and weight estimation device 200;
[0023] The broiler image acquisition device 100 includes: a cage 101, a top-view camera 107, a back-view camera 108, and an acquisition triggering device; the cage 101 has an opening 109 on one side, through which the broiler to be tested enters the cage 101; the cage 101 has a partition 102 inside, which divides the interior of the cage 101 into a feed storage area 103 and a broiler standing area 104; the feed storage area 103 has a feed container 105; the broiler standing area 104 can... The cage can accommodate a single broiler chicken standing upright; a feeding hole 106 is provided on the partition 102; the feeding hole 106 restricts the broiler chicken to only be able to pass its head through the feeding hole 106 to complete feeding; a top view camera 107 is fixed to the top of the cage 101 to collect top view images of the broiler chicken feeding; a rear view camera 107 is fixed to the outside of the cage 101, near the cage opening 109, to collect side rear view images of the broiler chicken feeding; a triggering device is used to trigger the acquisition of images and perform weight estimation.
[0024] The weight estimation device 200 is used to estimate the weight of the broiler chicken under test using two-view images;
[0025] The two-view images include: a top view image and a side-rear view image of the broiler chickens being fed.
[0026] The structure of the broiler chicken image acquisition device disclosed in this utility model is shown in top view as follows: Figure 1 As shown, the rear view diagram is as follows: Figure 2 As shown. In this embodiment, the broiler weight estimation system based on two-view images is applicable to scenarios involving weight estimation of the object to be estimated, such as weight estimation in broiler breeding. Based on extensive experimental observations by the inventors, in most cases, driven by their natural instincts, broilers will immediately rush towards food upon entering the cage, causing them to stretch their necks to pass their heads through the cage partitions, maintain an upright posture, and keep their wings tucked in. Therefore, in most cases, within 1 to 3 seconds of entering the cage, the broiler is likely in the aforementioned photographic condition. At this time, triggering the shooting opportunity by calling the camera to capture 3 to 5 sets of two-view images is sufficient. Those skilled in the art should know how to implement the triggering shooting action using relevant existing technologies, including but not limited to manual triggering using a switch and automatic triggering using a sensor; this embodiment does not specifically limit the methods used.
[0027] The conditions under which photography is permitted are:
[0028] The broiler chickens enter the cage, put their heads through the partition, and begin to eat; the broiler chickens are in an upright position; and the broiler chickens' wings are tucked in.
[0029] The sensors include, but are not limited to, infrared sensors, optical sensors, bioelectric sensors, and weighing scales, etc., and are not specifically limited in this embodiment.
[0030] In an optional embodiment, the top-view camera 107 and the rear-view camera 108 are configured to: be in standby mode by default; when the acquisition trigger device triggers the shooting action, they wait 1 second before taking pictures; after taking 3 to 5 sets of pictures at 1-second intervals, the top-view camera 107 and the rear-view camera 108 are automatically switched back to standby mode. In this optional embodiment, by taking advantage of the characteristic that broilers are likely to be in the shooting conditions for 1 to 3 seconds after entering the cage, and keeping the cameras in standby mode at other times, system power consumption can be effectively reduced.
[0031] Those skilled in the art should understand that, considering the significant weight fluctuations of broilers throughout their growth process, the camera's installation position must ensure that it can capture a complete image of the older broilers' bodies. Those skilled in the art can make corresponding adjustments based on the performance parameters of the selected camera, the size of the rearing cage, and other factors; specific limitations are not imposed in this embodiment.
[0032] In an optional embodiment, the acquisition triggering device includes: an RFID tag worn on the broiler chicken to be tested and an RFID sensor; when the broiler chicken wearing the RFID tag enters the cage 101 to eat, the RFID sensor triggers the acquisition of images and performs weight estimation.
[0033] In this optional embodiment, the radio frequency identification tag can contain the identification information of the broiler chicken to be tested, so as to realize long-term automatic weight data tracking for individual broilers.
[0034] In this optional embodiment, the broiler chickens are equipped with electronic tags, and the sensor is an electronic tag sensor. By combining the electronic tag sensor with the electronic tag, it is possible not only to detect when the broiler chicken enters a designated area within the cage, but also to obtain the chicken's identity information, facilitating weight tracking throughout the broiler's growth cycle. The electronic tag includes, but is not limited to, RFID tags, NFC tags, Bluetooth tags, and ZigBee tags, etc., and is not specifically limited in this embodiment.
[0035] In an optional embodiment, a supplementary light is provided on the top of the cage 101. In this optional embodiment, the supplementary light can improve the image quality of the acquired images, thereby improving the accuracy of weight assessment.
[0036] In an optional embodiment, the weight estimation device 200 further includes a result display device for displaying the weight estimation result in real time. In this optional embodiment, displaying the weight estimation result in real time facilitates manual recording of weight data.
[0037] In an optional embodiment, the top-view camera 107 and the rear-view camera 108 are consumer-grade infrared cameras. In this optional embodiment, the imaging device consists of two consumer-grade infrared cameras, which are respectively mounted to the side and above the weight estimation device to acquire two-view images of a single broiler chicken in its natural feeding state from both rear-view and top-view perspectives, forming a set of two-view images of the broiler chicken. Using two consumer-grade infrared cameras for image acquisition helps improve the quality of the acquired images under low-light conditions.
[0038] To verify the effectiveness of the broiler weight estimation system based on two-view images disclosed in this utility model, the following verification experiment was conducted. A total of 1122 broilers were collected, and the two-view images were analyzed. Figure 1 A total of 45,258 images were taken. Immediately after the chickens were photographed, staff weighed them and recorded the weight as the actual value.
