An environmental control method based on eggplant visual monitoring

Abstract

This invention relates to the field of image recognition technology, and more particularly to an environmental control method based on visual monitoring of watermarked eggs. The method includes the following steps: acquiring translucent images at various times and identifying the watermark region; acquiring the total projected area of ​​the egg and calculating the proportion of the watermark area at each time; constructing a neighborhood window to determine the spatial gradient modulus and calculating the normalized permeability of the watermark region at the current time; acquiring the edge of the watermark region and dividing it into several arc segments, determining the non-uniformity coefficient at each time; predicting the saturation risk after a set time period, and generating an environmental regulation increment based on the saturation risk and the time change rate of the normalized permeability for environmental regulation control. This invention characterizes the eggshell moisture permeation process through multidimensional visual indicators and establishes a mapping relationship from biological phenotype evolution to environmental regulation by combining a prediction model, effectively reducing the lag probability of environmental regulation and improving the accuracy of poultry house microclimate regulation.

Description

Technical Field

[0001] This invention relates to the field of image recognition technology, and in particular to an environmental control method based on visual monitoring of watermarked eggs. Background Technology

[0002] In large-scale intensive egg-laying hen farming, maintaining a stable balance of environmental parameters within the poultry house is physically significant for ensuring egg quality and improving egg production performance. Watermarked eggs, a biological phenotype characterized by semi-transparent spots on the eggshell surface, are strongly modulated by physical environmental factors such as fluctuations in humidity and cold air stress. To achieve dynamic closed-loop regulation of the poultry house microclimate, a monitoring method based on the visual characteristics of watermarked eggs is introduced. This method can provide real-time feedback on the stress state of individual organisms, thereby translating visual perception results into environmental regulation commands to address the risk of egg quality degradation due to inadequate environmental control.

[0003] Existing technologies mostly use image sensors to acquire translucent images of eggs and employ conventional segmentation algorithms to identify watermark regions. By obtaining the proportion of the watermark area in a single frame image and comparing it with a preset threshold, the adjustment action of mechanisms such as fans or water curtains is triggered, thus achieving passive intervention.

[0004] However, considering that the formation and expansion of watermarks is a dynamic physical process continuously modulated by the external environment, its phenotypic evolution exhibits significant non-stationarity and spatial unevenness. Existing control logic typically focuses on extracting the static geometric features of watermarks, making it difficult to accurately assess the drastic evolution of watermarks over time or to finely identify the differentiated expansion trends of watermarks in space. This results in the generated regulatory commands often failing to accurately adapt to the actual physical evolution requirements when faced with instantaneous changes in environmental parameters or local environmental anomalies. Due to the lack of in-depth assessment of the evolutionary inertia of watermarks, the intervention actions generated by environmental control agencies often exhibit a certain lag, increasing the probability of over- or under-regulation of environmental parameters. Furthermore, the stability of regulatory accuracy needs further improvement when dealing with local microclimate differences in complex poultry house environments. Summary of the Invention

[0005] To address the technical problem that existing environmental monitoring logic struggles to effectively calculate the evolutionary inertia of watermark phenotypes in the temporal domain and the heterogeneity of their spatial distribution, leading to delayed environmental control commands and a mismatch with actual physical evolution trends, this invention provides an environmental control method based on visual monitoring of watermarked eggs. This method includes the following steps:

[0006] The translucent images of the eggshell surface at preset sampling intervals are acquired, and the watermark region in the translucent images at each time moment is identified. Taking any given time moment as the current time, the total projected area of ​​the egg is obtained, and the watermark area ratio is calculated based on the ratio of the watermark area area to the total projected area of ​​the egg. The spatial gradient magnitude of each pixel within the watermark region is obtained, and the normalized permeability of the watermark region at the current time moment is determined based on the grayscale deviation and spatial gradient magnitude of each pixel at the current and previous times. The edge of the watermark region is obtained, divided into several arc segments, and the normalized permeability of each arc segment is calculated. The difference between the normalized permeability of each arc segment and the normalized permeability of the watermark region is statistically analyzed to determine the non-uniformity coefficient at each time moment. Based on the watermark area ratio, the normalized permeability of the watermark region, and the non-uniformity coefficient, the saturation risk after a set time period is predicted. Based on the saturation risk and the time change rate of the normalized permeability of the watermark region, an environmental regulation increment is generated. Environmental regulation control is performed using the saturation risk and the environmental regulation increment.

[0007] This invention comprehensively utilizes the static distribution characteristics and dynamic evolution patterns of watermarks on eggshell surfaces. First, it establishes a temporal initial state benchmark using the watermark area ratio and introduces spatial gradient modulus to weight and modulate grayscale fluctuations to determine the normalized permeability. Then, it assesses the spatial heterogeneity of moisture diffusion by dividing arc segments to determine the non-uniformity coefficient. This achieves a multidimensional characterization of eggshell moisture infiltration intensity in both spatiotemporal dimensions, accurately capturing the phenotypic driving effect of the poultry house environment. Based on this, it combines a nonlinear mapping function to predict the saturation risk after a set time period and uses the time change rate of normalized permeability to characterize the urgency of environmental deterioration to determine the environmental regulation increment. This achieves deep coupling between visual perception results and environmental regulation commands, reduces the lag probability of poultry house environmental regulation relative to watermark evolution, and improves the feedback accuracy to biological stress caused by local microclimate anomalies.

