Camera dirty identification method and device of range hood, range hood and storage medium

Abstract

The present application relates to the technical field of kitchen appliances, and provides a camera dirt identification method and device for a range hood, a range hood and a storage medium, the camera dirt identification method for the range hood first acquires n smoke images collected by a camera in n sampling periods, determines a respective dirt level corresponding to each of the n smoke images, then acquires a first quantity of a first target dirt level representing the presence of dirt in each dirt level, records the number of times of dirt based on the first quantity, and finally determines whether the camera is dirty according to the recorded number of times of dirt. According to the above method, by collecting multiple groups of images, the data is more complete, which is beneficial to the accuracy and reliability of the results. The number of times of dirt based on statistics is used to determine whether the camera is dirty, rather than a single number of times of dirt, which is more objective and can improve the anti-interference ability, reduce misidentification, effectively improve the identification accuracy, and better improve the user experience.

Description

Technical Field

[0001] This invention relates to the field of kitchen appliance technology, and in particular to a method, device, range hood, and storage medium for identifying dirt and grime using a camera on a range hood. Background Technology

[0002] Kitchen appliances such as range hoods and cooktops are essential appliances in modern kitchens, providing convenience and assistance for users' cooking. The range hood is an important component of the range hood and cooktop, responsible for removing cooking fumes from the kitchen.

[0003] With the diversification of cooking methods, the amount of oil fumes produced during cooking varies. When the amount of oil fumes is high, people often need to manually operate a button to increase the fume level to prevent the fumes from escaping. When the amount of oil fumes is low, they need to manually operate the button to decrease the level to avoid wasting energy. This method has a low level of automation and a relatively poor user experience.

[0004] In light of this, some range hoods have added intelligent modules for visually recognizing the concentration of cooking fumes. These modules use cameras to monitor changes in cooking fumes and adjust the fan speed in real time. However, kitchens are constantly exposed to harsh environments with high temperatures, high levels of cooking fumes, and high humidity. During long-term operation, grease and grime easily adhere to the camera lens. As usage time increases, the lens becomes increasingly dirty, directly causing blurred images and image distortion. This severely interferes with the accuracy of fume recognition. Although there are methods in the industry to identify and detect the dirtiness of camera lenses, most existing methods rely on a single, independent image for dirt determination. This single, fixed method is easily affected by environmental factors and cannot accurately determine the true degree of dirtiness. Ultimately, this still leads to problems such as inaccurate fume concentration recognition and delayed or inaccurate fan speed control, significantly reducing the operational stability and actual effectiveness of intelligent fume control systems. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art.

[0006] Therefore, one objective of this invention is to propose a method for identifying dirt in a range hood's camera. This method involves continuously acquiring images of cooking fumes collected over multiple sampling periods. First, it determines the dirt level corresponding to each image. Then, it obtains the number of first target dirt levels representing the presence of dirt from the obtained dirt levels. Next, it records the number of dirt occurrences based on the obtained first number. Finally, it determines whether the camera is dirty based on the recorded number of dirt occurrences and promptly reminds the user to wipe the lens. Thus, this method for identifying dirt in a range hood's camera, by acquiring multiple sets of images, provides more comprehensive data, which improves the accuracy and reliability of the results and ensures the accuracy of subsequent oil fume identification by the range hood. In addition, this method uses the cumulative dirt frequency from multiple historical statistics as the core criterion for identifying the dirt status, abandoning the traditional single-discrimination mode that relies solely on a single instantaneous dirt data for identification. Based on comprehensive judgment using multiple statistical data, the judgment criteria are made more rigorous and objective. This effectively reduces the interference caused by external uncertainties such as ambient light fluctuations and moisture obstruction, significantly reducing the occurrence of abnormal identifications such as misjudgment of dirt and grime, further enhancing anti-interference capabilities, reducing misidentification, effectively improving recognition accuracy, and better enhancing the user experience.

[0007] Therefore, a second objective of the present invention is to provide a camera dirt recognition device for a range hood.

[0008] Therefore, the third objective of this invention is to provide a smoke hood.

[0009] Therefore, a fourth object of the present invention is to provide a computer-readable storage medium.

[0010] To achieve the above objectives, a first aspect of the present invention provides a method for identifying dirt in a range hood camera. The camera is used to collect images of oil fumes generated by a cooktop. The method includes the following steps: during the operation of the cooktop, acquiring n oil fume images collected by the camera in n sampling periods; determining the dirt level corresponding to each of the n oil fume images; acquiring a first number of first target dirt levels representing the presence of dirt in each dirt level; recording the number of dirt occurrences based on the first number; and determining whether the camera is dirty based on the recorded number of dirt occurrences.

[0011] According to the present invention, the method for identifying dirt in a range hood camera first acquires n images of oil fumes collected by the camera in n sampling cycles during the operation of the stove. Then, it acquires the dirt level corresponding to each of the n oil fume images. Next, it acquires a first number of first target dirt levels representing the presence of dirt based on each dirt level. Then, it records the number of times dirt occurs based on the first number. Finally, it determines whether the camera is dirty based on the recorded number of times dirt occurs. By acquiring multiple sets of images, the data becomes more comprehensive, which is beneficial to the accuracy and reliability of the results and ensures the accuracy of subsequent oil fume identification by the range hood. In addition, the method uses the cumulative dirt frequency of multiple sets of historical statistics as the core judgment basis to identify the dirt status, abandoning the traditional single judgment mode that relies solely on a single instant of dirt data to complete the identification. Based on comprehensive judgment using multiple statistical data, the judgment criteria are made more rigorous and objective. This effectively reduces the interference caused by external uncertainties such as ambient light fluctuations and moisture obstruction, significantly reducing abnormal identification situations such as misjudgment and incorrect judgment due to dirt and grime. It further enhances anti-interference capabilities, reduces misidentification, and effectively improves recognition accuracy. It also prevents the continuous increase in lens dirt from directly causing blurred visual acquisition images and image distortion, which seriously interferes with the accuracy of oil fume recognition and judgment, leading to problems such as lag, misjudgment, and inaccurate adjustment of range hood settings, thus better improving the user experience.

