Water volume monitoring method and system applied to pet water dispenser

Abstract

The invention relates to the technical field of water volume monitoring, in particular to a water volume monitoring method and system applied to a pet water dispenser, and the method comprises the steps: obtaining multi-view image data of each monitoring area in the pet water dispenser, the multi-view image data comprises a top top view image and a side front view image, the overlook water volume coverage of the pet water dispenser is determined according to the top overlook image, and the front water volume coverage of the pet water dispenser is determined according to the side front image; and determining the consumption active state of the water body in each monitoring area according to the temperature and humidity of the environment where the pet water dispenser is located and all the water volume coverage degrees. The water volume coverage is obtained through the multi-view images, the consumption process of the water body can be accurately monitored and simulated in combination with environmental factors such as temperature and humidity, the water volume consumption can be monitored in real time through the dynamic simulation, the change of the water body can be predicted in advance, and it is guaranteed that the water volume management of the pet water dispenser is more scientific and accurate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of water quantity monitoring, in particular to a water quantity monitoring method and system applied to a pet water dispenser. BACKGROUND

[0002] In recent years, the pet industry has rapidly grown, especially as pets have become an important member of the family, the health management of pets has become increasingly important, and the performance and intelligent level of pet water dispensers, as part of pet health management, directly affect the quality of life and health level of pets.

[0003] Currently, traditional methods rely on manual inspection and judgment of water consumption, and lack the ability of automation and real-time monitoring. This method is easily affected by human error and cannot quickly respond to changes. It often relies on simple water level measurement or flow sensors, with limited precision, and cannot dynamically monitor and predict details, especially when the water consumption rate changes rapidly. Traditional methods are difficult to provide accurate prediction and early warning.

[0004] In addition, traditional methods usually ignore the influence of environmental factors and cannot adjust and predict water consumption according to these changes, which may lead to inaccurate prediction of water consumption when the environment changes significantly. Generally, only the overall water consumption is focused on, without fine-grained analysis of the consumption dynamics of specific monitoring areas, so it cannot provide targeted water replenishment recommendations, especially in areas with rapid consumption. Moreover, it usually cannot predict changes in water quantity in advance, and only when the water level falls below a certain critical point will it be discovered. Such delays can lead to water shortage in pets and affect their health. SUMMARY

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solution: a water quantity monitoring method applied to a pet water dispenser, comprising: Obtaining multi-view image data of each monitoring area in the pet water dispenser, wherein the multi-view image data includes a top-down image and a side view image, determining the top-down water quantity coverage of the pet water dispenser according to the top-down image, and determining the front view water quantity coverage of the pet water dispenser according to the side view image; Determining the consumption activity state of the water body in each monitoring area according to the temperature and humidity of the environment where the pet water dispenser is located and all water quantity coverages, dynamically simulating the consumption process of the water body in the pet water dispenser through all consumption activity states, and obtaining the water quantity consumption trajectory of each monitoring area; The water consumption distribution of the water in the pet water dispenser is obtained by fusing and reorganizing all the water consumption trajectories based on the position distance between the monitoring areas, and the water consumption distribution is used to determine a plurality of water quantity warning areas of the pet water dispenser, wherein the water quantity warning area is used to represent the area in which the water consumption rate of each monitoring area of the pet water dispenser is higher than the average level. The water flow dynamic data of each water quantity warning area is measured, and the water consumption flux of the water in each water quantity warning area in a unit time is determined according to the flow velocity gradient between the water flow dynamic data, wherein the water consumption flux is used to represent the parameter value of the water consumption amount per unit time through the water quantity warning area. When the water consumption flux is greater than the water quantity threshold, a water replenishment reminder is sent to the mobile terminal of the pet owner.

[0006] Preferably, the top-down water quantity coverage of the pet water dispenser is determined according to the top-down image, and the front view water quantity coverage of the pet water dispenser is determined according to the side view image, including: The first water area and the first non-water area in the top-down image are extracted by using an image segmentation algorithm; The first area ratio of the first water area in the total area of the top-down image is calculated, and the first area ratio is determined as the top-down water quantity coverage of the pet water dispenser; The second water area and the second non-water area in the side view image are extracted by using an image segmentation algorithm; The second area ratio of the second water area in the total area of the side view image is calculated, and the second area ratio is determined as the front view water quantity coverage of the pet water dispenser.

[0007] Preferably, the consumption active state of the water in each monitoring area is determined according to the temperature and humidity of the environment in which the pet water dispenser is located and all the water quantity coverages, including: The coordinate position of each monitoring area and the temperature and humidity of the environment in which the pet water dispenser is located are obtained; The water quantity deviation degree of each monitoring area is determined according to all the water quantity coverages; The consumption active state of the water in each monitoring area is determined according to the temperature and humidity and all the water quantity deviation degrees.

[0008] Preferably, the consumption process of the water in the pet water dispenser is dynamically simulated by all the consumption active states, and the water consumption trajectory of each monitoring area is obtained, including: The water flow velocity of the monitoring area and a plurality of consumption vectors of the monitoring area are determined; The water consumption velocity of the monitoring area is determined according to the consumption active state corresponding to the monitoring area and the water flow velocity; determining a water consumption trajectory of a monitoring area according to the water consumption speed and the consumption vector.

[0009] Preferably, all water consumption trajectories are recombined based on the position distance between each monitoring area to obtain a water consumption distribution map of the water body in the pet water dispenser, including: determining the position distance between each monitoring area; determining a water consumption network of the pet water dispenser according to all position distances; selecting one monitoring area as a selected monitoring area; recombining the water consumption trajectory of the selected monitoring area according to the water flow speed of the water consumption network and the selected monitoring area to obtain a recombined water consumption trajectory of the selected monitoring area; continuing to recombine the water consumption trajectories of the remaining monitoring areas; determining a water consumption distribution map of the water body in the pet water dispenser according to all recombined water consumption trajectories.

[0010] Preferably, a plurality of water consumption warning areas of the pet water dispenser are determined through the consumption distribution map, including: identifying an area with a water consumption rate higher than the average level based on the consumption distribution map; determining the identified area as a plurality of water consumption warning areas of the pet water dispenser.