[0008] Preferably, the normalized permeability of the watermark region satisfies the following relationship:

[0009] ;

[0010] in, yes Normalized permeability of the watermark area; It is the total projected area of ​​the egg; yes The total number of pixels within the watermark area; yes The first time within the watermark area The grayscale value of each pixel; The first one is the one mentioned Pixels in The grayscale value at any given moment; It is the sampling time interval; yes Time of the first Spatial gradient magnitude of each pixel; It is the preset first minute value; It is the first Area contribution factor of each pixel; It is the absolute value symbol.

[0011] This invention introduces an area contribution factor into the normalized permeability calculation formula for the watermark region. This maps the temporal grayscale jump characteristics at the pixel level to the linear expansion displacement determined by the spatial gradient magnitude, effectively transforming microscopic pixel displacement into a macroscopic area expansion ratio. Because this scheme uses the total area of ​​the egg projection for normalization, the final determined normalized permeability of the watermark region is reduced to the area change rate per unit time. This achieves dimensional alignment between the water permeation rate and the watermark distribution ratio in physical logic, effectively eliminating the numerical calculation bias risk caused by dimensional inconsistencies in subsequent saturation risk prediction models.

[0012] Preferably, obtaining the total projected area of ​​the egg includes: using an edge extraction algorithm to obtain the outline boundary of the egg in the translucent image; traversing the closed region enclosed by the outline boundary, counting the total number of pixels contained in the closed region, and recording it as the total projected area of ​​the egg.

[0013] Preferably, the step of calculating the watermark area ratio based on the ratio of the area of ​​the watermark region to the total projected area of ​​the egg includes: counting the total number of pixels contained in the watermark region; and recording the ratio of the total number of pixels to the total projected area of ​​the egg as the watermark area ratio.

[0014] Preferably, the non-uniformity coefficient satisfies the following relationship:

[0015] ;

[0016] in, yes Non-uniformity coefficient at time; It is the total number of arc segments; yes Time of the first Normalized penetration rate of each arc segment; yes Normalized permeability of the watermark area; It is the preset second minute value; It is the absolute value symbol.

[0017] This invention calculates the deviation intensity of the normalized permeability of each arc segment relative to the normalized permeability of the watermark region, transforming the asymmetric expansion characteristics of physical space into a fluctuation evaluation index. This decouples the non-uniform evaluation logic from the absolute geometric dimensions of the watermark, and helps to identify phenotypic abrupt changes caused by local duct directional stress or airflow stagnation areas.

[0018] Preferably, obtaining the edge of the watermark region and dividing it into several arc segments includes: extracting the edge of the watermark region using an edge detection algorithm; determining the geometric center point of the edge of the watermark region; using the geometric center point as the origin of polar coordinates, dividing the edge of the watermark region into several arc segments according to a preset angle bisector step size.

[0019] Preferably, the saturation risk satisfies the following relationship:

[0020] ;

[0021] in, yes The risk of saturation at any given moment; yes The percentage of the watermark area at any given moment; It is the current moment; It sets the duration; yes Normalized permeability of the watermark area; yes Non-uniformity coefficient at time; It is the environmental buffer coefficient; It is an integral variable.

[0022] This invention introduces a cumulative term with time-domain integral characteristics into the saturated risk prediction model and utilizes the consistency between the normalized permeability of the watermark region and the proportion of the watermark area in terms of dimensions. This achieves the linear superposition of the future expansion kinetic energy of the watermark and the current static damage level within the same numerical space. Considering the modulating effect of the environmental buffer coefficient on the adaptive convergence characteristics of biological phenotypes, this invention transforms the time-dimensional permeation accumulation into a bounded probability distribution through a nonlinear mapping function. This serves to predict the risk based on the watermark diffusion trend before the physical critical point is reached, reducing the probability of distortion in judgment conclusions caused by conflicts in physical dimensions in the prediction logic.

[0023] Preferably, generating the environmental regulation increment based on the saturation risk and the time change rate of the normalized permeability of the watermark region includes: calculating the absolute value of the difference between the normalized permeability of the watermark region at the current time and the previous sampling time, and recording the ratio of the absolute value of the difference to the sampling time interval as the time change rate; multiplying the product of the saturation risk and the time change rate by a preset gain coefficient to obtain the environmental regulation increment.