[0012] In some embodiments, determining whether the camera is dirty based on the number of times the dirt is recorded includes: determining that the camera is dirty if the cumulative number of times the dirt is recorded reaches or exceeds a first preset threshold; or, determining that the camera is dirty if the number of times the dirt is recorded within a preset time reaches or exceeds a second preset threshold.

[0013] In some embodiments, recording the number of times dirt occurs based on the first quantity includes: calculating a first dirt coverage rate to characterize the frequency of dirt occurrence based on the first quantity and the number n of the oil fume images; and recording the number of times dirt occurs once when the first dirt coverage rate is greater than or equal to a first preset coverage threshold.

[0014] In some embodiments, the method for identifying dirt in a range hood camera further includes: when the range hood is turned on, acquiring m images of oil fumes collected by the camera in m sampling cycles; determining the dirt level corresponding to each of the m oil fume images; acquiring a second number of each dirt level that represents a degree of dirt reaching a second target dirt level; and determining whether the camera is dirty based on the second number.

[0015] In some embodiments, determining whether the camera is dirty based on the second quantity includes: calculating a second dirt coverage rate to characterize the frequency of dirt occurrence based on the second quantity and the number m of the oil fume images; and determining that the camera is dirty when the second dirt coverage rate is greater than or equal to a second preset coverage threshold.

[0016] In some embodiments, the method for identifying dirt in the camera of a range hood, after determining that the camera is dirty, further includes: controlling an alarm device to issue an alarm message indicating that the camera is dirty; and / or controlling the fan of the range hood to run at the current fan speed until the recorded number of times the camera is dirty is cleared.

[0017] In some embodiments, determining the oil fume level corresponding to each of the n oil fume images includes: determining the dirt level corresponding to each oil fume image based on a pre-calibrated two-dimensional mapping relationship between oil fume images and dirt levels.

[0018] To achieve the above objectives, a second aspect of the present invention provides a dirt recognition device for a range hood camera. The camera is used to collect images of oil fumes generated by a cooktop. The dirt recognition device for the range hood camera includes: a first acquisition module, used to acquire n oil fume images collected by the camera in n sampling cycles during the operation of the cooktop; a determination module, used to determine the dirt level corresponding to each of the n oil fume images; a second acquisition module, used to acquire a first number of first target dirt levels representing the presence of dirt in each dirt level; a recording module, used to record the number of times dirt is collected based on the first number; and a processing module, used to determine whether the camera is dirty based on the recorded number of times dirt is collected.

[0019] The range hood camera dirt recognition device according to an embodiment of the present invention is used to implement the range hood camera dirt recognition method of the above embodiment of the present invention. During the operation of the stove, n oil fume images collected by the camera in n sampling cycles are first acquired, and then the dirt level corresponding to each of the n oil fume images is acquired. Next, based on each dirt level, a first number of first target dirt levels representing the presence of dirt is acquired, and then the number of times dirt is recorded based on the first number. Finally, the camera is judged to be dirty based on the recorded number of times dirt is collected. In this way, by acquiring multiple sets of images, the data is more substantial, which is conducive to the accuracy and reliability of the results to ensure the accuracy of subsequent range hood oil fume recognition. In addition, the method uses the cumulative dirt frequency of multiple sets of historical statistics as the core judgment basis to carry out the identification of dirt status, abandoning the traditional single discrimination mode that only relies on a single instant of dirt data to complete the identification. Based on comprehensive judgment using multiple statistical data, the judgment criteria are made more rigorous and objective. This effectively reduces the interference caused by external uncertainties such as ambient light fluctuations and moisture obstruction, significantly reducing abnormal identification situations such as misjudgment and incorrect judgment due to dirt and grime. It further enhances anti-interference capabilities, reduces misidentification, and effectively improves recognition accuracy. It also prevents the continuous increase in lens dirt from directly causing blurred visual acquisition images and image distortion, which seriously interferes with the accuracy of oil fume recognition and judgment, leading to problems such as lag, misjudgment, and inaccurate adjustment of range hood settings, thus better improving the user experience.

[0020] To achieve the above objectives, a third aspect of the present invention provides a range hood, comprising: a range hood body; a camera disposed on the range hood body for capturing images of oil fumes generated by a cooktop; and a camera dirt recognition device for the range hood as described in the above embodiments of the present invention; or, a controller for executing a camera dirt recognition method for the range hood as described in the above embodiments of the present invention.

[0021] According to an embodiment of the present invention, a range hood includes a range hood body; a camera disposed on the range hood body for capturing images of oil fumes generated by a cooktop; and a camera dirt recognition device for the range hood as described in the above embodiment of the present invention; or, a controller, the controller being used to execute a camera dirt recognition method for the range hood as described in the above embodiment of the present invention. During cooktop operation, the controller first acquires n oil fume images captured by the camera in n sampling cycles, then acquires the dirt level corresponding to each of the n oil fume images, then acquires a first quantity representing a first target dirt level based on each dirt level, then records the number of times dirt is detected based on the first quantity, and finally determines whether the camera is dirty based on the recorded number of times dirt is detected. This method, by acquiring multiple sets of images, makes the data more comprehensive, improving the accuracy and reliability of the results to ensure the accuracy of subsequent oil fume recognition by the range hood. Furthermore, this method uses the cumulative dirt frequency from multiple historical statistics as the core criterion for identifying the dirt state, abandoning the traditional single-discrimination mode that relies solely on a single instant of dirt data for identification. Based on comprehensive judgment using multiple statistical data, the judgment criteria are made more rigorous and objective. This effectively reduces the interference caused by external uncertainties such as ambient light fluctuations and moisture obstruction, significantly reducing abnormal identification situations such as misjudgment and incorrect judgment due to dirt and grime. It further enhances anti-interference capabilities, reduces misidentification, and effectively improves recognition accuracy. It also prevents the continuous increase in lens dirt from directly causing blurred visual acquisition images and image distortion, which seriously interferes with the accuracy of oil fume recognition and judgment, leading to problems such as lag, misjudgment, and inaccurate adjustment of range hood settings, thus better improving the user experience.