[0011] Preferably, the consumption flux of the water body in each water consumption warning area per unit time is determined according to the flow velocity gradient between each water flow dynamic data, including: determining the flow velocity gradient between each water flow dynamic data; selecting one water consumption warning area as a selected water consumption warning area; determining a consumption flux base value of the selected water consumption warning area according to the flow velocity gradient; determining the consumption flux of the water body in the selected water consumption warning area per unit time based on the consumption flux base value and the water flow dynamic data of the selected water consumption warning area; continuing to determine the consumption flux of the water body in the remaining water consumption warning areas per unit time.

[0012] Preferably, after determining the consumption flux of the water body in the selected water consumption warning area per unit time based on the consumption flux base value and the water flow dynamic data of the selected water consumption warning area, the method further includes: determining each target area in the multi-view image data, wherein each target area includes a water body area and each interference area; establishing a pixel matrix of each target area in the multi-view image data, wherein the pixel matrix is a digital image matrix composed of a plurality of pixels; determine a correlation between each of the interference regions and the water body region in the multi-view image data based on the pixel matrix; determine whether there is an interference region that has a correlation with the water body region reaching a preset interference threshold based on the correlation; If yes, correct the consumption flux based on a third area ratio of the interference region that has a correlation reaching the threshold in the total area of the multi-view image data.

[0013] Preferably, the correlation between each of the interference regions and the water body region in the multi-view image data based on the pixel matrix is determined, comprising: For the pixel matrix of each of the target regions, the pixel matrix of each of the target regions is respectively subjected to convolution calculation, frequency spectrum transformation, cross-correlation calculation and entropy value calculation to obtain a time domain feature matrix, a frequency domain feature matrix, a cross-correlation matrix and an entropy value matrix; The pixel matrix, the time domain feature matrix, the frequency domain feature matrix, the cross-correlation matrix and the entropy value matrix are summarized to obtain a target matrix of the target region; The target matrix of each of the target regions in each of the first base images is input into a pre-trained correlation model to obtain the correlation between each of the interference regions and the water body region in each of the first base images; The training method of the pre-trained correlation model comprises: Collect multi-view image data samples under different interference scenarios of the pet water dispenser, and label the water body region and the interference region in the multi-view image data samples; Construct a pixel matrix and a corresponding target matrix of the multi-view image data samples; Label the real correlation between the interference region and the water body region in each of the multi-view image data samples as a sample label; Take the target matrix as input and the sample label as output, train an initial neural network model, adjust the model parameters until convergence, and obtain the correlation model.

[0014] The water quantity monitoring system applied to the pet water dispenser is suitable for the water quantity monitoring method applied to the pet water dispenser, comprising: A water quantity monitoring unit is configured to acquire multi-view image data of each monitoring region in the pet water dispenser, wherein the multi-view image data comprises a top-down image and a side view image, determine a top-down water quantity coverage of the pet water dispenser based on the top-down image, and determine a front view water quantity coverage of the pet water dispenser based on the side view image; A consumption simulation unit is configured to determine the consumption active state of the water body in each monitoring area according to the temperature and humidity of the environment where the pet water dispenser is located and the water coverage of all the monitoring areas, to dynamically simulate the consumption process of the water body in the pet water dispenser through all the consumption active states, and to obtain the water consumption trajectory of each monitoring area; A water consumption warning unit is configured to fuse and recombine all the water consumption trajectories based on the position distance between the monitoring areas, to obtain the consumption distribution map of the water body in the pet water dispenser, and to determine the water consumption warning areas of the pet water dispenser through the consumption distribution map, wherein the water consumption warning areas are used to represent the areas where the water consumption rate in each monitoring area of the pet water dispenser is higher than the average level. A consumption calculation unit is configured to measure the water flow dynamic data of each water consumption warning area, to determine the consumption flux of the water body in each water consumption warning area per unit time according to the flow velocity gradient between the water flow dynamic data, and to use the consumption flux to represent the parameter value of the water body consumption per unit time through the water consumption warning area. A water replenishment reminding unit is configured to send a water replenishment reminder to the mobile terminal of the pet owner when the consumption flux is greater than the water threshold.

[0015] Compared with the prior art, the present application has the following advantages: The present application can accurately monitor and simulate the consumption process of the water body by obtaining the water coverage from the multi-view images and combining the environmental factors such as temperature and humidity. This dynamic simulation can not only monitor the water consumption in real time, but also predict the water body changes in advance, so as to ensure that the water management of the pet water dispenser is more scientific and accurate. Through the analysis of the water flow dynamic data to obtain the consumption flux of the water body, when the water consumption exceeds the set threshold, the pet owner can be actively reminded to replenish water. This function helps the pet owner to replenish water in time, ensures that the pet always has sufficient water supply, and avoids health problems caused by insufficient water supply for the pet. The present application can accurately obtain the water coverage of each monitoring area by extracting the water body area and the non-water body area and calculating the area ratio. At the same time, the consumption active state of the water body is determined by combining the temperature and humidity with the water coverage, so as to improve the accuracy of water body monitoring and reduce the manual intervention in the traditional method. Through the position relationship between the monitoring areas, the water consumption trajectories of all the monitoring areas are fused to generate the water consumption distribution map, and the water consumption warning areas are further divided. This intelligent warning can effectively identify the areas where the water consumption speed is accelerated, provide an early warning for the pet owner, and avoid unnecessary waste or insufficient water supply for the pet. The application can effectively process interference factors, improve the accuracy and robustness of water quantity monitoring, further improve the accuracy of water quantity consumption through the analysis of flow velocity gradient and the correction of interference area consumption flux, and ensure that the water consumption of the pet water dispenser can reflect the changes in each area in real time and accurately. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 It is a step flow diagram of the overall method in an embodiment of the application. Figure 2 It is a system architecture diagram of the overall system in an embodiment of the application.

