Cat litter quality detection method based on intelligent sensor

Abstract

The application discloses a cat litter quality detection method based on intelligent sensors and belongs to the technical field of pet product quality detection. The method solves the problem that the existing cat litter automatic detection only relies on static end-point parameters to carry out evaluation and cannot completely reflect the working state of the litter under real working conditions. The method releases a quantitative simulation liquid into a detection container containing a cat litter sample to be detected, collects humidity response signals and gas response signals in the process of cat litter absorbing the simulation liquid through an intelligent sensor array, constructs a diffusion map of the internal humidity of the cat litter evolving with time according to the humidity response signals and extracts a moisture absorption performance parameter, constructs a release curve of the gas concentration above the cat litter changing with time according to the gas response signals and extracts a gas control performance parameter, and carries out weighted fusion on the two types of parameters to obtain a comprehensive quality index. The application can completely capture the dynamic evolution process of the cat litter performance, fully restore the real use working conditions and significantly improve the accuracy of the detection results.

Description

Technical Field

[0001] This invention relates to the field of pet product quality testing technology, and in particular to a method for testing cat litter quality based on intelligent sensors. Background Technology

[0002] With the rapid development of the pet breeding industry, the quality requirements for cat litter products continue to rise. However, the quality testing of cat litter generally adopts the static endpoint parameter collection method. This existing technology collects terminal steady-state data such as the final water absorption, steady-state water absorption time, and equilibrium odor concentration at the end of the testing process through sensing devices, and completes the judgment and rating of cat litter quality solely based on the above static endpoint parameters.

[0003] However, cat litter involves a continuous process of liquid infiltration and gas release during actual use. Its actual performance is determined by the dynamic state throughout the entire operating cycle. Performance evaluation based solely on the final steady-state results cannot fully reflect the working state of the litter under real-world conditions, resulting in a significant discrepancy between the test evaluation results and the actual user experience. Summary of the Invention

[0004] This invention provides a cat litter quality detection method based on intelligent sensors to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a cat litter quality detection method based on intelligent sensors, comprising:

[0006] S1. Release a quantitative amount of simulated liquid into a test container containing a cat litter sample to be tested, and collect the humidity response signal and gas response signal generated by the cat litter during the absorption of the simulated liquid through an intelligent sensor array;

[0007] S2. Based on the humidity response signal, construct a diffusion map of the humidity evolution inside the cat litter sample over time, and extract the moisture absorption performance parameters from the diffusion map;

[0008] S3. Based on the gas response signal, construct the gas concentration release curve of the space above the cat litter sample over time, and extract the gas control performance parameters from the release curve;

[0009] S4. Weighted fusion of moisture absorption performance parameters and control performance parameters yields a comprehensive quality index.

[0010] Preferably, the step of releasing a quantitative amount of simulated liquid into the detection container containing the cat litter sample to be tested includes:

[0011] Obtain the preset base release total, and decompose the base release total into multiple sub-release amounts to generate an initial release sequence;

[0012] The first sub-release quantity is released according to the initial release sequence, and after the release is completed, the initial humidity response signal returned by the humidity sensor in the smart sensor array is read.

[0013] Instantaneous slope analysis was performed on the initial humidity response signal to obtain the humidity influence coefficient characterizing the instantaneous absorption capacity of cat litter;

[0014] Based on the humidity influence coefficient, the remaining sub-release amounts in the initial release sequence and the waiting time between adjacent release actions are numerically adjusted to generate a differentiated release sequence.

[0015] The remaining sub-release amounts are released sequentially according to the differentiated release sequence.

[0016] Preferably, the step of collecting humidity response signals and gas response signals generated by cat litter during the absorption of simulated liquid using a smart sensor array includes:

[0017] Simultaneously activate the humidity sensor and gas sensor to acquire the raw humidity sequence and raw gas sequence;

[0018] Bidirectional decoupling correction was performed on the original humidity sequence and the original gas sequence to obtain the humidity response signal and the gas response signal;

[0019] The bidirectional decoupling correction includes: removing humidity measurement artifacts caused by gas concentration fluctuations in the original humidity sequence based on the original gas sequence, and removing gas measurement artifacts caused by humidity fluctuations in the original gas sequence based on the original humidity sequence.

[0020] Preferably, constructing a diffusion map of humidity evolution over time within the cat litter sample based on the humidity response signal includes:

[0021] Perform a sliding window cross-correlation operation on the humidity response signals of every two humidity sensors in the smart sensor array to obtain the time delay matrix between each pair of sensors;

[0022] Based on the time delay matrix and the preset spatial coordinates of each humidity sensor, the local propagation velocity vector at each sensor location is calculated to generate a propagation velocity vector field.

[0023] By integrating the velocity vector in the propagation velocity vector field along the time axis, the cumulative change path of the humidity value at each sensor location over time is reconstructed, and a diffusion map is obtained.

[0024] Preferably, the extraction of hygroscopic performance parameters from the diffusion spectrum includes:

[0025] The velocity amplitude at the sensor location with the largest velocity amplitude is extracted from the propagation velocity vector field and used as the peak moisture absorption rate.

[0026] The average moisture absorption rate is obtained by weighting the velocity amplitudes of all sensor locations in the propagation velocity vector field with the reciprocal of the distance from each sensor location to the simulated liquid release point.

[0027] The peak moisture absorption rate and the average moisture absorption rate are used as the moisture absorption rate parameters in the moisture absorption performance parameters.

[0028] Preferably, the extraction of hygroscopic performance parameters from the diffusion spectrum further includes:

[0029] Calculate the direction deviation angle of the velocity vector between each pair of adjacent sensors in the propagation velocity vector field, and generate a set of direction deviation angles;

[0030] The angular dispersion is obtained by statistically analyzing the set of directional deviation angles.