[0022] To achieve the above objectives, a fourth aspect of the present invention provides a computer-readable storage medium storing a dirt detection program for a range hood camera, wherein the dirt detection program for the range hood camera, when executed by a processor, implements the dirt detection method for the range hood camera as described in the above embodiments of the present invention.

[0023] According to an embodiment of the present invention, a computer-readable storage medium stores a dirt recognition program for a range hood camera. When executed by a processor, the dirt recognition program implements the dirt recognition method for a range hood camera as described in the above embodiment of the present invention. During the operation of the stove, n oil fume images collected by the camera in n sampling cycles are first acquired. Then, the dirt level corresponding to each of the n oil fume images is acquired. Next, a first quantity of a first target dirt level representing the presence of dirt is acquired based on each dirt level. Then, the number of times dirt is detected is recorded based on the first quantity. Finally, the camera is judged to be dirty based on the recorded number of times dirt is detected. By acquiring multiple sets of images, the data is made more substantial, which is beneficial to the accuracy and reliability of the results to ensure the accuracy of subsequent oil fume recognition by the range hood. In addition, the method uses the cumulative dirt frequency of multiple sets of historical statistics as the core judgment basis to identify the dirt status, abandoning the traditional single judgment mode that relies only on a single instant of dirt data to complete the identification. Based on comprehensive judgment using multiple statistical data, the judgment criteria are made more rigorous and objective. This effectively reduces the interference caused by external uncertainties such as ambient light fluctuations and moisture obstruction, significantly reducing abnormal identification situations such as misjudgment and incorrect judgment due to dirt and grime. It further enhances anti-interference capabilities, reduces misidentification, and effectively improves recognition accuracy. It also prevents the continuous increase in lens dirt from directly causing blurred visual acquisition images and image distortion, which seriously interferes with the accuracy of oil fume recognition and judgment, leading to problems such as lag, misjudgment, and inaccurate adjustment of range hood settings, thus better improving the user experience.

[0024] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0025] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a flowchart of a method for identifying dirt in a range hood camera according to an embodiment of the present invention; Figure 2 This is a flowchart of a method for identifying dirt in a range hood camera according to another embodiment of the present invention; Figure 3 This is a flowchart of a method for identifying dirt and grime in a range hood camera according to a specific embodiment of the present invention. Figure 4 This is a flowchart of a power-on detection algorithm for lens contamination according to a specific embodiment of the present invention; Figure 5 This is a flowchart of a cooking process detection lens dirt algorithm according to a specific embodiment of the present invention; Figure 6This is a schematic diagram of the structure of a dirt recognition device for a range hood camera according to an embodiment of the present invention. Figure 7 This is a schematic diagram of the structure of a range hood according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the structure of a range hood according to another embodiment of the present invention.

[0026] Figure label: 1000-Dirty identification device for the range hood camera; 1100-First acquisition module; 1200-Determination module; 1300-Second acquisition module; 1400-Recording module; 1500-Processing module; 2000-Rain hood; 2100-Rain hood body; 2200-Camera; 2300-Controller. Detailed Implementation

[0027] The embodiments of the present invention are described in detail below. The embodiments described with reference to the accompanying drawings are exemplary. The embodiments of the present invention are described in detail below.

[0028] The following is for reference. Figures 1-8 This invention describes a method, apparatus, and computer-readable storage medium for identifying dirt in a range hood using a camera, according to embodiments of the present invention.

[0029] In one embodiment of the present invention, a method for identifying dirt in a range hood using a camera is provided. The range hood includes a main body and a camera, wherein the camera is used to capture images of oil fumes generated by the cooktop.

[0030] In a specific embodiment, the main body of the smoke hood includes a housing, and the housing is equipped with components such as a fan, which includes a motor and a fan wheel.

[0031] In a specific embodiment, a camera for image acquisition is fixedly installed on the side of the range hood facing the stove, ensuring that the lens is directly facing the stove area and that all daily cooking scenarios can be completely captured. After the camera is powered on, it continuously and periodically acquires real-time images of the kitchen, including the oil fume scene, and synchronously transmits the acquired image data to the main control chip. Based on the identified lens smudges or obstructions, the range hood provides corresponding feedback, thus automatically detecting and determining the oil smudge obstruction status of the range hood camera using a real-time oil fume image acquisition method.

[0032] Figure 1 This is a flowchart of a method for identifying dirt in a range hood camera according to an embodiment of the present invention, as shown below. Figure 1 As shown, the dirt detection method of the range hood's camera includes: Step S1: During the operation of the stove, acquire n images of cooking fumes collected by the camera in n sampling periods. In a specific embodiment, a fixed image sampling period is set, and the stove is continuously captured by a high-definition camera at preset time intervals. Multiple frames of oil fume images corresponding to multiple sampling periods are collected in sequence. The collected multiple oil fume images are summarized and stored in a unified manner to provide complete time-series image data for subsequent lens dirt level determination.

[0033] Specifically, the dirt recognition method of the range hood's camera first sets a uniform image sampling time interval as the sampling period. When the stove is in the cooking state, the camera continuously captures images of the cooking area. After each sampling period, a frame of oil fume image is stored. The image acquisition work is completed in sequence for n consecutive sampling periods, and finally a group of n oil fume images is obtained. This ensures that the acquired images can completely restore the camera's shooting images at different times, providing complete time-series image data for subsequent lens dirt level determination.

[0034] Step S2: Determine the level of dirtiness corresponding to each of the n oil fume images.

[0035] Specifically, the dirt recognition method of this range hood uses a pre-trained dirt level recognition model. This model is trained using a large number of oil fume images. By inputting the acquired oil fume images, it can output the corresponding dirt level. After collecting n sets of oil fume images over n cycles, the model is detected and calibrated to classify corresponding dirt levels, such as: no dirt, light dirt, moderate dirt, and heavy dirt. This accurately determines the actual dirt level generated under the current cooking conditions, providing reliable data values ​​for subsequent calculations.

[0036] Step S3: Obtain the first number of the first target dirt level that represents the presence of dirt in each dirt level.