[0024] This invention characterizes the acceleration of watermark expansion in real time by calculating the time-varying rate of normalized permeability in the watermark region. This rate is then coupled with the predicted saturation risk through a product, enabling adaptive adjustment of intervention intensity based on the urgency of environmental degradation. Under this feedback logic, when the watermark is in its rapid, rapid expansion phase, the adjustment increment increases in the same direction as the acceleration term, achieving strong negative feedback intervention. Conversely, when the watermark evolution enters a steady-state phase, the adjustment intensity decreases to maintain the lifespan of the actuators. This allows for flexible allocation of regulatory resources over time, reducing the risk of insufficient or excessive regulation of the poultry house microclimate.

[0025] Preferably, the environmental regulation control using saturation risk and environmental regulation increment includes: when the saturation risk falls into a preset control threshold range, adjusting environmental parameters using the environmental regulation increment.

[0026] Preferably, the identification of the watermark region in the translucent image at each time step includes: obtaining the grayscale value of a preset background reference; traversing each pixel in the translucent image, calculating the difference between the grayscale value of the corresponding pixel and the grayscale value of the preset background reference, and marking the set of pixels whose difference exceeds a preset deviation threshold as candidate regions; obtaining the gradient distribution sequence of the edge of the candidate region along the normal direction, calculating the spatial span covered by the non-zero gradient values ​​in the gradient distribution sequence, and recording it as the gradient support width; in response to the gradient support width falling into a preset diffusion width range, and the gradient distribution sequence showing a single-peak smooth transition trend, confirming the candidate region as a watermark region.

[0027] The beneficial effects of this invention are as follows: First, by introducing a preset background benchmark and diffusion distribution discrimination mechanism, this invention utilizes the physical permeation law of water in porous media to accurately identify the watermark area, reducing the interference of eggshell surface texture or sensor noise on the identification results. By establishing a feature evaluation system based on area normalization and spatial gradient modulation, this invention achieves a globally unified characterization of the permeation rate of different individual egg samples, eliminating individual sample differences and suppressing computational divergence. Furthermore, by analyzing the discrete intensity of arc segments and combining it with a time-domain integral prediction model, this invention achieves advanced assessment of future saturation risks and real-time determination of the urgency of environmental degradation, enhancing the accuracy of adapting environmental control commands to actual physical evolution trends. This invention maps complex image features to demarcated risk probabilities and adjustment increments, achieving deep collaboration between biological phenotypic monitoring and environmental actuator actions through a threshold triggering mechanism. This effectively reduces the risk of egg quality degradation due to environmental fluctuations or delayed regulation, improving the intelligence level and operational stability of the poultry house environmental monitoring system. Attached Figure Description

[0028] Figure 1 A flowchart of an environmental control method based on visual monitoring of watermarked eggs provided in an embodiment of the present invention;

[0029] Figure 2 This is a schematic diagram of permeability and non-uniformity coefficient curves provided in an embodiment of the present invention;

[0030] Figure 3 This is a schematic diagram comparing the risk warning effects provided in an embodiment of the present invention. Detailed Implementation

[0031] This invention provides an environmental control method based on visual monitoring of watermarked eggs, such as... Figure 1 As shown, the method includes steps S100-S400:

[0032] Step S100: Obtain the light transmission image of the eggshell surface at each time interval according to the preset sampling interval, and identify the watermark area in the light transmission image at each time interval.

[0033] It should be noted that environmental control based on visual monitoring of watermarked eggs relies on the accurate perception of the physical process of moisture penetration into the eggshell surface. The watermark, visually represented, exhibits localized grayscale heterogeneity in the image due to abrupt changes in transmittance. To establish a physical anchor for subsequent evaluation logic, visual information acquisition is required first. Determining a preset background baseline provides a static reference point for transmittance intensity in watermark recognition. Considering the physical differences in eggshell thickness and ambient light intensity among different batches of eggs, this baseline helps eliminate interference from environmental background noise and normal eggshell texture, ensuring a unified comparison benchmark for subsequent calculations of watermark deviation intensity. Furthermore, by identifying the diffusion distribution characteristics of edge gradients, the watermark area can be effectively distinguished from stains or sensor noise on the eggshell surface. This is because moisture penetration in the porous eggshell medium follows physical diffusion laws, and its edges exhibit a continuous, gradual grayscale transition characteristic, rather than a step-like characteristic.

[0034] Specifically, firstly, industrial cameras are deployed inside the poultry house. These cameras can be positioned above the egg collection line or the light transmission monitoring station, and will be used according to a preset sampling cycle. Continuously acquire translucent images of the eggshell surface. As a preferred embodiment, the sampling period... It is recommended to set it to 60 to 300 seconds to match the physical response frequency of the poultry house environmental control system.

[0035] Then, each pixel in the translucent image is traversed, and the difference between the gray value of the corresponding pixel and the preset background benchmark is calculated. Pixels with differences exceeding a preset deviation threshold are marked as candidate regions. The acquisition of the preset background benchmark includes: acquiring several sets of known dry and watermark-free translucent images of healthy eggs, extracting the pixel gray values ​​in a preset central area on the eggshell surface and calculating the average value, and defining the average value as the preset background benchmark.