[0017] In the figure: 1, water quantity monitoring unit; 2, consumption simulation unit; 3, water quantity early warning unit; 4, consumption calculation unit; 5, water replenishment reminding unit. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0019] Embodiment one, please refer to Figure 1 The application provides a technical solution: a water quantity monitoring method applied to a pet water dispenser, comprising: S1, acquiring multi-view image data of each monitoring area in the pet water dispenser, wherein the multi-view image data comprises a top-down image and a side view image, determining a top-down water quantity coverage of the pet water dispenser according to the top-down image, and determining a front view water quantity coverage of the pet water dispenser according to the side view image; S2, determining the consumption active state of the water body in each monitoring area according to the temperature and humidity of the environment where the pet water dispenser is located and all the water quantity coverages, dynamically simulating the consumption process of the water body of the pet water dispenser through all the consumption active states, and obtaining water quantity consumption trajectories of each monitoring area; S3, fusing and recombining all the water quantity consumption trajectories based on the position distance between each monitoring area, obtaining a consumption distribution map of the water body in the pet water dispenser, and determining a plurality of water quantity early warning areas of the pet water dispenser through the consumption distribution map, wherein the water quantity early warning area is used to represent the area where the water body consumption rate in each monitoring area of the pet water dispenser is higher than the average level; S4, measure the water flow dynamic data of each water quantity early warning area, and determine the consumption flux of the water body in each water quantity early warning area in a unit time according to the flow velocity gradient between each water flow dynamic data, wherein the consumption flux is used to represent the parameter value of the water body consumption amount through the water quantity early warning area in a unit time; S5, when the consumption flux is greater than the water quantity threshold value, a water supplement reminder is sent to the mobile terminal of the pet owner.

[0020] It should be noted that the top-down image and the side elevation image are images of two perspectives, which respectively provide the top and side perspectives of the water dispenser; the top-down image is used to observe the water surface coverage of the pet water dispenser and estimate the distribution of the water quantity on the surface of the water dispenser (i.e. the water surface area); according to the water surface area, the top-down water quantity coverage is calculated, which represents the distribution degree of the water; the side elevation image is used to observe the height of the water level and the change of the water surface level, and estimate the coverage of the water quantity on the water level from the side; this can reflect the depth and consumption rate of the water; the temperature and humidity of the environment will affect the evaporation of the water body and the drinking behavior of the pet; combined with these environmental data and the water quantity coverage, the activity degree of the water consumption of each monitoring area can be inferred; for example, when the temperature is higher, the water consumption may accelerate; when the humidity is higher, the evaporation speed may decrease; if the pet is active in drinking water, the water consumption will accelerate; According to the water quantity coverage and the consumption activity state of each monitoring area, the trajectory of the water consumption is calculated through dynamic simulation; this can be understood as the process of the gradual decrease of the water level over time; the position distance between each monitoring area refers to the spatial relationship of each area of the water dispenser; by analyzing the consumption trajectory of the water body in different areas, the consumption trajectories of each area can be fused into a whole to generate a consumption distribution map; this map can show the water quantity consumption situation of different areas, for example, the consumption speed of some areas is faster and the water level decreases faster; according to the consumption distribution map, the areas with faster water quantity consumption speed can be identified, which may become water quantity early warning areas; these areas consume water faster, which may cause the water level to decrease and the water concentration to decrease; for example, the water consumption in some areas of the water dispenser is particularly rapid, which may indicate that the pet frequently drinks water or the evaporation in this area is faster; the water flow dynamic data refers to the changes of the flow velocity, flow rate, etc. of the water body; these data can be used to analyze the consumption situation of the water body, and the consumption flux of each water quantity early warning area is calculated through the flow velocity gradient; the consumption flux is a description of the consumption amount of the water body in a unit time, which represents the rate of water loss; if the consumption flux of a certain area is greater than the set water quantity threshold value, it means that the water consumption in this area is too fast, and water supplement may be needed; when the system detects that the consumption flux of a certain water quantity early warning area exceeds the water quantity threshold value, the system will send a water supplement reminder to the mobile terminal of the pet owner; this can remind the pet owner to supplement water for the pet in time to avoid the low water quantity affecting the health of the pet.

[0021] In an optional embodiment, the top view water coverage of the pet water dispenser is determined according to the top view image, and the front view water coverage of the pet water dispenser is determined according to the side view image, comprising: The first water region and the first non-water region in the top view image are extracted by using an image segmentation algorithm; The first area ratio of the first water region in the total area of the top view image is calculated, and the first area ratio is determined as the top view water coverage of the pet water dispenser; The second water region and the second non-water region in the side view image are extracted by using an image segmentation algorithm; The second area ratio of the second water region in the total area of the side view image is calculated, and the second area ratio is determined as the front view water coverage of the pet water dispenser.

[0022] It should be noted that image segmentation is a computer vision technology that divides an image into different regions, each containing pixels with similar features such as color, brightness, texture, etc.; here, image segmentation is used to separate the water region from the non-water region in the image; the first water region refers to the part of the top view image that is identified as water (e.g. the water surface); the first non-water region refers to the area of the top view image that does not contain water (e.g. the container part of the water dispenser or other background parts); after identifying the first water region by the image segmentation algorithm, the area ratio in the entire top view image can be calculated; the specific steps are as follows: calculate the area of the first water region: this is usually the number of pixels or the actual area occupied by the water region in the image; calculate the area of the total region: that is, the area of the entire image region of the top view image; area ratio: then, calculate the area ratio of the water region, i.e. first area ratio = area of first water region ÷ area of total region × 100%; this ratio value is the top view water coverage, reflecting the proportion of the water surface in the top view of the water dispenser; Similar to the top view image, the water and non-water regions in the side view image also need to be extracted by image segmentation; the second water region: the area in the side view image that is identified as water (e.g. the part of the water surface seen from the side); the second non-water region: the part of the side view image that does not contain water (e.g. the frame of the water dispenser or other objects); once the second water region is extracted, the ratio in the side view image can be calculated; the steps are similar to those of the top view image: calculate the area of the second water region: this is the number of pixels or the actual area of the water surface region in the side view image; calculate the area of the total region: that is, the area of the entire image region of the side view image; area ratio: second area ratio = area of second water region ÷ area of total region × 100%; this ratio value is the front view water coverage, reflecting the proportion of the water region in the side view of the water dispenser; The top view water coverage and the front view water coverage are important indicators for measuring the distribution of water in the pet water fountain. The top view water coverage reflects the coverage of the water surface from the top, which can represent the distribution and range of the water on the surface of the water fountain. The front view water coverage reflects the coverage of the water surface from the side, which can represent the depth of the water and the change of the water level. These data can be used for subsequent water consumption monitoring and early warning mechanisms to help pet owners understand the water level in the water fountain in real time.