[0031] The angular dispersion is compared with a preset dispersion threshold. When the angular dispersion is less than the threshold, the cat litter is determined to be isotropic and uniformly diffused in terms of moisture absorption.

[0032] The angular dispersion or its normalized value is used as the moisture absorption uniformity parameter in the moisture absorption performance parameters.

[0033] Preferably, constructing a release curve of the gas concentration in the space above the cat litter sample over time based on the gas response signal includes:

[0034] Threshold detection is performed on the humidity response signal of each humidity sensor to determine the arrival time of the liquid front at each sensor location, and the earliest arrival time is selected as the time reference point for gas signal alignment.

[0035] Starting from the time reference point, the gas response signal at each sensor location is extracted, and the effective gas signal segment after the time reference point is retained.

[0036] The baseline release curve is obtained by averaging all the retained gas response signal segments.

[0037] A polynomial smoothing filter is applied to the baseline release curve to obtain the release curve of gas concentration changing over time.

[0038] Preferably, the extraction of gas control performance parameters from the release curve includes:

[0039] The time axis of the release curve is nonlinearly scaled based on the average moisture absorption rate, so that the scaled time axis is inversely proportional to the moisture absorption rate, thus obtaining the normalized moisture absorption release curve.

[0040] The area under the normalized release curve of the hygroscopicity is calculated within a preset integral interval to obtain the total amount of gas released. The preset integral interval is the cutoff time point from the starting point to the point where the signal first decays from the peak value to a preset percentage of the peak value.

[0041] Extract the first time length required for the hygroscopic normalized release curve to rise from the starting point to the peak concentration, and the second time length required for the concentration to decay to a preset ratio from the peak concentration. Use the first time length and the second time length as the gas release rate characteristic values.

[0042] The total gas release amount and the characteristic value of the gas release rate are used as the release capacity parameters in the gas control performance parameters.

[0043] Preferably, the extraction of gas control performance parameters from the release curve further includes:

[0044] Identify the decay phase of the release curve from the peak concentration and extract the signal segment of this decay phase.

[0045] Perform exponential decay fitting on the signal segment during the decay phase to obtain the decay time constant;

[0046] The reciprocal of the decay time constant is used as the characteristic value of the gas adsorption rate, and this characteristic value of the gas adsorption rate is used as the adsorption capacity parameter in the gas control performance parameters.

[0047] Preferably, the weighted fusion of the moisture absorption performance parameters and gas control performance parameters to obtain a comprehensive quality index includes:

[0048] The first fusion weight is determined based on the degree of deviation between the angular dispersion and the preset uniformity benchmark value, wherein the greater the degree of deviation, the smaller the first fusion weight;

[0049] The ratio of the average moisture absorption rate to the peak moisture absorption rate is calculated and used as the moisture absorption efficiency factor to characterize the spatial difference in moisture absorption rate.

[0050] Obtain the preset initial weights of gas performance, and multiply them by the moisture absorption efficiency factor as the modulation coefficient to obtain the dynamic second fusion weights;

[0051] Based on the first fusion weight and the dynamic second fusion weight, the moisture absorption performance parameters and gas control performance parameters are weighted and summed to obtain the comprehensive quality index.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] 1. By constructing a diffusion spectrum of the internal humidity of cat litter over time and a release curve of the gas concentration in the space above the cat litter over time, the corresponding moisture absorption performance parameters and gas control performance parameters are extracted and weighted and fused. This allows for the complete capture of the dynamic process of moisture absorption and gas control of cat litter throughout its entire working cycle, fully restoring the working state of the litter under real working conditions. The extracted performance parameters accurately characterize the actual usage characteristics of the product, significantly improving the consistency between quality test results and real user experience.

[0054] 2. By adapting the liquid release method to the characteristics of the sample being tested and cross-correcting the signals of multiple physical quantities, the detection requirements of cat litter with different characteristics can be accurately matched, ensuring the consistency between the detection conditions and the actual use scenario. At the same time, it effectively eliminates the mutual interference between different physical quantity signals and improves the fidelity of the original collected data. The synergistic effect of the two further enhances the accuracy and stability of the detection results. Attached Figure Description

[0055] Figure 1 This is a flowchart illustrating the cat litter quality detection method based on intelligent sensors provided by the present invention.

[0056] Figure 2 This is a flowchart illustrating the release curve construction method provided by the present invention. Detailed Implementation

[0057] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0059] Reference Figures 1-2 As shown, in this embodiment, the cat litter quality detection method based on smart sensors includes:

[0060] S1. Release a quantitative amount of simulated liquid into a test container containing a cat litter sample to be tested, and collect the humidity response signal and gas response signal generated by the cat litter during the absorption of the simulated liquid through an intelligent sensor array;

[0061] S2. Based on the humidity response signal, construct a diffusion map of the humidity evolution inside the cat litter sample over time, and extract the moisture absorption performance parameters from the diffusion map;

[0062] S3. Based on the gas response signal, construct the gas concentration release curve of the space above the cat litter sample over time, and extract the gas control performance parameters from the release curve;

[0063] S4. Weighted fusion of moisture absorption performance parameters and control performance parameters yields a comprehensive quality index.

[0064] In actual testing, this method relies on simulating real excretion conditions and simultaneously collecting and analyzing multi-dimensional signals to comprehensively evaluate the water absorption, clumping, odor suppression, and deodorization performance of cat litter.

[0065] In practice, a standardized, sealed, and temperature-controlled testing container is first constructed to effectively isolate external temperature, humidity, and environmental odors, ensuring consistent testing conditions for each group. Before testing, the cat litter to be tested—whether it's tofu litter, bentonite litter, or a mixture of both—is evenly spread inside the testing container, ensuring uniform thickness and compactness to avoid testing errors caused by differences in sample placement.