[0036] Finally, the watermark area is obtained. Considering that the edges of scratches are usually steep and their gradient support width is narrow, while water penetration is diffuse and its edges are gradual and its gradient support width is wide, in order to distinguish between real water penetration and rigid scratches or stains on the eggshell surface, this embodiment adopts the edge gradient analysis method.

[0037] Specifically, the grayscale gradients of the edge pixels of the candidate region along the normal direction are obtained, and a gradient distribution sequence is constructed. The number of pixels or physical distance covered by the non-zero gradient values ​​in the gradient distribution sequence is calculated and defined as the gradient support width. If the gradient support width falls within a preset diffusion width range, and the gradient distribution sequence exhibits a single-peak smooth transition trend, then the region is determined to satisfy the diffusion distribution law of water infiltration characteristics and is recorded as a watermark region.

[0038] It should be further noted that the values ​​of the preset center area size, preset deviation threshold, and diffusion width range need to be set according to the actual visual environment and the physical characteristics of the eggshell: For scenes with high background noise or rough eggshell texture, the preset deviation threshold can be appropriately increased, such as set to 15 to 20 gray levels, to enhance the ability to suppress artifact noise; for scenes with uniform lighting and a smooth eggshell surface, it can be reduced to 5 to 10 gray levels to improve the sensitivity of capturing initial weak penetration features. The diffusion width range should be adjusted in conjunction with the eggshell porosity and imaging magnification: for eggshells with large pores and strong penetration diffusion, the diffusion width range can be set to 5 to 10 pixels; for eggshells with high density and limited diffusion, this range can be reduced to 2 to 5 pixels to prevent edge features from being mistakenly eliminated by the smoothing algorithm. The size of the preset center region needs to balance statistical representativeness and computational efficiency: for varieties with poor eggshell thickness uniformity, it can be expanded to 150 pixels by 150 pixels to obtain a more robust background average; for varieties with high thickness uniformity, it can be reduced to 80 pixels by 80 pixels. In this embodiment, the preset center region is preferably 100 pixels by 100 pixels, the preset deviation threshold is preferably 10, and the diffusion width range is preferably 3 to 8 pixels.

[0039] At this point, the translucent images and watermark areas corresponding to each moment have been obtained.

[0040] Step S200: Taking any time as the current time, obtain the total projected area of ​​the egg, and calculate the watermark area ratio based on the ratio of the area of ​​the watermark region to the total projected area of ​​the egg; obtain the spatial gradient magnitude of each pixel in the watermark region, and determine the normalized permeability of the watermark region at the current time based on the grayscale deviation and spatial gradient magnitude of each pixel at the current time and the previous time.

[0041] It should be noted that individual eggs differ in physical geometry. Directly using absolute pixel displacement to characterize water penetration intensity can introduce systematic biases due to variations in imaging magnification or sample size. To eliminate interference from individual morphology, it is necessary to obtain the total projected area of ​​each sample to establish a global spatial normalization benchmark, thereby ensuring that subsequently acquired features accurately reflect the nature of environmental stress.

[0042] First, the proportion of the watermark area at each time point is obtained. It should be noted that the evolution of the watermark on the eggshell surface is a dynamic physical process of continuous water infiltration along the micropores of the eggshell. Calculating the ratio of the watermark area to the total projected area of ​​the eggshell serves to establish a time-domain initial state baseline, characterizing the cumulative damage caused by water infiltration at the current time point, and providing necessary static feature input for subsequently capturing the dynamic evolution rate within adjacent sampling intervals.

[0043] Specifically, for each translucent image, an edge extraction algorithm is used to identify and lock the outer contour boundary of the individual egg in the translucent image. By traversing the closed region enclosed by the contour boundary, the total number of pixels contained therein is counted and recorded as the total projected area of ​​the egg. The pixels contained in the watermark region are traversed and the total number of pixels is counted. The total number of pixels is compared with the previously calculated total projected area of ​​the egg to obtain the watermark area ratio at the corresponding time.

[0044] Then, for each pixel within the watermark area, its neighborhood window is constructed, and the spatial gradient magnitude of each pixel is determined. It should be noted that the spatial gradient magnitude characterizes the degree of drastic evolution of local grayscale in the spatial dimension, serving to distinguish the diffusion edges caused by water infiltration from the inherent texture of the eggshell background. By introducing this metric, a spatial morphological reference benchmark can be provided for grayscale temporal fluctuations, thereby effectively suppressing numerical fluctuations caused by background noise and improving the robustness of penetration determination.

[0045] Specifically, a neighborhood window is constructed centered on each pixel within the watermark area. The gradient operator is used to extract features from the pixel distribution within the neighborhood window to determine the spatial gradient magnitude of the current pixel.

[0046] It should be noted that the neighborhood window size needs to be set according to the actual working conditions: for scenes with large porosity features on the eggshell surface and wide water penetration diffusion bands, the window can be appropriately increased to 20 to 30 pixels to fully encompass the local diffusion gradient features and ensure the accuracy of background statistics when calculating local contrast; for low-magnification imaging or scenes with extremely fine eggshell textures and narrow penetration edges, the window can be appropriately reduced to 9 to 11 pixels to avoid introducing too much interference from irrelevant textures due to an excessively large window. In this embodiment, the preferred neighborhood window size is [not specified]. Pixel.