[0039] The two-view images acquired by the broiler image acquisition device 100 are input into the weight estimation device 200 for weight estimation, obtaining the estimated weight of the broiler to be tested. The estimated value is then compared with the actual value. The weight estimation results and related data are shown in Table 3.
[0040] Table 3. Weight estimation results
[0041] Model MAE Params FLOPs CNN 17.180g 3.27M 5.15G
[0042] Table 3 defines MAE (Mean Absolute Error) as the average of the absolute differences between predicted and actual values. The formula is:
[0043]
[0044] Among them, y i This is the actual value. is the predicted value, and n is the number of samples.
[0045] Function: MAE is used to measure the average absolute difference between the model's predicted values and the actual values.
[0046] In this test, the MAE index was used to evaluate the estimation results. When the weight estimation device 200 used the CNN network commonly used in the prior art, the MAE obtained was 17.180g, which proves the effectiveness of the broiler weight estimation system based on two-view images disclosed in this utility model.
[0047] The specific process of inputting the two-view images acquired by the broiler image acquisition device 100 into the weight estimation device 200 for weight estimation is as follows:
[0048] In response to the start acquisition signal issued by the acquisition trigger device, a set of two-view images of the broiler chicken is acquired by the top-view camera 107 and the rear-view camera 108. The set of two-view images is then filtered using a commonly used image classification network to obtain images that meet the conditions for capture, such as the ShuffleNetV2 network. Semantic segmentation processing is then performed on the two-view images that meet the capture conditions using a commonly used image segmentation network, such as the SeaFormer network, to obtain segmented two-view images. Based on the segmented two-view images, a commonly used neural network is input to predict the weight of the broiler chicken, such as a CNN network.
[0049] The system structure of the broiler weight estimation system based on two-view images is as follows: Figure 3 As shown, during the test, two consumer-grade 2D infrared cameras were installed on the side, rear, and top of the broiler image acquisition device 100 to capture two-view images of broilers in a natural feeding state while standing on the broiler image acquisition device 100. The two-view images of the broilers were obtained by parsing the RTSP protocol stream using a Python program, resulting in at least one two-view image of the object to be estimated. This enables data interaction between the broiler image acquisition device 100 and the weight estimation device 200.
[0050] The 2D infrared camera follows the RTSP protocol. The specific coordinates of the camera are as follows: when viewed from above, the camera is 8.7 cm from the left pivot of the iron plate; when viewed from behind, the iron plate is 18.8 cm from the top of the iron door; and the camera is 15.3 cm from the edge of the iron plate.
[0051] In this embodiment, those skilled in the art should understand that, considering the significant weight fluctuations of broilers throughout their growth process, the camera's installation position must ensure that it can capture a complete image of the older broilers' bodies. Those skilled in the art can make corresponding adjustments based on the performance parameters of the selected camera, the size of the rearing cage, and other factors; however, no specific limitations are imposed in this embodiment.
[0052] Compared to existing technologies that typically rely on 2D or 3D cameras to capture side or top images of animals, this invention uses consumer-grade camera equipment to acquire post-view and top-view images. It fully considers the characteristics of broilers compared to pigs, cattle, and sheep, such as increased activity, significant body size variations during the rearing cycle, and the substantial impact of standing posture and wing position on weight prediction. By using images acquired from two perspectives, a more accurate prediction result is obtained. The weight estimation results and related data are shown in Table 3. For example, the MAE is 17.180g, demonstrating that the predicted value obtained by the broiler weight estimation system based on two-view images disclosed in this invention has a high degree of fit with the actual value and can achieve a usable prediction effect.
Claims
1. A broiler weight estimation system based on two-view images, characterized in that, include: Broiler image acquisition device (100) and weight estimation device (200); The broiler image acquisition device (100) includes: a cage (101), a top-view camera (107), a back-view camera (108), and an acquisition triggering device; the cage (101) has an opening (109) on one side, and the broiler to be tested enters the cage (101) through the opening (109); the cage (101) is equipped with a partition (102), which divides the interior of the cage (101) into a feed storage area (103) and a broiler standing area (104); the feed storage area (103) is equipped with a feed container (105); the broiler standing area is equipped with a feed container (105). The zone (104) can accommodate a single broiler chicken standing upright; a feeding hole (106) is provided on the partition (102); the feeding hole (106) restricts the broiler chicken to only be able to put its head through the feeding hole (106) to complete feeding; a top view camera (107) is fixed to the top of the cage (101) to collect top view images of the broiler chicken feeding; a rear view camera (108) is fixed to the outside of the cage (101), near the cage opening (109), to collect side rear view images of the broiler chicken feeding; a triggering device is used to trigger the acquisition of images and perform weight estimation; The weight estimation device (200) is used to estimate the weight of the broiler chicken under test using two-view images; The two-view images include: a top view image and a side-rear view image of the broiler chickens being fed.
2. The broiler weight estimation system based on two-view images according to claim 1, characterized in that, The acquisition triggering device includes: an RFID tag worn on the broiler chicken to be tested and an RFID sensor; when the broiler chicken wearing the RFID tag enters the cage (101) to eat, the RFID sensor triggers the acquisition of images and performs weight estimation.
3. The broiler weight estimation system based on two-view images according to claim 1, characterized in that, The cage (101) is equipped with a supplementary light on its top.
4. The broiler weight estimation system based on two-view images according to claim 1, characterized in that, The weight estimation device (200) also includes a result display device for displaying the weight estimation results in real time.
5. The broiler weight estimation system based on two-view images according to claim 1, characterized in that, The top-view camera (107) and the rear-view camera (108) are consumer-grade infrared cameras.