[0066] After the litter is applied, a simulated liquid is released quantitatively into the litter sample area in the container to simulate the cat's excretion behavior. Throughout the process of the litter absorbing the liquid, a distributed intelligent sensor array is used to continuously collect dynamic signals. The sensor array includes humidity sensors and gas sensors deployed at multiple points, respectively located inside the litter body and in the cavity above the litter body, which can capture real-time changes in moisture absorption and diffusion and odor release from all directions.

[0067] After acquiring high-precision humidity response signals and gas response signals, the diffusion spectrum of humidity evolution over time and the release curve of gas concentration change over time are reconstructed respectively. Based on this, moisture absorption performance parameters and gas control performance parameters that can quantify cat litter quality are extracted. Finally, a unified comprehensive quality index is calculated through an adaptive weighted fusion algorithm to achieve standardized and precise quantitative rating of cat litter quality.

[0068] In this embodiment of the invention, releasing a quantitative amount of simulated liquid into a detection container containing a cat litter sample to be tested includes:

[0069] Obtain the preset base release total, and decompose the base release total into multiple sub-release amounts to generate an initial release sequence;

[0070] The first sub-release quantity is released according to the initial release sequence, and after the release is completed, the initial humidity response signal returned by the humidity sensor in the smart sensor array is read.

[0071] Instantaneous slope analysis was performed on the initial humidity response signal to obtain the humidity influence coefficient characterizing the instantaneous absorption capacity of cat litter;

[0072] Based on the humidity influence coefficient, the remaining sub-release amounts in the initial release sequence and the waiting time between adjacent release actions are numerically adjusted to generate a differentiated release sequence.

[0073] The remaining sub-release amounts are released sequentially according to the differentiated release sequence.

[0074] In practice, injecting a large amount of liquid at once can cause the surface of the cat litter to clump together quickly, preventing the liquid from penetrating and resulting in insufficient internal moisture absorption, which affects the accuracy of the test data.

[0075] Therefore, this solution provides a preferred implementation method, which adopts an adaptive phased liquid release method to match the actual usage conditions of cats urinating in multiple times and cat litter continuously absorbing moisture.

[0076] In practice, the system pre-stores the basic release volume adapted to the single urination volume of a typical cat. This value can be calibrated through a large amount of pet behavior data and is preferably set to 8mL to adapt to industrial quality inspection and laboratory standard testing scenarios.

[0077] During the test, the preset basic release total is first broken down into multiple equal sub-release amounts to form an initial release sequence. For example, 8 mL of simulated liquid can be divided into four 2 mL sub-liquid volumes. In the initial state, the interval between two adjacent releases is uniformly preset to 10 seconds.

[0078] After initial sequence configuration is completed, the first sub-release operation is executed. This scheme uses sterile saline as the simulated liquid to accurately simulate the physicochemical properties of urine and to reproduce the real-world usage scenario to the greatest extent possible. After a single release is completed, real-time data from all humidity sensors is immediately read and integrated to form a complete initial humidity response signal.

[0079] Subsequently, an instantaneous slope analysis was performed on the signal. Specifically, the humidity change data within 0.5 seconds after the liquid release was completed was extracted to calculate the slope of the change. This yielded a humidity influence coefficient that characterizes the instantaneous water absorption activity of the cat litter. The larger the slope, the faster the instantaneous moisture absorption response and the better the water absorption activity of the cat litter.

[0080] It is understandable that cat litter of different materials has significantly different moisture absorption characteristics. Bentonite cat litter has a fast water absorption speed and high instantaneous activity, while tofu cat litter has a relatively slow water absorption process, and its fixed release rhythm cannot meet the testing requirements of all categories.

[0081] Therefore, the system can dynamically determine and adjust based on preset high and low response thresholds. When the humidity influence coefficient is higher than the high response threshold, it indicates that the cat litter has excellent moisture absorption performance, and the amount of residual ions released can be appropriately increased and the release interval time can be shortened. When the humidity influence coefficient is lower than the low response threshold, it indicates that the cat litter absorbs water slowly and has poor activity, and the amount of liquid released at one time can be reduced and the interval waiting time can be extended to avoid liquid accumulation and leakage.

[0082] After adjustment, a differentiated release sequence adapted to the characteristics of the cat litter being tested is generated. Subsequently, the remaining liquid is released in strict accordance with the optimized sequence to ensure that the release rhythm matches the actual moisture absorption capacity of the cat litter throughout the process.

[0083] In this embodiment of the invention, a smart sensor array is used to collect humidity response signals and gas response signals generated by cat litter during the absorption of simulated liquid, including:

[0084] Simultaneously activate the humidity sensor and gas sensor to acquire the raw humidity sequence and raw gas sequence;

[0085] Bidirectional decoupling correction was performed on the original humidity sequence and the original gas sequence to obtain the humidity response signal and the gas response signal;

[0086] The bidirectional decoupling correction includes: removing humidity measurement artifacts caused by gas concentration fluctuations in the original humidity sequence based on the original gas sequence, and removing gas measurement artifacts caused by humidity fluctuations in the original gas sequence based on the original humidity sequence.

[0087] In the specific implementation process, in order to solve the problems of crosstalk between humidity signals and gas signals, and the problem of multiple detection artifacts and low data accuracy caused by sensor environmental drift, this solution provides a preferred implementation method, namely, a signal processing method of synchronous acquisition and bidirectional decoupling correction, to obtain a pure and effective detection signal.

[0088] In practice, at the start of the simulated liquid release, all humidity and gas sensors are activated simultaneously to ensure that the acquisition timelines of the two types of signals are fully aligned and the timing is consistent. Dynamic data of the entire process of cat litter moisture absorption are continuously collected, and original humidity and gas sequences that change continuously over time are generated respectively.