[0023] In an optional embodiment, the consumption activity of the water in each monitoring area is determined according to the temperature and humidity of the environment where the pet water fountain is located and all water coverages, including: Obtaining the coordinate positions of each monitoring area and the temperature and humidity of the environment where the pet water fountain is located; Determining the water deviation degree of each monitoring area according to all water coverages; Determining the consumption activity of the water in each monitoring area according to the temperature and humidity and all water deviation degrees.

[0024] It should be noted that the coordinate positions of each monitoring area refer to the positions of the monitoring points set in different areas of the water fountain (such as the top, side, bottom, etc.). Each monitoring area records its position in the image through accurate coordinates to ensure that the data collection is accurate. These coordinates may be pixel positions in the image or may be defined by sensors in the actual space. The temperature and humidity refer to the temperature and humidity conditions of the environment where the pet water fountain is located. These data are usually obtained in real time by environmental sensors. Temperature and humidity are key factors affecting water consumption, for example, high temperature environment may cause water evaporation to accelerate, and low humidity may accelerate water consumption, etc. Water coverage is an indicator for measuring the distribution of water in each monitoring area. For example, the water coverage extracted from the top view and the side view reflects the area ratio of the water surface under different angles. Water deviation degree is used to measure the difference between actual water distribution and expected water distribution. For each monitoring area, compare the water coverage with the ideal or expected water coverage. If the actual water coverage is low or high, it means that there is a deviation in the water quantity, which may be related to the use of the pet water fountain, environmental changes or equipment problems; The consumption active state refers to the consumption rate of the water body and the state of water level change, which is jointly affected by temperature, humidity, and water quantity deviation: temperature and humidity directly affect the evaporation rate of water and the usage frequency of the pet water fountain; high temperature environment usually leads to accelerated evaporation of the water body, while low humidity environment may cause more intense evaporation of the water body; lower temperature or higher humidity environment may slow down the consumption speed of water; water quantity deviation reflects the distribution of the water body in each monitoring area; if the water quantity deviation of a certain area is large, it may indicate that the water body is unevenly distributed or there is a problem with the equipment, causing abnormal consumption of the water body in that area; the water quantity deviation combined with environmental conditions can help determine whether the consumption of the water body in a certain area is in a normal state; by comprehensively considering temperature, humidity, and water quantity deviation, the consumption active state of each monitoring area can be evaluated; for example, in a high temperature and low humidity environment, if the water quantity deviation of a certain area is large, it may mean that the water consumption in that area is abnormally active, and water source needs to be replenished or the working status of the water fountain needs to be checked.

[0025] In an optional embodiment, the consumption process of the water body of the pet water fountain is dynamically simulated through all the consumption active states, and the water consumption trajectory of each monitoring area is obtained, including: determining the water flow speed of the monitoring area and multiple consumption vectors of the monitoring area; determining the water consumption speed of the monitoring area according to the consumption active state corresponding to the monitoring area and the water flow speed; determining the water consumption trajectory of the monitoring area according to the water consumption speed and the consumption vector.

[0026] It should be noted that the water flow speed refers to the flow speed of the water body in a certain monitoring area; it is usually closely related to factors such as the flow direction, flow rate, temperature, humidity, and pressure of the water body; for example, the water in some areas may flow quickly due to the influence of gravity, while the water flow speed in other areas may be slow due to equipment design or environmental factors; by monitoring the state of the water flow, the flow speed of the water body in each area can be determined; specifically, the water flow speed is usually related to the consumption active state of the water; for example, in a higher temperature and lower humidity environment, water may evaporate or be consumed more easily, which may cause the water flow to accelerate; by detecting the speed and direction of the water flow, the dynamic changes of the water body can be better understood; the consumption vector refers to the direction and size of the consumption of the water body in a certain monitoring area; the consumption vector can be multiple because the consumption of the water body is not a single direction or location process; for example, water may evaporate, be drunk by pets, or leak due to equipment problems; through the data of multiple monitoring points, the direction of water consumption (such as the direction of pet drinking or the location of evaporation) can be depicted, and the consumption process can be quantified; the multiple consumption vectors of each monitoring area can help analyze which factors cause the consumption of water; The consumption active state reflects the degree of water consumption, which is closely related to environmental temperature and humidity, equipment working state, and water coverage, etc. For example, if the temperature in a certain area is very high and the humidity is very low, the water in the area may evaporate quickly or be consumed by pets, so the consumption active state is high. Combined with the flow speed and the consumption active state, the water consumption speed in a certain area can be estimated. If the water flow speed is fast and the consumption active state is high (for example, in hot weather, pets are active in drinking water), the water consumption speed will be fast. Conversely, if the flow speed is slow and the consumption active state is low (for example, in a low-temperature environment, the frequency of drinking water is low), the consumption speed will be relatively slow. By comprehensively considering these two factors, the water consumption speed of each monitoring area can be determined, that is, the consumption rate of water in the area per unit time. The water consumption trajectory refers to the specific path and change trajectory of water consumption in the monitoring area. Through the previously calculated consumption speed and consumption vector, the spatial distribution and temporal evolution of water consumption can be inferred. For example, if the consumption active state of a certain area is high and the flow speed is fast, the consumption trajectory of the water body may be that the water body starts from a specific location in the area, gradually reduces over time, and expands to the surrounding area. By monitoring these trajectories, not only can the water consumption trend of different monitoring areas be evaluated, but also the dynamic changes of water consumption can be tracked in real time, thereby providing data support for further optimizing the design and use of pet water dispensers.