[0047] Finally, the normalized permeability of the watermark region at each time step is obtained. It should be noted that the dynamic essence of watermark expansion is the temporal evolution of the transmittance of neighboring pixels, and its evolution rate directly reflects the driving effect of environmental factors on the intensity of water infiltration. To calculate this dynamic feature, it is necessary to eliminate computational biases caused by sensor noise or drastic changes in local lighting. By introducing spatial gradient distribution features to adaptively weight the grayscale fluctuations of the pixel array, numerical divergence is suppressed and the robustness of feature description is enhanced, thereby achieving a mapping from static image space to dynamic physical evolution space.

[0048] Based on the above logic, the normalized permeability of the watermark region satisfies the following relationship:

[0049] ;

[0050] in, yes Normalized permeability of the watermark area; It is the total projected area of ​​the egg; yes The total number of pixels within the watermark area; yes The first time within the watermark area The grayscale value of each pixel; The first one is the one mentioned Pixels in The grayscale value at any given moment; It is the sampling time interval; yes Time of the first Spatial gradient magnitude of each pixel; It is a preset first tiny value used to prevent the denominator from being zero, and is preferably set to 0.001; It is the first The area contribution factor of each pixel is used to map the pixel-level line expansion intensity to the global area expansion ratio. It is the absolute value symbol.

[0051] In this relation, It captures the jump in transmittance caused by water infiltration, reflecting the dynamic intensity of penetration. This is used to characterize the intensity of the reference grayscale change required for a unit displacement at a specific pixel in a transparent image; by introducing a spatial gradient magnitude. This effectively suppresses the risk of numerical divergence caused by missing local gradients in the image, ensuring the stability of the normalized permeability calculation for the watermark region under different illumination intensities. and The product processing transforms complex pixel grayscale fluctuations into area expansion ratios with clear physical meaning, achieving alignment between visual features and dynamic indicators.

[0052] It should be noted that the area contribution factor The value of the area contribution factor needs to be dynamically adjusted based on the geometric center of the egg: for areas where the watermark edge is located at the large or small end of the egg where projection distortion is more obvious, the corresponding area contribution factor should be increased through coordinate transformation. This is to compensate for the underestimation of area contribution caused by viewpoint compression. In this embodiment, it is preferable to generate the corresponding weight value linearly based on the radial distance of the pixel from the center point of the egg.

[0053] At this point, the normalized permeability of the watermark area at each time point was obtained.

[0054] Step S300: Obtain the edge of the watermark area, divide it into several arc segments, calculate the normalized permeability of each arc segment, statistically analyze the difference between the normalized permeability of each arc segment and the normalized permeability of the watermark area, and determine the non-uniformity coefficient at each time point.

[0055] It should be noted that the diffusion of moisture in the micropores of the eggshell is affected by the local microclimate. When there is directional cold wind stress or local high humidity dead zones, the watermark boundary will exhibit an asymmetrical and abnormal advance. By introducing a coefficient of variation structure, the interference of the average rate base on the evaluation of spatial distribution can be eliminated. Thus, at any stage of watermark formation, the non-uniformity of its spatial expansion can be measured with a uniform standard, providing a physical basis for identifying local environmental anomalies.

[0056] First, the normalized permeability of each arc segment is obtained. It should be noted that, in order to capture the local evolution details of the permeation process in the spatial dimension, the continuous watermark boundaries need to be discretized. By extracting the local expansion features at different spatial locations, the direction of environmental stressors can be accurately located, thus providing refined input features for subsequent targeted environmental regulation.

[0057] Specifically, an edge detection algorithm is used to extract the closed edges of the watermark region, and the geometric center point of the closed edges is determined. Using the geometric center point as the origin, a preset angle bisector step size is calculated. Divide the polar coordinate space into A fan-shaped region divides the edge into There are corresponding arc segments, among which For 360 degrees and The ratio. Referring to the logic for determining the normalized permeability, calculations are performed on the boundary pixel matrix corresponding to each arc segment to determine the corresponding time point. Normalized penetration rate of each arc segment.

[0058] It should be noted that the angle bisector step length The value needs to be set according to the actual working conditions: for scenarios where the expected watermark shape is more complex or there is local airflow interference, the angle bisector step size can be reduced. For example, setting it to 10 to 20 degrees would increase the total number of arc segments. This enhances the accuracy of capturing long-range, continuous protrusion structures; for scenes with relatively regular watermark shapes or low imaging resolution, the angle bisector step size can be appropriately increased. For example, set the angle to 45 degrees to 90 degrees and reduce the total number of arc segments. This improves computational efficiency and filters out minor detail noise. In this embodiment, the angle bisector step size... The preferred setting is 30 degrees, corresponding to the total number of arc segments. The optimal value is set to 12.