[0089] It should be noted that the unprocessed raw sequence will have cross-interference issues. Small fluctuations in gas concentration will be misidentified as humidity changes by the humidity sensor, resulting in humidity measurement artifacts. At the same time, the drastic humidity fluctuations caused by the moisture absorption of the sand body will also interfere with the concentration recognition accuracy of the gas sensor, resulting in gas measurement artifacts. The superposition of the two types of artifacts will directly affect the accuracy of subsequent parameter extraction.

[0090] Therefore, this scheme implements bidirectional decoupling correction processing. Specifically, it can rely on the fluctuation characteristics of the original gas sequence to reversely eliminate artifact noise caused by gas disturbances in the original humidity sequence; at the same time, it can rely on the variation pattern of the original humidity sequence to correct the detection error caused by humidity fluctuations in the original gas sequence.

[0091] After bidirectional cross-correction and smoothing denoising, we finally obtained crosstalk-free, high-fidelity humidity response and gas response signals, providing a reliable data foundation for subsequent spectrum reconstruction, curve fitting, and performance parameter extraction.

[0092] In this embodiment of the invention, a diffusion map of humidity evolution over time within a cat litter sample is constructed based on the humidity response signal, including:

[0093] Perform a sliding window cross-correlation operation on the humidity response signals of every two humidity sensors in the smart sensor array to obtain the time delay matrix between each pair of sensors;

[0094] Based on the time delay matrix and the preset spatial coordinates of each humidity sensor, the local propagation velocity vector at each sensor location is calculated to generate a propagation velocity vector field.

[0095] By integrating the velocity vector in the propagation velocity vector field along the time axis, the cumulative change path of the humidity value at each sensor location over time is reconstructed, and a diffusion map is obtained.

[0096] In the specific implementation process, by reconstructing the humidity diffusion spectrum through the humidity response signal, the dynamic process of liquid penetration and diffusion between cat litter particles can be intuitively quantified, replacing the traditional crude evaluation method that relies solely on the single moisture absorption time, and realizing a refined quantitative analysis of the moisture absorption process.

[0097] In practice, the high-precision humidity response signal that has been calibrated is retrieved first, and then a sliding window cross-correlation operation is performed on any two paired humidity sensor signals in the sensor array.

[0098] It should be noted that the sliding window cross-correlation operation is a mature existing signal processing technology in this field. This solution does not improve the algorithm itself, but only uses this technology to accurately extract the time offset of the two timing signals. The sliding window duration can be preferably set to 0.5s, which can adapt to the dynamic response speed of cat litter moisture absorption and diffusion.

[0099] It is understandable that liquid diffusion will cause the humidity sensors at different locations to have signal response time differences. The greater the distance, the more obvious the response time difference. By cross-correlation calculation, the signal propagation time difference between each pair of paired sensors can be calculated in batches. Then, by integrating the time difference data of all paired sensors, a complete global time delay matrix can be constructed.

[0100] Specifically, this solution pre-calibrates and stores the three-dimensional spatial coordinates of all humidity sensors during the equipment manufacturing or testing initialization phase. The relative spacing and spatial position of each sensor are known fixed parameters, and there is no need for real-time calculation.

[0101] After obtaining the time delay matrix, the propagation speed and direction of the humidity signal can be calculated point by point by combining the known spatial distance of the sensors and the corresponding signal propagation time difference. Then, the local humidity propagation velocity vector corresponding to each sensor position can be obtained by integrating them. After summing the velocity vectors of all points, a propagation velocity vector field covering the entire sand body detection area can be generated.

[0102] Then, all velocity vectors in the vector field are integrated along the time axis to accumulate the humidity diffusion changes at each moment, restore the continuous change law of humidity at each sensor location over time, and finally reconstruct the complete diffusion spectrum of internal humidity of the cat litter sample over time, fully showing the spatiotemporal variation characteristics of liquid penetration range, diffusion speed, and diffusion uniformity.

[0103] In this embodiment of the invention, extracting hygroscopic performance parameters from the diffusion spectrum includes:

[0104] The velocity amplitude at the sensor location with the largest velocity amplitude is extracted from the propagation velocity vector field and used as the peak moisture absorption rate.

[0105] The average moisture absorption rate is obtained by weighting the velocity amplitudes of all sensor locations in the propagation velocity vector field with the reciprocal of the distance from each sensor location to the simulated liquid release point.

[0106] The peak moisture absorption rate and the average moisture absorption rate are used as the moisture absorption rate parameters in the moisture absorption performance parameters.

[0107] In the specific implementation process, the moisture absorption rate parameter is extracted based on the constructed propagation speed vector field. The speed of cat litter water absorption is comprehensively evaluated from two dimensions: the ultimate moisture absorption capacity and the overall average moisture absorption capacity, so as to avoid the problem of one-sided evaluation by a single parameter.

[0108] In practice, the velocity amplitude corresponding to all sensor points in the propagation velocity vector field is first traversed, and the maximum value in the entire field is selected and defined as the peak moisture absorption rate. This parameter can effectively characterize the local optimal instantaneous water absorption capacity of cat litter. For example, high-quality bentonite cat litter has a uniform pore structure and high water absorption activity, and its peak moisture absorption rate is significantly higher than that of low-quality cat litter with inferior powder doping and pore blockage.

[0109] Considering the optimal moisture absorption conditions around the liquid release point and the spatial differences in moisture absorption lag in the distant sand body, this scheme does not use a simple arithmetic average, but instead uses a distance-weighted average to calculate the overall moisture absorption rate.

[0110] Specifically, the reciprocal of the straight-line distance from each sensor location to the center point of liquid release is used as the weight. The closer the distance, the higher the weight, and the farther the distance, the lower the weight. The velocity amplitude of all sensors in the field is weighted and integrated through a weighted average algorithm to obtain the average moisture absorption rate that can truly reflect the overall moisture absorption efficiency of the sand body.