[0027] In an optional embodiment, all water consumption trajectories are reorganized based on the position distances between the monitoring areas to obtain a water consumption distribution map of the pet water dispenser, including: Determining the position distances between the monitoring areas; Determining the water consumption network of the pet water dispenser according to all position distances; Selecting a monitoring area as a selected monitoring area; Reorganizing the water consumption trajectory of the selected monitoring area according to the water consumption network and the water flow speed of the selected monitoring area to obtain the reorganized water consumption trajectory of the selected monitoring area; Continuing to reorganize the water consumption trajectories of the remaining monitoring areas; Determining the water consumption distribution map of the pet water dispenser according to all reorganized water consumption trajectories.

[0028] It is necessary to clarify the relative positions between the monitoring areas; these areas can be different positions inside or around the water dispenser, and the water consumption in each position can be different; by measuring the physical distance between each area (e.g., the straight-line distance between areas), the relationship between different areas can be known; this distance can help understand the flow of water between different areas and the degree of consumption; once the position distance between each monitoring area is determined, a "water consumption network" can be established; this means that the correlation of water consumption between each monitoring area needs to be analyzed; for example, the distance between some areas is closer, which may cause the water to flow faster, while the farther areas may have different trends in water consumption due to slower flow speed; in this way, a whole consumption network can be formed to help understand the relationship between water consumption in different areas; After analyzing the entire water consumption network, a specific monitoring area is selected as the "selected monitoring area"; this area can be the most concerned place or a representative area for data analysis; by selecting a monitoring area as a reference, it can be analyzed more deeply to observe its water consumption trajectory and trend; after selecting the monitoring area, the water consumption process in this area will be analyzed based on the "water consumption network" and the water flow speed of the selected area; the flow speed of water directly affects the consumption rate of water, while the water consumption network reveals the relationship between water consumption in different areas; by combining these two factors, the water consumption trajectory of the area can be reorganized to obtain more accurate consumption data; next, the water consumption trajectories of other monitoring areas also need to be reorganized; the water flow speed, consumption state, and positional relationship between different areas may be different, so the consumption trajectory of each area may be different; by performing similar reorganization analysis on all remaining areas, the water consumption trajectory of each area can be obtained; By integrating the water consumption trajectories of all reorganized monitoring areas, a distribution map of water consumption in the pet water dispenser can be drawn; this map will show the water consumption in different areas, showing which areas consume more and which areas consume less; this is of great significance for optimizing water distribution, improving the efficiency of the water dispenser, and adjusting the water distribution strategy according to the consumption needs of different areas.

[0029] In an optional embodiment, a plurality of water consumption warning areas of the pet water dispenser are determined based on the consumption distribution map, including: Identifying areas with water consumption rates higher than the average level based on the consumption distribution map; Determining the identified areas as the plurality of water consumption warning areas of the pet water dispenser.

[0030] It is necessary to explain that in the water consumption distribution map, the water consumption of each area will be different; low water consumption area refers to those areas with less water consumption, which is characterized by low consumption rate or low consumption itself; by looking at the consumption distribution map, these areas can be found; low water consumption area may be caused by various reasons, such as location is more edge, low frequency of use, or slow water flow speed; the water consumption of these areas is relatively low, which may mean that these areas are not frequently used or have less water supply in water distribution; they are usually represented as "cold" areas in the consumption map, which may appear as lighter or lower consumption value areas; once the low water consumption area is identified, the next step is to set the "water warning area" according to the situation of these areas; water warning area refers to when the water consumption of a certain area reaches a lower threshold, the system will trigger an alarm to remind the manager that it may be necessary to adjust the water distribution, check whether the equipment is normal, or take other measures to ensure the continuous supply of drinking water machine; After determining which areas are low water consumption areas, multiple water warning areas can be planned according to the distribution of these areas; these warning areas will usually cover low water consumption areas and set a certain range around them to ensure that when water consumption further decreases, measures can be taken in time; each low water consumption area will have a corresponding water consumption threshold; when the consumption of a certain area falls below this threshold, the system will send an alarm signal; the setting of this threshold is usually based on the actual drinking water demand and water usage frequency; the role of the warning area is to find the abnormal place of water consumption as soon as possible and remind the management personnel to take measures; this can help to adjust the water distribution in time to avoid some areas from being unable to use the drinking water machine due to lack of water, and to ensure that the pet drinking fountain always operates efficiently and stably.

[0031] In an optional embodiment, the consumption flux of the water body in each water flow dynamic data is determined according to the flow velocity gradient between the water flow dynamic data, comprising: determining the flow velocity gradient between the water flow dynamic data; selecting a water warning area as a selected water warning area; determining the consumption flux basic value of the selected water warning area according to the flow velocity gradient; determining the consumption flux of the water body in the selected water warning area per unit time based on the consumption flux basic value and the water flow dynamic data of the selected water warning area; continue to determine the consumption flux of the water body in the remaining water warning area per unit time.

[0032] It is necessary to note that the flow rate gradient refers to the rate of change of water flow speed between different areas; the speed of water flow may vary in different water quantity warning areas; by comparing the flow rates of these areas, the flow rate gradient can be calculated; the flow rate gradient reflects the trend of water flow between areas, and generally a larger flow rate gradient indicates that the water flow changes rapidly in certain areas, while a smaller flow rate gradient indicates that the water flow is relatively stable; the flow rate gradient is very important for subsequent calculation of consumption flux, as it affects the transfer and consumption speed of water between areas; a high flow rate gradient may mean that water flow is more likely to be rapidly consumed, while a low flow rate gradient means that water flow may be more gradual and slower to consume; according to the flow rate gradient and consumption, a representative or most needed water quantity warning area is selected as the basis for subsequent calculation; this selected area may be an area with a larger flow rate gradient or more significant water consumption; the selected water quantity warning area is usually an area where water consumption is more concentrated or at risk, so as to focus on monitoring; The selected criteria may include factors such as water flow speed, consumption, frequency of area use, etc.; by considering these factors comprehensively, it is ensured that the selected area is representative and can effectively reflect the trend of water consumption; the consumption flux base value refers to the preliminary water consumption of the selected water quantity warning area under the influence of the flow rate gradient; this is a base value used to estimate the water flow consumption of the area; generally, the consumption flux base value takes into account the flow rate, water quantity of the area, and other dynamic data such as temperature, humidity, etc.; the flow rate gradient affects the speed of water flow, thereby affecting the consumption of water quantity; in areas with a larger flow rate gradient, water flow is faster and the consumption flux base value is higher; in areas with a smaller flow rate gradient, water flow is slower and the consumption flux base value is lower; once the consumption flux base value is determined, the consumption flux of the area per unit time can be further refined based on the dynamic data of water flow (such as water flow rate, flow, etc.); the dynamic data of water flow helps to understand how water is consumed in a given time; in addition to flow rate, water body consumption in the water quantity warning area is also affected by other factors such as flow time, water quantity fluctuation of the area, etc.; based on these data, the consumption of water in the selected water quantity warning area per unit time can be further determined; this data is crucial for water quantity warning and scheduling; After determining the consumption flux of the selected water quantity warning area, similar calculations need to be performed for the remaining water quantity warning areas; the water flow dynamic data of each warning area may be different, so the consumption flux of each area may also be different; for each remaining water quantity warning area, its water consumption flux per unit time is calculated based on its respective flow rate gradient and water flow dynamic data; this process may involve different calculation models or empirical formulas to ensure that the consumption of each area can be accurately evaluated.