[0037] In a specific embodiment, based on the various dirt levels corresponding to the n sets of detected oil fume images, the main control chip calculates and determines the first number of target dirt levels representing the presence of dirt in each dirt level through a model. It should be noted that "various dirt levels" here refers to the result obtained after the camera sequentially acquires n oil fume images within n sampling periods, and the pre-trained dirt level recognition model judges each image. That is, each oil fume image corresponds to one dirt level, and the two form a one-to-one correspondence in quantity. For example, if 10 oil fume images are acquired, the pre-trained dirt level recognition model will output 10 corresponding dirt levels.

[0038] Specifically, the dirt recognition method of the range hood camera reduces the image data volume by statistically analyzing the first number of the first target dirt level that represents dirt in each dirt level, eliminating dirt-free data that does not affect visual recognition, and only filtering the first target dirt level and counting the first number, thereby improving the accuracy and computational efficiency of dirt recognition.

[0039] Step S4: Record the number of times the item gets dirty based on the first count.

[0040] In a specific embodiment, after completing the screening of the first target dirt level, the first number of people reaching the first target dirt level is counted, and then the first number value is calculated as a percentage. The calculated value is judged, and when the calculated value reaches a preset threshold, it is recorded as one dirt count.

[0041] Specifically, the dirt recognition method of this range hood's camera, after completing the data screening for the first target dirt level, uses the main control chip to count the number of samples in the screening results that meet the requirements of the first target dirt level, obtaining the corresponding first quantity. Then, using all statistical samples as the overall base, a percentage calculation is performed on this first quantity to obtain the corresponding percentage value. The calculated percentage value is then compared with a pre-stored judgment threshold within the device. When the percentage value meets and reaches the preset threshold condition, the system automatically completes the cumulative recording of one dirt recognition count. By calculating the percentage based on the overall sample, abandoning single-data judgment, the judgment result is more objective and reliable, avoiding misjudgments caused by instantaneous abnormal data. Simultaneously, accumulating and recording the number of dirt recognitions allows for continuous tracking of the camera's dirt accumulation trend, facilitating precise control of cleaning timing.

[0042] Step S5: Determine if the camera is dirty based on the number of times it gets dirty.

[0043] In a specific embodiment, the real-time calculated percentage value is compared with the judgment threshold preset in the device. When the percentage value is detected to reach or exceed the preset threshold condition, the system automatically completes the cumulative recording of the number of dirt recognitions. Then, the main control chip judges the number of reported dirt. When the number of reported dirt exceeds the preset threshold condition, the camera is judged to be dirty. If it does not exceed the threshold, the number of reported times continues to be accumulated.

[0044] Specifically, the dirt detection method for the range hood's camera records the number of times dirt accumulates and calculates the cumulative dirt percentage. Only when a preset threshold is reached or exceeded is the camera lens considered dirty. In other words, the system only alerts the user to a dirty lens on the display screen after a certain number of dirt reports, avoiding frequent reporting that could negatively impact user experience. Furthermore, setting a fixed threshold as a judgment criterion standard unifies the evaluation criteria and ensures consistent and stable program operation. Thus, this method, by collecting multiple sets of images, provides more comprehensive data, improving the accuracy and reliability of the results. It determines dirt based on the statistical number of dirt occurrences, rather than a single occurrence, resulting in a more objective judgment benchmark, improved anti-interference capabilities, reduced false recognition, and enhanced recognition accuracy, ultimately improving the user experience. Specifically, the dirt recognition method for the range hood's camera first acquires n images of oil fumes collected by the camera in n sampling cycles during the operation of the stove. Then, it acquires the dirt level corresponding to each of the n oil fume images. Next, it acquires a first number of first target dirt levels representing the presence of dirt based on each dirt level. Then, it records the number of times dirt occurs based on the first number. Finally, it determines whether the camera is dirty based on the recorded number of times dirt occurs. By acquiring multiple sets of images, the data becomes more comprehensive, which is beneficial to the accuracy and reliability of the results and ensures the accuracy of subsequent oil fume recognition by the range hood. In addition, this method uses the cumulative dirt frequency of multiple sets of historical statistics as the core judgment criterion to identify the dirt status, abandoning the traditional single judgment mode that relies solely on a single instant of dirt data to complete the recognition. Based on comprehensive judgment using multiple statistical data, the judgment criteria are made more rigorous and objective. This effectively reduces the interference caused by external uncertainties such as ambient light fluctuations and moisture obstruction, significantly reducing abnormal identification situations such as misjudgment and incorrect judgment due to dirt and grime. It further enhances anti-interference capabilities, reduces misidentification, and effectively improves recognition accuracy. It also prevents the continuous increase in lens dirt from directly causing blurred visual acquisition images and image distortion, which seriously interferes with the accuracy of oil fume recognition and judgment, leading to problems such as lag, misjudgment, and inaccurate adjustment of range hood settings, thus better improving the user experience.

[0045] In one embodiment of the present invention, step S5 determines whether the camera is dirty based on the number of times the dirt is recorded, including: if the cumulative number of times the dirt is recorded reaches or exceeds a first preset threshold, the camera is determined to be dirty; or, if the number of times the dirt is recorded within a preset time reaches or exceeds a second preset threshold, the camera is determined to be dirty.

[0046] In a specific embodiment, the main control chip compares the real-time calculated percentage of the first target level of dirt with a pre-stored threshold value inside the device. When the percentage of the first target level of dirt reaches or exceeds the preset threshold, the system automatically records the cumulative number of dirt identifications. Then, it records and judges the number of reported dirt. When the cumulative number of reported dirt exceeds the first preset threshold, the camera is determined to be dirty. If it does not exceed the threshold, the number of reported dirt continues to accumulate until the threshold is reached. Alternatively, if people walk in front of the stove or kitchen utensils move and briefly block the camera, generating momentarily invalid data, the system can determine that the camera is dirty when the number of reported dirt reaches or exceeds the second preset threshold within a preset time period.