[0111] Ultimately, the peak moisture absorption rate and the average moisture absorption rate are combined to form the moisture absorption rate parameter, which serves as the core quantitative indicator for evaluating how quickly cat litter absorbs moisture.

[0112] In this embodiment of the invention, extracting hygroscopic performance parameters from the diffusion spectrum further includes:

[0113] Calculate the direction deviation angle of the velocity vector between each pair of adjacent sensors in the propagation velocity vector field, and generate a set of direction deviation angles;

[0114] The angular dispersion is obtained by statistically analyzing the set of directional deviation angles.

[0115] The angular dispersion is compared with a preset dispersion threshold. When the angular dispersion is less than the threshold, the cat litter is determined to be isotropic and uniformly diffused in terms of moisture absorption.

[0116] The angular dispersion or its normalized value is used as the moisture absorption uniformity parameter in the moisture absorption performance parameters.

[0117] In practice, by quantifying the angular discrete characteristics of moisture absorption and diffusion, the uniformity of cat litter moisture absorption can be accurately evaluated, and the clumping regularity and water absorption stability of cat litter can be effectively judged.

[0118] In practice, the velocity vector data of all adjacent sensors in the propagation velocity vector field are retrieved, and the directional deviation angle of each pair of adjacent vectors is calculated one by one. This angle can intuitively reflect the consistency of the liquid diffusion direction in the local area. The smaller the deviation angle, the more uniform the diffusion direction in the adjacent area and the more uniform the moisture absorption of the sand body. The larger the deviation angle, the more disordered the local diffusion and the uneven distribution of dry and wet.

[0119] The directional deviation angles corresponding to all adjacent sensors are summarized to form a complete set of angles. The global angular dispersion is obtained by calculating the standard deviation. The larger the angular dispersion value, the worse the overall moisture absorption and diffusion uniformity.

[0120] The system is configured with an adaptively calibrated discreteness judgment threshold. This threshold is calibrated by collecting global angular discreteness data of compliant high-quality cat litter samples from the market in batches. The upper limit of the discrete angle distribution of high-quality qualified products is used as the critical benchmark for uniform diffusion. It can be adaptively calibrated and updated according to different cat litter types and detection accuracy requirements. For example, under normal working conditions, 15° can be selected as a general judgment benchmark to assist in implementation, ensuring that the judgment standard fits the actual performance boundary of qualified products.

[0121] In specific testing and judgment, if the measured angular dispersion is less than the preset threshold, it can be determined that the moisture absorption process of the cat litter under test is isotropic and uniform, the liquid penetration is uniform, the clumping is regular, and the user experience is good; if the angular dispersion is greater than or equal to the preset threshold, it is determined that the moisture absorption and diffusion is uneven, the cat litter has the defects of local water absorption being too fast and local water absorption being delayed, and the overall clumping performance is poor.

[0122] During the parameter input and calculation stage, the original angular dispersion can be used directly, or the angular dispersion can be normalized and mapped to the 0-1 interval. The normalized value is used as the moisture absorption uniformity parameter, which facilitates subsequent multi-parameter fusion calculation. The maximum and minimum extreme value intervals of the normalization are adaptively calibrated based on the measured extreme values ​​of the samples with good and bad performance across the entire range, ensuring that the test results of different batches are consistent and comparable.

[0123] In this embodiment of the invention, based on the gas response signal, a release curve of the gas concentration in the space above the cat litter sample over time is constructed, including:

[0124] Threshold detection is performed on the humidity response signal of each humidity sensor to determine the arrival time of the liquid front at each sensor location, and the earliest arrival time is selected as the time reference point for gas signal alignment.

[0125] Starting from the time reference point, the gas response signal at each sensor location is extracted, and the effective gas signal segment after the time reference point is retained.

[0126] The baseline release curve is obtained by averaging all the retained gas response signal segments.

[0127] A polynomial smoothing filter is applied to the baseline release curve to obtain the release curve of gas concentration changing over time.

[0128] In the specific implementation process, in order to solve the problem of misalignment of gas signal start time and curve distortion caused by inconsistent liquid infiltration time at different spatial locations, this solution adopts the humidity front alignment method to construct a standardized gas release curve to ensure the accuracy of gas characteristic extraction.

[0129] In practice, the humidity response signals of all humidity sensors are first thresholded. The system is configured with an adaptively calibrated humidity change judgment threshold. This threshold is calibrated based on the steady-state baseline fluctuation range of the sensor. A value slightly larger than the no-load steady-state fluctuation amplitude is selected as the change threshold to accurately distinguish between environmental baseline noise and real liquid penetration signals, avoiding false triggering and missed detection problems. For example, under normal working conditions, 1.2 times the steady-state baseline value can be used as the judgment benchmark.

[0130] When the real-time humidity value of the sensor exceeds the threshold, it can be determined that the liquid penetration front has reached the sensor location, and the arrival time of each sensor is recorded synchronously.

[0131] Then, by comparing the arrival times of all points, the minimum time is selected as the unified time reference point for the global gas signal. This moment is the true starting point for cat litter moisture absorption and gas production, which can completely eliminate the signal deviation caused by spatial moisture absorption lag.

[0132] Using this time reference point as a unified starting point, the raw signals of all gas sensors are extracted, and blank noise segments that have no gas production significance before the reference point are removed, while the effective gas signal segments after the reference point are retained.

[0133] Then, the effective signal segments of all sensors are averaged to eliminate single-point sensor sampling errors and obtain a smooth reference release curve.