[0033] In an optional embodiment, after determining the consumption flux of the water body in the selected water quantity warning area per unit time based on the consumption flux base value and the water flow dynamic data of the selected water quantity warning area, the method further comprises: determining each target area in the multi-view image data, wherein each target area comprises a water body area and each interference area; establishing a pixel matrix of each target area in the multi-view image data, wherein the pixel matrix is a digital image matrix composed of multiple pixels; determining the correlation between each interference area and the water body area in the multi-view image data based on the pixel matrix; judging whether there is an interference area that has a correlation with the water body area reaching a preset interference threshold according to the correlation; if yes, correcting the consumption flux based on the third area ratio of the interference area with a correlation exceeding the threshold in the total area of the multi-view image data.

[0034] It should be noted that in the multi-view image data, the target area includes a "water body area" and an "interference area"; the water body area refers to the part of the image containing the water body, while the interference area refers to other areas that may affect the analysis of the water body or have an impact on the water body area, such as pollutants, buildings or other objects; these target areas are identified and labeled through image recognition technology (such as image segmentation, edge detection, etc.); each area is classified as a water body area or an interference area according to its characteristics (such as color, texture, shape, etc.); this process may involve the use of computer vision algorithms, such as convolutional neural networks (CNN), etc.; each target area is represented as a pixel matrix; a digital image is composed of multiple pixels, and the values of these pixels reflect the brightness, color, etc. information of different areas in the image; the pixel matrix is a matrix structure arranged in digital form, containing all the pixel data of the area in the image; the pixel matrix provides detailed data of the target area, which can help further analyze the properties within the area, such as color, spectral characteristics, etc.; through these matrices, the spatial distribution characteristics of each target area can be analyzed in depth, and the basis for subsequent correlation analysis is provided; The correlation relationship refers to the similarity or proximity between the interference region and the water body region in terms of space and features in the image. For example, the interference region may have some similarity with the water body region in terms of color, texture or other physical properties, or the interference region may be directly in contact with the water body region. The correlation relationship can be determined by calculating the pixel similarity between the interference region and the water body region. Common methods include pixel value difference analysis, color histogram comparison, texture analysis, etc. If the pixels of the interference region and the pixels of the water body region are similar or close in some features, it can be considered that there is a certain correlation between the two. A preset threshold is set to determine whether the correlation relationship between the interference region and the water body region is significant. When the correlation relationship between the interference region and the water body region exceeds the threshold, it is considered that the interference region has a greater impact on the water body region and may need to be corrected. If the correlation between the interference region and the water body region is higher than the preset threshold (for example, the pixel value of the interference region is similar to that of the water body region), the interference region is considered to be a significant interference factor. At this time, further processing is needed. Once it is determined that the correlation relationship of certain interference regions exceeds the interference threshold, correction needs to be made in the consumption flux calculation of the water body region. The influence of the interference region may cause the speed of water body consumption to change, so it is very important to correct the consumption flux. The basis for correction is the proportion of these interference regions in the image. If the third area proportion of these interference regions that exceed the threshold in the total area of the multi-view image data is large, it indicates that the influence of these interference regions on the overall water body consumption flux is more significant. Therefore, according to the degree of influence of these interference regions, the consumption flux of the water body region is adjusted, which may reduce consumption or dynamically correct through other methods.

[0035] In an optional embodiment, the correlation relationship between each interference region and the water body region in the multi-view image data is determined based on the pixel matrix, which includes: For the pixel matrix of each target region, the pixel matrix of the target region is respectively subjected to convolution calculation, frequency spectrum transformation, cross-correlation calculation and entropy value calculation to obtain a time domain feature matrix, a frequency domain feature matrix, a cross-correlation matrix and an entropy value matrix. The pixel matrix, the time domain feature matrix, the frequency domain feature matrix, the cross-correlation matrix and the entropy value matrix are summarized to obtain a target matrix of the target region. The target matrix of each target region in each first base image is input into a pre-trained correlation relationship model to obtain the correlation relationship between each interference region and the water body region in each first base image. The training method of the pre-trained correlation relationship model includes: Collect multi-view image data samples of pet water dispensers under different interference scenarios, and label the water area and interference area in the multi-view image data samples; Construct a pixel matrix of the multi-view image data samples and a corresponding target matrix; Label the true association relationship between the interference area and the water area in each multi-view image data sample as a sample label; Take the target matrix as input and the sample label as output, train the initial neural network model, adjust the model parameters until convergence, and obtain the association relationship model.