[0047] It should be noted that the first preset number threshold is used to determine whether the dirt accumulation has reached the level that should trigger an alarm. Its setting is based on the statistical relationship between the cumulative number of dirt events and the probability of false alarms, selecting a critical value. For example, setting the first preset number threshold to 3 times ensures that the cumulative number can reliably reach this value in real dirt scenarios, but is difficult to reach under occasional interference, thus balancing detection sensitivity and anti-interference capability. The preset time is set based on the temporal distribution characteristics of dirt events, selecting a continuous monitoring duration that can effectively capture persistent pollution behavior. For example, setting the preset time to 30 seconds requires that this duration can effectively capture persistent pollution behavior while avoiding random interference. The second preset number threshold is set based on the fact that this threshold should be higher than the average number of random interference events within its preset time, while being lower than the number of dirt events in real dirt scenarios. The second preset number threshold is, for example, 3 times, to distinguish between occasional interference and real pollution. The above parameters are set according to the oil fume pollution pattern, sensor characteristics, and anti-interference requirements to adapt to and achieve the technical effects of this invention.

[0048] Furthermore, the dirt recognition method of this range hood's camera can directly determine whether the dirt has reached or exceeded a first preset threshold number of times, or determine whether the dirt has reached or exceeded a second preset threshold number of times within a fixed statistical period. The judgment logic can be flexibly switched according to the actual kitchen environment. When there is little environmental interference and changes in grease are readily apparent, direct cumulative judgment is used for rapid identification. In complex scenarios with fluctuating lighting and where moisture and debris can easily cause temporary data anomalies, time-based frequency judgment is used. Relying on valid data within a time period to verify the results, this method can adapt to different cooking needs, streamline the judgment process, reduce computational redundancy, and flexibly balance recognition efficiency and anti-interference capabilities while ensuring recognition accuracy.

[0049] In one embodiment of the present invention, recording the number of times dirt occurs based on a first quantity includes: calculating a first dirt coverage rate to characterize the frequency of dirt occurrence based on the first quantity and the number n of oil fume images; and recording the number of times dirt occurs once when the first dirt coverage rate is greater than or equal to a first preset coverage rate threshold.

[0050] In a specific embodiment, by pre-programming different levels of dirtiness corresponding to oil fume images into the device, the real-time acquired oil fume images are compared and matched with each level of dirtiness. Then, the first number 'a' of the first target dirtiness level representing the presence of dirt is counted among the n dirtiness levels. The first target dirtiness level can include: heavy dirtiness, moderate dirtiness, and light dirtiness, i.e., the first number 'a' of these three dirtiness levels is counted. Based on the ratio of the counted first number 'a' to the number of oil fume images 'n', a first dirtiness coverage rate is calculated to represent the frequency of dirtiness occurrence. Here, the number of oil fume images used as the denominator is essentially the number of dirt level determination results. Specifically, the camera sequentially acquires n oil fume images within n sampling periods. After determining the dirt level of each oil fume image, a corresponding dirt level can be obtained, resulting in a total of n dirt level determination results. Therefore, each acquired oil fume image corresponds to one dirt level, and the number of oil fume images and the number of dirt level determination results are exactly equal, both being the same value n. Thus, the dirt level is classified into four categories during the dirt model determination. A threshold value representing the ratio of moderate to severe dirt levels to the total number of samples is used as a condition to determine whether the lens is dirty, improving the accuracy of the determination and minimizing the impact of random outputs on the dirt level determination.

[0051] Furthermore, after calculating the first dirt coverage rate used to characterize the frequency of dirt occurrence, it is determined whether the coverage rate is greater than or equal to a first preset coverage threshold. If the threshold is reached or exceeded, it is recorded as one instance of dirtiness. For example, if there are 3 heavily dirty images, 5 moderately dirty images, 8 lightly dirty images, and 10 undirty images, then the first number 'a' representing the presence of dirt in the first target level among the 26 oil fume images is 16, and 'n' is 26. The proportion is calculated to obtain the first dirtiness coverage rate used to represent the frequency of dirtiness: equal If the first preset coverage threshold is The system records a dirt event as a single occurrence when the coverage rate exceeds a first preset coverage threshold. This first preset coverage threshold distinguishes between valid dirt events and occasional interference. Its value should ensure that the first dirt coverage rate caused by random or transient interference is below this threshold, while the first dirt coverage rate in a real dirt scenario reaches or exceeds this threshold. In practical applications, this first preset coverage threshold can be set according to the sensor's detection accuracy and anti-interference requirements. For example, the first preset coverage threshold can be set to 15% to achieve the technical effect of accurately recording the number of dirt events in this invention. Therefore, the lens dirt detection output is calculated using an additive statistical method. A cleaning reminder is triggered only when the system has cumulatively identified a dirt state reaching a preset number of occurrences. This effectively avoids frequent false reminders caused by transient environmental interference, reduces irrelevant prompts, ensures the accuracy and validity of dirt detection results, and significantly improves the user's daily experience.

[0052] Figure 2 This is a flowchart of a method for identifying dirt in a range hood using a camera, according to another embodiment of the present invention. Figure 2As shown, in one embodiment of the present invention, a method for identifying dirt in a camera based on a range hood includes: when the range hood is turned on, acquiring m images of oil fumes collected by the camera in m sampling cycles; determining the dirt level corresponding to each of the m oil fume images; acquiring a second number representing the degree of dirt reaching a second target dirt level in each dirt level; and determining whether the camera is dirty based on the second number.

[0053] In a specific embodiment, the dirt recognition method for the range hood's camera first captures images of the cooking area and stores the collected oil fume images according to a set image sampling time interval when the range hood is turned on. This process is repeated for m consecutive sampling cycles, capturing m images to ensure that the captured images can completely reproduce the camera's view at startup, providing complete temporal image data for subsequent lens dirt level determination. Then, a pre-trained dirt level recognition model is used. This model is trained on a large number of oil fume images. By inputting the acquired oil fume images, the corresponding dirt level can be output. The m sets of oil fume images obtained after m cycles of acquisition are detected and calibrated to classify corresponding dirt levels, such as no dirt, light dirt, moderate dirt, and heavy dirt, accurately determining the dirt level of the camera lens at startup and providing reliable data values ​​for subsequent calculations.

[0054] Furthermore, after determining the dirt level corresponding to each of the m oil fume images, the second number of images in each dirt level that represent the degree of dirtiness reaching the second target dirt level is counted. The degree of dirtiness that meets the second target dirt level includes: moderate and heavy dirtiness. The second number of images that reach the second target dirt level is counted as b. This reduces the image data volume, removes no-dirt data and light-dirt data that do not affect visual recognition, and only filters the second target dirt level and counts the second number.