[0134] Finally, polynomial smoothing filtering can be applied to the baseline release curve. In this embodiment, second-order polynomial filtering is preferred. The smoothing characteristics of second-order polynomial can accurately adapt to the slow-gradual change characteristics of gas concentration, which is "slowly rising-peak-slowly falling". It can effectively filter out high-frequency noise caused by environmental disturbances and completely retain the true trend and characteristic inflection point of concentration change. The same type of smoothing algorithm can be flexibly replaced according to the signal noise intensity, and finally a standardized, high-fidelity release curve of gas concentration changing with time can be obtained.

[0135] In this embodiment of the invention, extracting gas control performance parameters from the release curve includes:

[0136] The time axis of the release curve is nonlinearly scaled based on the average moisture absorption rate, so that the scaled time axis is inversely proportional to the moisture absorption rate, thus obtaining the normalized moisture absorption release curve.

[0137] The area under the normalized release curve of the hygroscopicity is calculated within a preset integral interval to obtain the total amount of gas released. The preset integral interval is the cutoff time point from the starting point to the point where the signal first decays from the peak value to a preset percentage of the peak value.

[0138] Extract the first time length required for the hygroscopic normalized release curve to rise from the starting point to the peak concentration, and the second time length required for the concentration to decay to a preset ratio from the peak concentration. Use the first time length and the second time length as the gas release rate characteristic values.

[0139] The total gas release amount and the characteristic value of the gas release rate are used as the release capacity parameters in the gas control performance parameters.

[0140] In the specific implementation process, in order to avoid the difference in moisture absorption rate of different cat litters interfering with the accuracy of odor performance evaluation and to achieve fair benchmarking of odor control performance of different types and qualities of cat litters, this solution uses a combination of moisture absorption rate normalization correction and integral statistics to extract cat litter odor gas release capacity parameters in a comprehensive and multi-dimensional manner.

[0141] In practice, the calculated average moisture absorption rate is retrieved first, and this core parameter is used as the correction benchmark. Then, the time axis of the original gas release curve is subjected to nonlinear scaling transformation. It should be noted that the transformation rule is a fixed mapping relationship preset by the system, and the scaled time axis is strictly inversely proportional to the moisture absorption rate of the cat litter.

[0142] The preset mapping principle is in line with physical reality: the faster the cat litter absorbs moisture, the more concentrated the odor release cycle, and the time axis needs to be compressed to avoid the peak being overstretched; the slower the cat litter absorbs moisture, the more gradual the odor release process, and the time axis needs to be stretched to fully capture the decay characteristics.

[0143] Understandably, for cat litter with a fast moisture absorption rate, the time axis of its gas release curve can be appropriately compressed; for cat litter with a slow moisture absorption rate, the time axis of its gas release curve can be appropriately stretched. This is to unify the testing standards for cat litter with different characteristics, completely eliminate the detection interference caused by the difference in moisture absorption speed on the gas release time and concentration change curve, and finally obtain a standardized normalized moisture absorption release curve.

[0144] Subsequently, this scheme presets a 30% peak value ratio as the cutoff benchmark for the integration interval. It should be noted that this ratio parameter has a clear preset calibration basis. The specific preset method is as follows: by searching through a large number of measured gas decay curves, it is determined that after the gas concentration decays to 30% of the peak value, the odor release basically enters the weak background noise stage, and the effective odor release characteristics have been fully presented. Therefore, 30% is selected as a uniform cutoff ratio, which can stably cut off the effective gas production range and avoid the interference of ineffective background noise in the later stage.

[0145] Specifically, the starting point of the integration interval is the time reference point for gas signal alignment, and the ending point is the moment when the peak gas concentration of the normalized release curve first decays to 30% of the peak concentration.

[0146] After defining the integration interval, the normalized curve within that interval is integrated. The area under the curve is the total gas release of the cat litter. The lower this value, the less odorous gas the cat litter releases during the entire moisture absorption process, and the better its initial odor suppression performance.

[0147] Meanwhile, two sets of key time characteristic parameters were extracted from the hygroscopic normalized release curve. First, the duration from the start point to the peak concentration was extracted and defined as the first time length, which is used to characterize the intensity of the rapid release of odor gas. Then, the duration from the peak concentration to 30% of the peak concentration was extracted and defined as the second time length, which is used to characterize the natural dissipation rate of odor gas.

[0148] Understandably, the two sets of time parameters can reflect the speed of odor release and the ability to dissipate, respectively. Integrating the two can form the characteristic value of gas release rate. Finally, combining the total amount of gas released with the characteristic value of gas release rate forms the release capacity parameter in the gas control performance parameters, thus completing the comprehensive quantification of the odor release performance of cat litter.

[0149] In this embodiment of the invention, extracting gas control performance parameters from the release curve further includes:

[0150] Identify the decay phase of the release curve from the peak concentration and extract the signal segment of this decay phase.

[0151] Perform exponential decay fitting on the signal segment during the decay phase to obtain the decay time constant;

[0152] The reciprocal of the decay time constant is used as the characteristic value of the gas adsorption rate, and this characteristic value of the gas adsorption rate is used as the adsorption capacity parameter in the gas control performance parameters.

[0153] In the specific implementation process, the odor adsorption capacity of cat litter is accurately extracted by exponential decay fitting, which makes up for the deficiency that simply detecting the gas release amount cannot evaluate the subsequent deodorization performance and improves the gas control performance evaluation system.

[0154] In practice, the first step is to accurately identify the changing stages of the gas release curve, distinguish between the rising segment, the peak plateau segment, and the decaying segment, and separately extract the signal segment of the decaying segment where the concentration continues to fall from the peak, and remove irrelevant interval data to avoid interfering with the fitting accuracy.

[0155] For the extracted attenuation signal segments, a first-order exponential decay model is used for fitting. It should be noted that the first-order exponential model is a preset fixed fitting model of the system. Its preset principle is: the dynamic process of physical adsorption of cat litter porous media and natural diffusion and dissipation of gas completely conforms to the change law of first-order exponential decay. The model's adaptability has been verified by a large amount of experimental data and no custom correction is required.