[0036] It should be noted that four different calculations and transformations are performed on the pixel matrix of each target area: convolution is a common operation in image processing that can extract feature information from images; through convolution, important features in the target area can be identified, such as edges, textures, etc.; spectral transformation is a method of converting images from time domain to frequency domain, usually using Fourier transform; through spectral transformation, frequency features in the image can be extracted to help analyze the periodicity and details of the image; cross-correlation is a way to measure the similarity between two signals or image regions; by calculating the cross-correlation of the target area pixel matrix with other regions or standard templates, the correlation between the target area and other parts can be identified; entropy is an indicator of image complexity and uncertainty; by calculating the entropy value, the distribution of information within the target area can be evaluated, reflecting the complexity of the region; These calculations will respectively obtain four matrices: time domain feature matrix: reflects the features of the image in the time domain; frequency domain feature matrix: reflects the features of the image in the frequency domain; cross-correlation matrix: reflects the correlation between the target area and other areas; entropy value matrix: reflects the complexity and information amount of the target area; then, these four matrices are combined with the original pixel matrix to form a more comprehensive matrix, called the target matrix; this target matrix contains various feature information of the target area, including time domain, frequency domain, cross-correlation, and complexity, which can comprehensively describe the characteristics of the target area; each target area in each image will obtain a corresponding target matrix; next, each target matrix is input into the pre-trained association relationship model; the task of this model is to predict the association relationship between the target area and the water area or the interference area based on the features of the target area; specifically, the model needs to judge whether the relationship between the interference area and the water area is strong, thereby providing a basis for subsequent analysis and correction; To obtain this model, a large amount of sample data needs to be trained; the training process includes the following steps: collecting image data of pet water dispensers in different interference situations from different scenes; these data cover a variety of interference scenarios, including water area and interference area; in these image samples, the water area and interference area of each image need to be labeled to ensure that the model can identify these areas; extract the pixel matrix and target matrix from these labeled image samples; the target area of each image corresponds to a target matrix, which contains the feature information of the area; label the actual relationship between the interference area and the water area for each sample data; these labels are the target output of model learning, indicating the real relationship between each interference area and the water area, which may be "strong correlation", "weak correlation" or "no correlation"; use the target matrix as input and the label of the correlation relationship as output to train the initial neural network model; by continuously adjusting the parameters of the network, the model will learn how to predict the relationship between the interference area and the water area according to the target matrix; the training will continue until the prediction effect of the model reaches the predetermined accuracy; through the above training process, a trained and converged correlation relationship model is finally obtained; this model can accurately predict the correlation degree between each interference area and the water area according to the features in the target matrix; the model can be used for subsequent image analysis to help identify and correct the interference of the water area.

[0037] Embodiment two, please refer to Figure 2 The present application provides a technical solution: a water quantity monitoring system applied to a pet water dispenser, which is suitable for the water quantity monitoring method applied to a pet water dispenser and comprises: A water quantity monitoring unit 1 is configured to obtain multi-view image data of each monitoring area in the pet water dispenser, wherein the multi-view image data includes a top-down image and a side view image, the top-down water quantity coverage of the pet water dispenser is determined according to the top-down image, and the front view water quantity coverage of the pet water dispenser is determined according to the side view image. A consumption simulation unit 2 is configured to determine the consumption active state of the water body in each monitoring area according to the temperature and humidity of the environment where the pet water dispenser is located and all water quantity coverages, dynamically simulate the consumption process of the water body in the pet water dispenser through all consumption active states, and obtain the water quantity consumption trajectory of each monitoring area. A water quantity warning unit 3 is configured to fuse and recombine all water quantity consumption trajectories based on the position distance between each monitoring area, obtain the consumption distribution map of the water body in the pet water dispenser, and determine a plurality of water quantity warning areas of the pet water dispenser through the consumption distribution map, wherein the water quantity warning area is used to represent the area where the water consumption rate in each monitoring area of the pet water dispenser is higher than the average level. The consumption calculating unit 4 is configured to measure water flow dynamic data of each water quantity early warning area, and determine a consumption flux of the water body in each water quantity early warning area in a unit time according to a flow rate gradient between the water flow dynamic data, wherein the consumption flux is used to represent a parameter value of the water body consumption amount through the water quantity early warning area in a unit time; The water replenishment reminding unit 5 is configured to send a water replenishment reminder to the mobile terminal of the pet owner when the consumption flux is greater than the water quantity threshold value.

[0038] The above detailed description of the embodiments of the present application is made in combination with the accompanying drawings, but the present application is not limited thereto, and various changes can be made within the knowledge of those skilled in the art without departing from the purpose of the present application.

Claims

1. A method for monitoring water volume in pet water fountains, characterized in that, include: Acquire multi-view image data of each monitoring area in the pet water fountain, wherein the multi-view image data includes a top top view image and a side front view image. Determine the top view water coverage of the pet water fountain based on the top top view image, and determine the front view water coverage of the pet water fountain based on the side front view image. Based on the temperature and humidity of the environment where the pet water fountain is located and the water coverage, the active state of water consumption in each monitoring area is determined. The water consumption process of the pet water fountain is dynamically simulated through all active states of consumption to obtain the water consumption trajectory of each monitoring area. Based on the location distance between each monitoring area, all water consumption trajectories are fused and recombined to obtain a water consumption distribution map in the pet water fountain. Multiple water consumption warning zones of the pet water fountain are determined through the consumption distribution map. The water consumption warning zones are used to characterize areas in each monitoring area of ​​the pet water fountain where the water consumption rate is higher than the average level. Measure the dynamic data of water flow in each water volume warning zone, and determine the water consumption flux in each water volume warning zone per unit time based on the velocity gradient between the dynamic data of water flow. The consumption flux is used as a parameter value to characterize the amount of water consumed through the water volume warning zone per unit time. When the water consumption exceeds the water consumption threshold, a water replenishment reminder is sent to the pet owner's mobile device.

2. The water volume monitoring method for pet water fountains according to claim 1, characterized in that, Determining the top-view water coverage of the pet water fountain based on the top top view image, and determining the front-view water coverage of the pet water fountain based on the side front view image, including: An image segmentation algorithm is used to extract the first water body region and the first non-water body region from the top-view image; Calculate the first area ratio of the first water body area in the total area of ​​the top view image, and determine the first area ratio as the top view water coverage of the pet water fountain; The second water body region and the second non-water body region in the side front view image are extracted using an image segmentation algorithm; Calculate the second area ratio of the second water body region in the total area of ​​the side front view image, and determine the second area ratio as the front view water coverage of the pet water fountain.