[0055] Further, after completing the data screening for the second target dirt level, the number of samples meeting the requirements of the second target dirt level is counted, resulting in the corresponding second quantity b. Then, using all statistical samples m as the overall base, a proportion calculation is performed on this second quantity b to obtain the corresponding proportion value. The real-time calculated proportion value is compared with a pre-stored judgment threshold inside the range hood. When the proportion value meets and reaches the preset threshold condition, the camera lens is determined to be dirty, and the user is reminded to clean the lens. It should be noted that if the scenarios of "range hood being turned on" and "cooktop running" overlap in time, the range hood's camera dirt recognition method will prioritize the start-up detection logic. After the start-up detection is completed, it will automatically switch to the cooking process detection logic. Both scenarios use the same dirt status counter and flag, and will not repeatedly trigger alarms or repeatedly execute gear control actions due to different scenario names. That is, the range hood always outputs a unique control command based on the current dirt detection result and the real-time operating status (whether cooking is in progress), and there is no situation where two independent logics conflict in parallel. By using the above methods, the range hood can automatically perform a self-check for lens dirt upon startup, and proactively identify and prevent issues such as lens grease blockage. This effectively prevents problems such as decreased accuracy in fume recognition and inaccurate speed adjustment judgment caused by lens dirt after the user enables the fume recognition function, thus achieving early warning of faults and ensuring the stable and reliable operation of the fume recognition and control function from the source.

[0056] In one embodiment of the present invention, determining whether the camera is dirty based on a second quantity includes: calculating a second dirt coverage rate to characterize the frequency of dirt occurrence based on the second quantity and the number m of oil fume images; and determining that the camera is dirty when the second dirt coverage rate is greater than or equal to a second preset coverage rate threshold.

[0057] In a specific embodiment, a second quantity b representing the degree of dirtiness reaching a second target dirtiness level is determined based on the statistical analysis of m oil fume images corresponding to various dirtiness levels. The second target dirtiness level may include: heavy dirtiness and moderate dirtiness, i.e., the second quantity b representing these two dirtiness levels is statistically analyzed. The second dirtiness coverage rate, representing the frequency of dirtiness occurrence, is obtained by calculating the ratio of the statistically analyzed second quantity b to the number of oil fume images m. Here, the number of oil fume images used as the denominator is essentially the number of dirt level determination results. Specifically, the camera sequentially acquires m oil fume images within m sampling periods. After determining the dirt level of each oil fume image, a corresponding dirt level for that image can be obtained, resulting in a total of m dirt level determination results. Therefore, each acquired oil fume image corresponds to one dirt level, and the number of oil fume images and the number of dirt level determination results are exactly equal, both being the same value m. Thus, the dirt level is classified into four categories during the dirt model determination. A threshold value representing the ratio of moderate to severe dirt levels to the total number of samples is used as a condition to determine whether the lens is dirty, improving the accuracy of the determination and minimizing the impact of random outputs on the dirt level assessment.

[0058] Furthermore, after calculating the second dirt coverage rate used to characterize the frequency of dirt occurrence, it is determined whether this coverage rate is greater than or equal to a second preset coverage threshold. If the threshold is reached or exceeded, the camera is determined to be dirty, and the user is promptly reminded to clean the camera. This second preset coverage threshold is used to determine whether the dirt level has reached the critical level required to determine camera dirtiness. Its value should ensure that the second dirt coverage rate under slightly or critically polluted conditions is below the threshold, while the second dirt coverage rate under significantly polluted conditions reaches or exceeds the threshold. This second preset coverage threshold can be set according to alarm sensitivity requirements, false alarm tolerance, etc. For example, a second preset coverage threshold of 30% can be used to adapt to the technical effect of accurately determining whether a camera is dirty. For example, if there are 3 heavily dirty images, 5 moderately dirty images, 8 lightly dirty images, and 10 undirty images, then the second number b representing the second target level of dirtiness in the dirt level determination results corresponding to the 26 oil fume images is 8, and n is 26. The proportion is calculated to obtain the second dirt coverage rate used to represent the frequency of dirt occurrence. equal If the second preset coverage threshold is If the coverage rate exceeds a second preset coverage threshold, the camera is deemed dirty, and the user is reminded to clean it. By using overall data as a benchmark, the selected values ​​are statistically analyzed and their proportions are calculated, resulting in a more objective judgment and avoiding misjudgments based on a single value.

[0059] In one embodiment of the present invention, after determining that the camera is dirty, the method further includes: controlling the alarm device to issue an alarm message indicating that the camera is dirty; and / or controlling the fan of the range hood to run at the current fan speed until the recorded number of times the camera is dirty is cleared.

[0060] Specifically, when the range hood detects a dirty lens upon startup, the alarm control device will display an alarm message indicating the lens's dirtiness, such as a screen flash or a text message, reminding the user to clean the lens. Furthermore, during cooking, to prevent blurred or distorted images caused by lens dirt, which could lead to inaccurate smoke detection and inconsistent fan speeds, the controller will keep the range hood's fan at the current speed. Simultaneously, the alarm control device will display an alarm message indicating the lens's dirtiness until the user cleans the lens. For example, the user can manually wipe the lens to clean it. After wiping, the user can manually confirm cleaning through the user interface, triggering a reset command; alternatively, the range hood can automatically reset the count when the system detects a second consecutive instance of dirt coverage below a preset threshold for a preset duration, indicating the lens is now clean and automatically resetting the accumulated dirt count to zero.

[0061] Furthermore, by locking the current speed setting, the current smoke extraction power can be stabilized to ensure the smoke extraction effect, while avoiding situations such as speed jumping and sudden changes in speed caused by recognition errors. This ensures that the range hood operates more smoothly, avoids frequent changes in air volume that affect the user experience, and also prevents problems such as increased power consumption and unstable smoke extraction caused by incorrect speed adjustment.