[0156] It is understandable that the process of cat litter porous media adsorbing odor gases and the natural dissipation of gases completely conforms to the physical law of first-order exponential decay, and the fitting results are consistent with actual working conditions.

[0157] After fitting, a unique decay time constant can be obtained. This constant directly characterizes the rate of gas decay and the cat litter's adsorption capacity. The larger the decay time constant, the slower the gas decay and the weaker the cat litter's continuous adsorption and deodorization capacity. The smaller the decay time constant, the higher the deodorization adsorption efficiency.

[0158] To adapt to the normalization calculation requirements of subsequent weighted fusion, the reciprocal of the decay time constant can be further defined as the characteristic value of gas adsorption rate. The larger the characteristic value, the better the adsorption performance. This characteristic value can be used as the adsorption capacity parameter in the gas control performance parameters to achieve a quantitative evaluation of the dynamic deodorization performance of cat litter.

[0159] In this embodiment of the invention, the moisture absorption performance parameters and gas control performance parameters are weighted and fused to obtain a comprehensive quality index, including:

[0160] The first fusion weight is determined based on the degree of deviation between the angular dispersion and the preset uniformity benchmark value, wherein the greater the degree of deviation, the smaller the first fusion weight;

[0161] The ratio of the average moisture absorption rate to the peak moisture absorption rate is calculated and used as the moisture absorption efficiency factor to characterize the spatial difference in moisture absorption rate.

[0162] Obtain the preset initial weights of gas performance, and multiply them by the moisture absorption efficiency factor as the modulation coefficient to obtain the dynamic second fusion weights;

[0163] Based on the first fusion weight and the dynamic second fusion weight, the moisture absorption performance parameters and gas control performance parameters are weighted and summed to obtain the comprehensive quality index.

[0164] In the specific implementation process, a dynamic adaptive weighted fusion method is adopted to integrate moisture absorption performance and gas control performance, so that the overall quality indicators are more in line with the actual user experience.

[0165] In specific implementation, the system presets the moisture absorption uniformity benchmark value to 0. It should be noted that the principle of this benchmark value is: the theoretical optimal ideal state of the angular dispersion after normalization, that is, the ideal working condition in which the liquid diffuses completely isotropically and uniformly without any deviation in the diffusion direction, serves as a unified benchmark zero point for global comparison, which facilitates the quantitative calculation of the deviation between the measured sample and the optimal performance.

[0166] In practice, the deviation of the measured normalized angle dispersion from the benchmark value is calculated. The greater the deviation, the more obvious the defect in the uniformity of cat litter moisture absorption. The first fusion weight corresponding to the moisture absorption performance is lowered to weaken the negative impact of defect parameters on the comprehensive score and ensure the objectivity and fairness of the evaluation.

[0167] Then, the moisture absorption efficiency factor is calculated. Specifically, the ratio of the average moisture absorption rate to the peak moisture absorption rate is used to characterize the spatial uniformity of the moisture absorption rate. The closer the ratio is to 1, the more balanced the moisture absorption rate is and the better the performance consistency. The smaller the ratio, the greater the difference in local moisture absorption performance and the poorer the overall stability.

[0168] Among them, the initial weight of 0.5 for gas performance is the preset benchmark weight of the system. The preset calibration method of this weight is as follows: based on the big data statistics of user experience of a large number of commercial cat litter, the impact ratio of cat litter moisture absorption and clumping performance on user experience is close to 1:1. Therefore, 0.5 is preset as the initial weight for balance to ensure that the basic evaluation dimensions are balanced and unbiased.

[0169] Then, the moisture absorption efficiency factor is used as the dynamic modulation coefficient. The modulation coefficient is multiplied by the initial weight to adaptively generate a dynamically changing second fusion weight, so that the gas performance weight can be adjusted in real time according to the moisture absorption balance of the cat litter.

[0170] Finally, by using the first fusion weight and the dynamic second fusion weight, the normalized moisture absorption performance parameters and gas control performance parameters are weighted and summed to calculate the comprehensive quality index in the 0-1 range. The higher the index value, the better the comprehensive performance of the cat litter in terms of moisture absorption, clumping, odor suppression and deodorization, thereby achieving a precise quantitative rating of cat litter quality.

[0171] In the several embodiments provided by this invention, it should be understood that the disclosed method can be implemented in other ways.

[0172] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0173] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, and technology that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0174] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for detecting cat litter quality based on intelligent sensors, characterized in that, The method includes: S1. Release a quantitative amount of simulated liquid into a test container containing a cat litter sample to be tested, and collect the humidity response signal and gas response signal generated by the cat litter during the absorption of the simulated liquid through an intelligent sensor array; S2. Based on the humidity response signal, construct a diffusion map of the humidity evolution inside the cat litter sample over time, and extract the moisture absorption performance parameters from the diffusion map; S3. Based on the gas response signal, construct the gas concentration release curve of the space above the cat litter sample over time, and extract the gas control performance parameters from the release curve; S4. Weighted fusion of moisture absorption performance parameters and control performance parameters yields a comprehensive quality index.

2. The cat litter quality detection method based on intelligent sensors as described in claim 1, characterized in that, The process of releasing a quantitative amount of simulated liquid into the detection container containing the cat litter sample to be tested includes: Obtain the preset base release total, and decompose the base release total into multiple sub-release amounts to generate an initial release sequence; The first sub-release quantity is released according to the initial release sequence, and after the release is completed, the initial humidity response signal returned by the humidity sensor in the smart sensor array is read. Instantaneous slope analysis was performed on the initial humidity response signal to obtain the humidity influence coefficient characterizing the instantaneous absorption capacity of cat litter; Based on the humidity influence coefficient, the remaining sub-release amounts in the initial release sequence and the waiting time between adjacent release actions are numerically adjusted to generate a differentiated release sequence. The remaining sub-release amounts are released sequentially according to the differentiated release sequence.