[0059] Then, the non-uniformity coefficient at each time point is calculated. It should be noted that, to achieve comparability of the degree of abnormal expansion under different infiltration stages, a deviation evaluation mechanism independent of absolute rate needs to be established. This invention maps the asymmetric expansion characteristics of physical space into a dimensionless fluctuation evaluation index by calculating the deviation of the normalized permeability of each arc segment relative to the normalized permeability of the watermark region, thereby enhancing the accuracy of capturing sensitivity to environmental fluctuations.

[0060] Based on the above logic, the non-uniform coefficients satisfy the following relationship:

[0061] ;

[0062] in, yes Non-uniformity coefficient at time; It is the total number of arc segments; yes Time of the first Normalized penetration rate of each arc segment; yes Normalized permeability of the watermark area; It is a preset second tiny value used to prevent the denominator from being zero, and is preferably set to 0.0001; It is the absolute value symbol.

[0063] In this relation, Characterized the first The normalized permeability of each arc segment relative to the normalized permeability of the watermark area. The absolute value of the deviation represents the spatial discreteness intensity of the watermark boundary expansion. The denominator term introduces the normalized permeability of the watermark region. A normalization process was established to transform the indicator into a relative volatility, thus decoupling the non-uniform evaluation from the absolute size of the watermark. When the watermark expands synchronously in all radial directions, the deviation term approaches zero, and the non-uniformity coefficient remains low. If abnormal acceleration occurs in a local area due to environmental blind spots, the numerator term will rise sharply due to the expansion of the local rate deviation, thus serving as a physical early warning. Through this dimensionless mapping, the divergent numerical space is constrained to a unified proportional space, ensuring accurate identification of watermark cases of different severity.

[0064] Thus, the non-uniformity coefficients at each time point were obtained.

[0065] Step S400: Based on the watermark area ratio, the normalized permeability of the watermark region, and the non-uniformity coefficient, predict the saturation risk after a set time; generate an environmental regulation increment based on the saturation risk and the time change rate of the normalized permeability of the watermark region; and use the saturation risk and the environmental regulation increment for environmental regulation control.

[0066] It should be noted that stress regulation in poultry farming environments exhibits both physical inertia and biological lag. Frequent start-up and shutdown of actuators such as fans and evaporative cooling pads not only waste energy but may also trigger secondary stress in poultry due to drastic fluctuations in wind speed or humidity. By mapping the temporal trend of watermark evolution to a saturation risk probability, it is possible to predict the physical critical point before it is reached, thus providing the necessary buffer time for environmental regulation. This ensures that the control logic can both capture instantaneous deterioration trends and maintain stable output from the actuators.

[0067] First, based on the watermark area ratio, the normalized permeability of the watermark region, and the non-uniformity coefficient, the saturation risk after a set time period is predicted. It should be noted that the formation of watermarked eggs is a process of continuous water infiltration and accumulation within the porous eggshell, and its transition from a small area to full-area saturation exhibits a non-linear accelerating characteristic. To prevent regulatory lag during the watermark outbreak period, it is necessary to extrapolate the future saturation risk by combining the aforementioned determined watermark area ratio, normalized permeability of the watermark region, and non-uniformity coefficient. By introducing an exponential decay term to simulate the adaptive convergence characteristics of organisms in response to stress environments, the probability of quality damage in future time periods can be accurately characterized, thereby transforming divergent visual evolution observations into a basis for risk assessment.

[0068] Based on the above logic, the saturation risk satisfies the following relationship:

[0069] ;

[0070] in, yes The saturation risk at any given moment ranges from 0 to 1. yes The percentage of the watermark area at any given moment; It is the current moment; The duration is set, preferably 30 minutes, but the implementers can adjust it according to their needs; yes Time of the first Normalized penetration rate of each arc segment; yes Non-uniformity coefficient at time; It is the environmental buffer coefficient; It is an integral variable.

[0071] In this relation, Used to cumulatively assess the potential watermark diffusion kinetic energy on a time-domain scale; among which, The term amplifies the risk of spatial heterogeneity. When the watermark exhibits an asymmetrical and abnormal surge, the value of this term increases and drives up the integral result, indicating a higher risk of local saturation. This term is used to characterize the hysteretic convergence properties of biological phenotypes in response to environmental stress. By introducing this attenuation term, the prediction logic can adaptively adjust the weight of attention to long-term cumulative risk.

[0072] It should be noted that the environmental buffer coefficient The value needs to be set according to the actual breeding cycle and equipment control characteristics: For poultry groups in the brooding period or with fragile eggshells, the organism's ability to cope with stress is weak, and the environmental buffer coefficient is low. The environmental buffer coefficient can be appropriately lowered, such as to 0.05 to 0.1, to enhance sensitivity to the cumulative risk of long-term stress; for adult poultry or breeds with high eggshell density, it can be appropriately increased, such as to 0.2 to 0.3. In this embodiment, the environmental buffer coefficient... The optimal value is set to 0.15.