[0062] In one embodiment of the present invention, determining the oil fume level corresponding to each of the n oil fume images includes: determining the dirt level corresponding to each oil fume image based on a pre-calibrated two-dimensional mapping relationship between oil fume images and dirt levels.

[0063] In a specific embodiment, the oil fume images and dirt levels are pre-calibrated, with each oil fume image assigned a corresponding dirt level. Alternatively, a pre-trained dirt level recognition model can be used, trained on a large number of oil fume images. Inputting an acquired oil fume image will output the corresponding dirt level. During the offline calibration phase, features are extracted and labeled from oil fume images with known dirt levels. During actual detection, the same features are extracted from real-time acquired images, and the closest dirt level is automatically matched by comparison with the calibration sample library. This establishes a two-dimensional mapping relationship between various feature parameters of the oil fume image and different dirt levels of the lens. It should be noted that the "two-dimensional" in this mapping relationship does not refer to a geometric two-dimensional space, but rather to two dimensions: the first dimension: various feature parameters of the oil fume image; the second dimension: the dirt level of the lens (e.g., discrete or continuous values ​​such as level 1, 2, 3, 4, etc.). This two-dimensional mapping relationship clearly defines the dirt level corresponding to each type of image imaging state. When the range hood is working normally, the camera captures images of cooking fumes in real time and compares them with a pre-calibrated two-dimensional mapping relationship to quickly identify the actual level of dirt on the lens corresponding to the currently captured image of cooking fumes, thus completing the basic level determination.

[0064] Furthermore, by leveraging a pre-calibrated two-dimensional mapping and association system, a significant amount of cumbersome real-time computation can be eliminated. Simply substituting the real-time image into the established corresponding rules allows for efficient classification and determination of the level of dirt. This method unifies the judgment criteria and evaluation standards throughout the process, is unaffected by external temporary environmental factors, effectively improves the overall speed and efficiency of lens dirt level identification, ensures consistency in multiple test results, and makes the dirt level determination results more closely reflect the actual oil and dirt adhesion on the lens.

[0065] The following describes the method for identifying dirt in the camera of a range hood according to the above embodiments of the present invention with reference to specific embodiments. In this specific embodiment, a method for identifying dirt in the camera of a range hood is provided, and the method for identifying dirt in the camera of a range hood is implemented through the range hood.

[0066] Figure 3 This is a control flowchart of a method for identifying dirt in a range hood camera according to a specific embodiment of the present invention, as shown below. Figure 3 As shown in this specific embodiment, the method for identifying dirt and grime using the camera of the range hood includes the following steps: S10: Initialization; S20: AI Function Activation S30: The camera is not installed or is installed incorrectly; S40: Screen flashing or buzzer sound indicates to the user that the camera is not installed or is not installed correctly; S50: Camera fault diagnosis (whether communication is normal); S60: The screen flashes to display fault codes to alert the user; S70: Camera detected as being obstructed by a baffle; S80: Screen flashes to remind the user to turn on the camera; S90: Camera lens is dirty or blurry; S100: Screen flashes to remind the user to wipe the screen; S110: The range hood operates using AI functional logic; S120: Detects current lens dirt.

[0067] In this specific embodiment, the control flow of the dirt recognition method for the range hood's camera is described as follows: First, proceed to S10, then execute S20. If S30 is true, execute S40; otherwise, execute S50. If S50 is true, execute S60; otherwise, execute S70. If S70 is true, execute S80; otherwise, execute S90. If S90 is true, execute S100; otherwise, execute S110. If S120 is true, execute S100.

[0068] Figure 4 This is a flowchart of a lens contamination detection algorithm based on a specific embodiment of the present invention, such as... Figure 4 As shown in this specific embodiment, step S90, which involves powering on and detecting whether the lens is dirty, includes: S91: Initialization; S92: Camera captures current data. Images of each sampling period , , ..., , ; S93: Detection of collected samples Lens dirt level corresponding to each sampling period of images , , ..., , The levels are divided into four categories (S - heavily soiled; A - moderately soiled; B - lightly soiled; C - no soiling). S94: Count the number of S and A. ; S95: Judgment If yes, proceed to step S96; otherwise, proceed to step S92. S96: The camera lens is determined to be dirty.

[0069] Figure 5 This is a flowchart of a cooking process lens dirt detection algorithm according to a specific embodiment of the present invention, such as... Figure 5As shown, in this specific embodiment, step S120, detecting current lens contamination, includes: S121: Initialization; S122: Camera captures current data. Images of each sampling period , , ..., , ; S123: Detection of collected samples Lens dirt level corresponding to each sampling period of images , , ..., , The levels are divided into four categories (S - heavily soiled; A - moderately soiled; B - lightly soiled; C - no soiling). S124: Count the total number of S, A, and B. ; S125: Judgment If yes, proceed to step S126; otherwise, proceed to step S123. S126: Record the number of times dirt is reported ; S127: Determine the number of reported dirt incidents ; S128: The camera lens is determined to be dirty; S129: The range hood is operating at the current fan speed setting; S1210: The controller controls the screen of the range hood to flash to remind the user that the lens is dirty.

[0070] In this specific embodiment, the dirt detection method of the range hood's camera has the following advantages compared to the prior art: (1) The dirt detection method of the camera of the range hood will detect whether the lens is dirty when the range hood is turned on, so as to avoid the situation that the oil fume recognition accuracy is too low due to the lens being dirty after the user turns on the oil fume recognition function, and has a pre-warning function. (2) The dirt recognition method of the camera of the range hood also detects the degree of dirt on the lens during the cooking process and performs cumulative algorithm processing on the output value of dirt. That is, the system will only remind the user of the lens dirt on the display screen after the system recognizes and reports dirt a certain number of times (obtained from the experiment), thus avoiding frequent reporting of dirt and affecting the user experience. (3) When the lens of the range hood is detected to be dirty during the cooking process, the range hood maintains the current fan speed setting, thus avoiding the fan speed from jumping back and forth due to inaccurate oil fume recognition caused by lens dirt. (4) The dirt identification method of the camera of the smoke machine classifies the dirt level into four categories when judging the dirt model. The method of judging whether the lens is dirty adopts the threshold of the ratio of the number of medium and severe dirt to the total number of samples as a condition, which reduces the impact of random output on dirt judgment to a greater extent. (5) The dirt recognition method of the camera of the smoke machine removes the non-dirty data that does not affect visual recognition when calculating the dirt coverage rate, and only filters the target dirt level and counts the number, thereby improving the accuracy and computational efficiency of dirt recognition. (6) The dirt recognition method of the camera of the range hood relies on the overall sample to calculate the proportion, abandons the judgment of single data, and the judgment result is more objective and reliable; (7) The dirt recognition method of the camera of the range hood accumulates and records the number of dirt recognitions, and can continuously track the dirt accumulation trend of the camera, and accurately control the cleaning time.