3. The cat litter quality detection method based on intelligent sensors as described in claim 1, characterized in that, The process of collecting humidity and gas response signals generated by cat litter during the absorption of simulated liquid using a smart sensor array includes: Simultaneously activate the humidity sensor and gas sensor to acquire the raw humidity sequence and raw gas sequence; Bidirectional decoupling correction was performed on the original humidity sequence and the original gas sequence to obtain the humidity response signal and the gas response signal; The bidirectional decoupling correction includes: removing humidity measurement artifacts caused by gas concentration fluctuations in the original humidity sequence based on the original gas sequence, and removing gas measurement artifacts caused by humidity fluctuations in the original gas sequence based on the original humidity sequence.

4. The cat litter quality detection method based on intelligent sensors as described in claim 1, characterized in that, The process of constructing a diffusion map of humidity evolution over time within a cat litter sample based on the humidity response signal includes: Perform a sliding window cross-correlation operation on the humidity response signals of every two humidity sensors in the smart sensor array to obtain the time delay matrix between each pair of sensors; Based on the time delay matrix and the preset spatial coordinates of each humidity sensor, the local propagation velocity vector at each sensor location is calculated to generate a propagation velocity vector field. By integrating the velocity vector in the propagation velocity vector field along the time axis, the cumulative change path of the humidity value at each sensor location over time is reconstructed, and a diffusion map is obtained.

5. The cat litter quality detection method based on intelligent sensors as described in claim 4, characterized in that, The extraction of hygroscopic performance parameters from the diffusion spectrum includes: The velocity amplitude at the sensor location with the largest velocity amplitude is extracted from the propagation velocity vector field and used as the peak moisture absorption rate. The average moisture absorption rate is obtained by weighting the velocity amplitudes of all sensor locations in the propagation velocity vector field with the reciprocal of the distance from each sensor location to the simulated liquid release point. The peak moisture absorption rate and the average moisture absorption rate are used as the moisture absorption rate parameters in the moisture absorption performance parameters.

6. The cat litter quality detection method based on intelligent sensors as described in claim 5, characterized in that, The extraction of hygroscopic performance parameters from the diffusion spectrum also includes: Calculate the direction deviation angle of the velocity vector between each pair of adjacent sensors in the propagation velocity vector field, and generate a set of direction deviation angles; The angular dispersion is obtained by statistically analyzing the set of directional deviation angles. The angular dispersion is compared with a preset dispersion threshold. When the angular dispersion is less than the threshold, the cat litter is determined to be isotropic and uniformly diffused in terms of moisture absorption. The angular dispersion or its normalized value is used as the moisture absorption uniformity parameter in the moisture absorption performance parameters.

7. The cat litter quality detection method based on intelligent sensors as described in claim 6, characterized in that, The step of constructing a release curve of the gas concentration in the space above the cat litter sample over time based on the gas response signal includes: Threshold detection is performed on the humidity response signal of each humidity sensor to determine the arrival time of the liquid front at each sensor location, and the earliest arrival time is selected as the time reference point for gas signal alignment. Starting from the time reference point, the gas response signal at each sensor location is extracted, and the effective gas signal segment after the time reference point is retained. The baseline release curve is obtained by averaging all the retained gas response signal segments. A polynomial smoothing filter is applied to the baseline release curve to obtain the release curve of gas concentration changing over time.

8. The cat litter quality detection method based on intelligent sensors as described in claim 7, characterized in that, The extraction of gas control performance parameters from the release curve includes: The time axis of the release curve is nonlinearly scaled based on the average moisture absorption rate, so that the scaled time axis is inversely proportional to the moisture absorption rate, thus obtaining the normalized moisture absorption release curve. The area under the normalized release curve of the hygroscopicity is calculated within a preset integral interval to obtain the total amount of gas released. The preset integral interval is the cutoff time point from the starting point to the point where the signal first decays from the peak value to a preset percentage of the peak value. Extract the first time length required for the hygroscopic normalized release curve to rise from the starting point to the peak concentration, and the second time length required for the concentration to decay to a preset ratio from the peak concentration. Use the first time length and the second time length as the gas release rate characteristic values. The total gas release amount and the characteristic value of the gas release rate are used as the release capacity parameters in the gas control performance parameters.

9. The cat litter quality detection method based on intelligent sensors as described in claim 8, characterized in that, The extraction of gas control performance parameters from the release curve also includes: Identify the decay phase of the release curve from the peak concentration and extract the signal segment of this decay phase. Perform exponential decay fitting on the signal segment during the decay phase to obtain the decay time constant; The reciprocal of the decay time constant is used as the characteristic value of the gas adsorption rate, and this characteristic value of the gas adsorption rate is used as the adsorption capacity parameter in the gas control performance parameters.

10. The cat litter quality detection method based on intelligent sensors as described in claim 9, characterized in that, The weighted fusion of moisture absorption performance parameters and gas control performance parameters yields a comprehensive quality index, including: The first fusion weight is determined based on the degree of deviation between the angular dispersion and the preset uniformity benchmark value, wherein the greater the degree of deviation, the smaller the first fusion weight; The ratio of the average moisture absorption rate to the peak moisture absorption rate is calculated and used as the moisture absorption efficiency factor to characterize the spatial difference in moisture absorption rate. Obtain the preset initial weights of gas performance, and multiply them by the moisture absorption efficiency factor as the modulation coefficient to obtain the dynamic second fusion weights; Based on the first fusion weight and the dynamic second fusion weight, the moisture absorption performance parameters and gas control performance parameters are weighted and summed to obtain the comprehensive quality index.

Images