[0073] Then, based on the saturation risk and the time-varying rate of change of the normalized permeability of the watermark area, an environmental regulation increment is generated. It should be noted that the effectiveness of poultry house environmental control depends not only on the magnitude of the deviation but also on the urgency of the worsening stress intensity. In scenarios with the same watermark area proportion, a rapidly expanding watermark represents a more severe risk of environmental runaway. By coupling the saturation risk with the rate of change of permeability through a product, the regulatory intensity can be adaptively allocated according to the acceleration of stress evolution, achieving strong intervention of regulatory resources during the outbreak phase and minimal maintenance during the stable phase.

[0074] Based on the above logic, The environmental adjustment increment at any given time satisfies the following relationship:

[0075] ;

[0076] in, yes Incremental environmental regulation over time; It is the preset gain coefficient; yes The risk of saturation at any given moment; , They are , Normalized permeability of the watermark area at any given time; It is the sampling period; It is the absolute value symbol.

[0077] In this relation, The rate of change of the normalized permeability of the watermarked area represents the acceleration of watermark expansion. A large value indicates that the infiltration process is in a rapid ramp-up phase, which, when compared with the saturation risk... The product effect can enhance the adjustment increment. The mapping strength ensures that the identified environmental stress conclusions have sufficient response weights.

[0078] It should be noted that the preset gain coefficient The value needs to be set according to the actual breeding cycle and equipment control characteristics: for breeding scenarios equipped with high-power, non-frequency variable speed control fans, the preset gain coefficient should be set accordingly. The gain coefficient should be set at a low level, such as 2 to 4, to prevent drastic environmental fluctuations caused by excessive adjustment. For scenarios equipped with precision variable frequency fans or multi-stage evaporative cooling pads, the gain coefficient can be appropriately increased, such as 8 to 15, to quickly offset environmental degradation using the equipment's adjustment precision. In this embodiment, the preset gain coefficient... The preferred value is 5.

[0079] Finally, environmental regulation control is implemented using saturation risk and environmental regulation increments. It should be noted that, in order to filter out subtle fluctuations at the visual perception level and ensure the lifespan of the environmental control unit, a robust threshold triggering mechanism needs to be established. By setting a control threshold range, low-risk background noise can be shielded, ensuring that environmental intervention is activated only when the saturation risk truly threatens egg quality, thus achieving deep coordination between biological perception logic and actuator actions.

[0080] Specifically, the saturation risk is compared with a preset control threshold range. If the saturation risk falls within the control threshold range, the environmental adjustment increment is output. The system then moves to the environmental control unit, driving the actuators to complete environmental optimization.

[0081] It should be further noted that the value of the control threshold range needs to be set according to the stress sensitivity and environmental control precision of the actual breeding scenario: for scenarios involving the raising of high-yield hatching eggs or eggs with fragile shells, the trigger lower limit of the control threshold range can be lowered, such as to 0.45, to enhance the sensitivity to early, subtle penetration risks; for scenarios where poultry have strong environmental adaptability or where natural temperature fluctuations are large due to geographical location, the trigger lower limit can be appropriately increased, such as to 0.75, to shield redundant equipment actions caused by minor fluctuations in the visual background and improve the robustness of the system. In this embodiment, the control threshold range is preferably set to 0.6 to 1.

[0082] like Figure 2 The figure shows a schematic diagram of permeability and non-uniformity coefficient curves. The horizontal axis represents time, the left vertical axis represents the permeability value, and the right vertical axis represents the non-uniformity coefficient value. The solid line corresponds to the normalized permeability of the watermark area, reflecting the jump in transmittance and the dynamic intensity of permeation caused by water infiltration. The dashed line corresponds to the non-uniformity coefficient, representing the spatial dispersion intensity of the watermark boundary expansion. As can be observed from the figure, the solid line exhibits a clear S-shaped growth trend, accurately capturing the physical process of water permeation from initial induction, mid-term acceleration, to late-term saturation and slowing down. During the critical period of most intense permeation, such as around 500 on the time axis, the dashed line shows significant fluctuations and forms a peak. This is consistent with the characteristic described in this invention of a localized area experiencing abnormal surges due to eggshell pores, leading to a sharp increase in the non-uniformity coefficient, demonstrating the invention's extremely high accuracy in identifying the spatial heterogeneity of water permeation on the eggshell surface.

[0083] like Figure 3 The diagram shows a comparison of risk warning effects. The horizontal axis represents time, and the vertical axis represents the indicator value. The solid line corresponds to the saturation risk probability of this invention, the dashed line corresponds to the saturation risk comparison value of the prior art, the dashed line corresponds to the environmental adjustment increment generated by this invention, and the shaded area corresponds to the preset control threshold range. Because this invention introduces a risk amplification term based on spatial heterogeneity and the rate of change of normalized permeability for time-domain cumulative evaluation, the solid line reaches the trigger lower limit of 0.6 around time 400, entering the control threshold range and activating the warning. The prior art only responds around time 850, resulting in a significant delay in warning identification and an inability to effectively capture permeation risks in the early stages. After the solid line enters the trigger zone, the dashed line rises rapidly and generates an environmental adjustment increment, confirming that this invention can assign sufficient response weight to the actuator after identifying environmental stress. Subsequently, as the risk approaches saturation, the dashed line smoothly falls back, demonstrating that the threshold triggering mechanism can achieve early and accurate intervention in egg quality risks while ensuring the lifespan of the actuator.