3. The water volume monitoring method for pet water fountains according to claim 2, characterized in that, The activity level of water consumption in each monitoring area was determined based on the temperature and humidity of the environment where the pet water fountain was located, as well as the coverage of all water volumes. This included: Obtain the coordinates of each monitoring area and the temperature and humidity of the environment where the pet water fountain is located; Determine the water volume deviation of each monitoring area based on all water volume coverage; The water consumption activity level in each monitoring area is determined based on the temperature, humidity, and all water volume deviations.

4. The water volume monitoring method for pet water fountains according to claim 3, characterized in that, The water consumption process of the pet water fountain is dynamically simulated by analyzing all active consumption states, resulting in the water consumption trajectory of each monitoring area, including: Determine the water flow velocity and multiple consumption vectors within the monitoring area; The water consumption rate of the monitoring area is determined based on the consumption activity status of the monitoring area and the water flow velocity. The water consumption trajectory of the monitoring area is determined based on the water consumption rate and the consumption vector.

5. The water volume monitoring method for pet water fountains according to claim 4, characterized in that, Based on the locational distances between various monitoring areas, all water consumption trajectories are fused and reconstructed to obtain a water consumption distribution map within the pet water fountain, including: Determine the locational distances between each monitoring area; The water consumption network of the pet water fountain is determined based on all location distances; Select a monitoring area as the designated monitoring area; Based on the water consumption network and the water flow velocity in the selected monitoring area, the water consumption trajectory of the selected monitoring area is reconstructed to obtain the reconstructed water consumption trajectory of the selected monitoring area. Continue to reconstruct the water consumption trajectory of the remaining monitoring areas; A water consumption distribution map of the pet water fountain was determined based on the water consumption trajectory of all recombined water volumes.

6. The water volume monitoring method for pet water fountains according to claim 5, characterized in that, The consumption distribution map identifies multiple water level warning zones for the pet water fountain, including: Based on the consumption distribution map, areas where the water consumption rate is higher than the average level are identified; The identified areas were designated as multiple water level warning zones for the pet water fountain.

7. The water volume monitoring method for pet water fountains according to claim 6, characterized in that, The water consumption flux per unit time within each water volume warning zone is determined based on the velocity gradient between various water flow dynamic data, including: Determine the velocity gradient between various dynamic water flow data; Select a water volume warning zone as the designated water volume warning zone; The baseline value of the consumption flux for the selected water volume early warning zone is determined based on the velocity gradient. The water consumption flux of the selected water volume warning area per unit time is determined based on the baseline value of the consumption flux and the dynamic water flow data of the selected water volume warning area. Continue to determine the water consumption flux per unit time in the remaining water volume warning area.

8. The water volume monitoring method for pet water fountains according to claim 7, characterized in that, After determining the water consumption flux per unit time of the selected water volume warning area based on the baseline value of consumption flux and the dynamic water flow data of the selected water volume warning area, the method further includes: Identify each target region in the multi-view image data, wherein each target region includes a water body region and each interference region; Establish a pixel matrix for each target region in the multi-view image data, wherein the pixel matrix is ​​a digital image matrix composed of multiple pixels; The correlation between each interference region and the water body region in the multi-view image data is determined based on the pixel matrix; Based on the aforementioned correlation, determine whether there exists an interference area whose correlation with the water body reaches a preset interference threshold; If so, the flux consumption is corrected based on the third area proportion of the interference region exceeding the threshold in the total area of ​​the multi-view image data.

9. The water volume monitoring method for pet water fountains according to claim 8, characterized in that, Determining the correlation between each interference region and the water body region in the multi-view image data based on the pixel matrix includes: For each pixel matrix of the target region, convolution calculation, spectral transformation, cross-correlation calculation and entropy calculation are performed on the pixel matrix of the target region to obtain the time domain feature matrix, frequency domain feature matrix, cross-correlation matrix and entropy matrix; The pixel matrix, temporal feature matrix, frequency domain feature matrix, cross-correlation matrix, and entropy matrix are summarized to obtain the target matrix of the target region. The target matrix of each target region in each first base image is input into a pre-trained association model to obtain the association between each interference region and the water region in each first base image. The training method for the pre-trained association model includes: Collect multi-view image data samples of pet water fountains under different interference scenarios, and label the water area and interference area in the multi-view image data samples; Construct the pixel matrix and corresponding target matrix of multi-view image data samples; The true correlation between the interference area and the water body area in each multi-view image data sample is labeled as the sample label; Using the target matrix as input and the sample labels as output, train an initial neural network model, adjust the model parameters until convergence, and obtain the correlation model.

10. A water volume monitoring system for pet water fountains, applicable to the water volume monitoring method for pet water fountains as described in any one of claims 1-9, characterized in that, include: A water volume monitoring unit is used to acquire multi-view image data of various monitoring areas in a pet water fountain. The multi-view image data includes a top top view image and a side front view image. The top top view image is used to determine the top water volume coverage of the pet water fountain, and the side front view image is used to determine the front water volume coverage of the pet water fountain. The consumption simulation unit is used to determine the active state of water consumption in each monitoring area based on the temperature and humidity of the environment where the pet water fountain is located and the water coverage of all the water volume. By dynamically simulating the water consumption process of the pet water fountain through all the active consumption states, the water consumption trajectory of each monitoring area is obtained. The water consumption warning unit is used to merge and reconstruct all water consumption trajectories based on the location distance between each monitoring area to obtain a water consumption distribution map in the pet water fountain. The water consumption distribution map is used to determine multiple water consumption warning zones of the pet water fountain. The water consumption warning zones are used to characterize areas in each monitoring area of ​​the pet water fountain where the water consumption rate is higher than the average level. The consumption calculation unit is used to measure the dynamic water flow data of each water volume warning zone, and determine the water consumption flux of each water volume warning zone per unit time based on the velocity gradient between the dynamic water flow data. The consumption flux is used as a parameter value to characterize the amount of water consumed through the water volume warning zone per unit time. The water replenishment reminder unit is used to send a water replenishment reminder to the pet owner's mobile terminal when the water consumption exceeds the water consumption threshold.

Images