[0084] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. An environmental control method based on visual monitoring of watermarked eggs, characterized in that, include: The translucent images of the eggshell surface at each time point are acquired according to a preset sampling interval, and the watermark area in the translucent images at each time point is identified. Taking any given moment as the current moment, obtain the total projected area of ​​the egg, calculate the watermark area ratio based on the ratio of the watermark area to the total projected area of ​​the egg, obtain the spatial gradient magnitude of each pixel within the watermark area, and determine the normalized permeability of the watermark area at the current moment based on the grayscale deviation and spatial gradient magnitude of each pixel at the current moment and the previous moment. Obtain the edge of the watermark region, divide it into several arc segments, calculate the normalized permeability of each arc segment, and statistically analyze the difference between the normalized permeability of each arc segment and the normalized permeability of the watermark region to determine the non-uniformity coefficient at each time point. Based on the proportion of watermark area, the normalized permeability of the watermark area, and the non-uniformity coefficient, the saturation risk after a set time is predicted; based on the saturation risk and the time change rate of the normalized permeability of the watermark area, an environmental regulation increment is generated. Environmental regulation and control are carried out by utilizing saturation risk and environmental regulation increments. Identifying watermark regions in the translucent image at various times includes: obtaining the grayscale value of a preset background reference; traversing each pixel in the translucent image, calculating the difference between the grayscale value of the corresponding pixel and the grayscale value of the preset background reference, and marking the set of pixels whose difference exceeds a preset deviation threshold as candidate regions; obtaining the gradient distribution sequence of the candidate region edge along the normal direction, calculating the spatial span covered by the non-zero gradient values ​​in the gradient distribution sequence, and recording it as the gradient support width; in response to the gradient support width falling into a preset diffusion width range, and the gradient distribution sequence showing a single-peak smooth transition trend, confirming the candidate region as a watermark region; The normalized permeability of the watermarked area satisfies the following relationship: ;in, yes Normalized permeability of the watermark area; It is the total projected area of ​​the egg; yes The total number of pixels within the watermark area; yes The first time within the watermark area The grayscale value of each pixel; It is the first Pixels in The grayscale value at any given moment; It is the sampling time interval; yes Time of the first Spatial gradient magnitude of each pixel; It is the preset first minute value; It is the first Area contribution factor of each pixel; It is the absolute value symbol; The non-uniformity coefficient satisfies the following relationship: ;in, yes Non-uniformity coefficient at time; It is the total number of arc segments; yes Time of the first Normalized penetration rate of each arc segment; It is the preset second minute value; Saturation risk satisfies the following relationship: ;in, yes The risk of saturation at any given moment; yes The percentage of the watermark area at any given moment; It is the current moment; It sets the duration; It is the environmental buffer coefficient; It is an integral variable; Based on the saturation risk and the time rate of change of the normalized permeability of the watermark area, an environmental regulation increment is generated, including: calculating the absolute value of the difference between the normalized permeability of the watermark area at the current time and the previous sampling time, and recording the ratio of the absolute value of the difference to the sampling time interval as the time rate of change; multiplying the product of the saturation risk and the time rate of change by a preset gain coefficient to obtain the environmental regulation increment.

2. The environmental control method based on visual monitoring of watermarked eggs according to claim 1, characterized in that, To obtain the total projected area of ​​the egg, including: The outline boundary of an egg in a translucent image is obtained using an edge extraction algorithm; Traverse the closed region enclosed by the outline boundary, count the total number of pixels contained in the closed region, and record it as the total projected area of ​​the egg.

3. The environmental control method based on visual monitoring of watermarked eggs according to claim 1, characterized in that, The watermark area percentage is calculated based on the ratio of the watermark area to the total projected area of ​​the egg, including: Count the total number of pixels contained within the watermark area; The ratio of the total number of pixels to the total projected area of ​​the egg is recorded as the watermark area percentage.

4. The environmental control method based on visual monitoring of watermarked eggs according to claim 1, characterized in that, Obtain the edge of the watermark area and divide it into several arc segments, including: Edge detection algorithms are used to extract the edges of the watermark region; Determine the geometric center point of the edge of the watermark area, and use the geometric center point as the origin of the polar coordinates to divide the edge of the watermark area into several arc segments according to the preset angle bisector step size.

5. The environmental control method based on visual monitoring of watermarked eggs according to claim 1, characterized in that, Environmental regulation and control are carried out using saturation risk and environmental regulation increments, including: When the saturation risk falls within the preset control threshold range, environmental parameters are adjusted incrementally using environmental regulation.

Images