[0071] In summary, the method for identifying dirt in a range hood camera according to the embodiments of the present invention first acquires n images of oil fumes collected by the camera in n sampling cycles during the operation of the stove, then acquires the dirt level corresponding to each of the n oil fume images, then acquires a first number of first target dirt levels representing the presence of dirt based on each dirt level, then records the number of times dirt occurs based on the first number, and finally determines whether the camera is dirty based on the recorded number of times dirt occurs. By acquiring multiple sets of images, the data is more comprehensive, which is beneficial to the accuracy and reliability of the results to ensure the accuracy of subsequent oil fume identification by the range hood. In addition, this method uses the cumulative dirt frequency of multiple sets of historical statistics as the core judgment basis to identify the dirt status, abandoning the traditional single judgment mode that relies only on a single instant of dirt data to complete the identification. Based on comprehensive judgment using multiple statistical data, the judgment criteria are made more rigorous and objective. This effectively reduces the interference caused by external uncertainties such as ambient light fluctuations and moisture obstruction, significantly reducing abnormal identification situations such as misjudgment and incorrect judgment due to dirt and grime. It further enhances anti-interference capabilities, reduces misidentification, and effectively improves recognition accuracy. It also prevents the continuous increase in lens dirt from directly causing blurred visual acquisition images and image distortion, which seriously interferes with the accuracy of oil fume recognition and judgment, leading to problems such as lag, misjudgment, and inaccurate adjustment of range hood settings, thus better improving the user experience.

[0072] In one embodiment of the present invention, a dirt identification device 1000 for a range hood camera is also provided for a range hood 2000, the range hood 2000 including a camera disposed on the range hood body.

[0073] Figure 6 This is a schematic diagram of the structure of a dirt detection device 1000 for a range hood camera according to an embodiment of the present invention. Figure 6As shown, in one embodiment of the present invention, the camera dirt recognition device 1000 of the range hood includes: a first acquisition module 1100, used to acquire n oil fume images collected by the camera in n sampling cycles during the operation of the stove; a determination module 1200, used to determine the dirt level corresponding to each of the n oil fume images; a second acquisition module 1300, used to acquire a first number of first target dirt levels representing the presence of dirt in each dirt level; a recording module 1400, used to record the number of times dirt is collected based on the first number; and a processing module 1500, used to determine whether the camera is dirty based on the recorded number of times dirt is collected.

[0074] In one embodiment of the present invention, the processing module 1500 is used to determine whether the camera is dirty based on the number of times the dirt is recorded, specifically including: if the cumulative number of times the dirt is recorded reaches or exceeds a first preset number threshold, the camera is determined to be dirty; or, if the number of times the dirt is recorded within a preset time reaches or exceeds a second preset number threshold, the camera is determined to be dirty.

[0075] In one embodiment of the present invention, the recording module 1400 is used to record the number of times dirt occurs based on a first quantity, specifically including: calculating a first dirt coverage rate to characterize the frequency of dirt occurrence based on the first quantity and the number n of oil fume images; when the first dirt coverage rate is greater than or equal to a first preset coverage threshold, the number of times dirt occurs is recorded.

[0076] In one embodiment of the present invention, the first acquisition module 1100 is further configured to: acquire m oil fume images collected by the camera in m sampling cycles when the range hood is turned on; the determination module 1200 is further configured to: determine the dirt level corresponding to each of the m oil fume images; the second acquisition module 1300 is further configured to: acquire a second number of each dirt level that represents the degree of dirt reaching the second target dirt level; and the processing module 1500 is further configured to: determine whether the camera is dirty based on the second number.

[0077] In one embodiment of the present invention, the processing module 1500 is used to determine whether the camera is dirty based on a second quantity, specifically including: calculating a second dirt coverage rate to characterize the frequency of dirt occurrence based on the second quantity and the number m of oil fume images; when the second dirt coverage rate is greater than or equal to a second preset coverage threshold, the camera is determined to be dirty.

[0078] In one embodiment of the present invention, after determining that the camera is dirty, the processing module 1500 further includes: controlling the alarm device to issue an alarm prompt message indicating that the camera is dirty; and / or controlling the fan of the smoke hood to run at the current fan speed until the recorded number of times the camera is dirty is cleared.

[0079] In one embodiment of the present invention, the determining module 1200 is used to determine the oil fume level corresponding to each of the n oil fume images, specifically including: determining the dirt level corresponding to each oil fume image based on a pre-calibrated two-dimensional mapping relationship between oil fume images and dirt levels.

[0080] It should be noted that the specific implementation of the dirt recognition device 1000 for the range hood camera in this embodiment of the invention is similar to the specific implementation described in the dirt recognition method for the range hood camera in the above embodiment of the invention. For details, please refer to the description in the section on dirt recognition method for the range hood camera. To reduce redundancy, it will not be repeated here.

[0081] Specifically, the range hood camera dirt recognition device 1000 according to an embodiment of the present invention is used to implement the range hood camera dirt recognition method of the above embodiment of the present invention. During the operation of the stove, it first acquires n oil fume images collected by the camera in n sampling cycles, then acquires the dirt level corresponding to each of the n oil fume images, then acquires a first number of first target dirt levels representing the presence of dirt based on each dirt level, then records the number of times dirt is detected based on the first number, and finally determines whether the camera is dirty based on the recorded number of times dirt is detected. In this way, by acquiring multiple sets of images, the data is more substantial, which is beneficial to the accuracy and reliability of the results to ensure the accuracy of subsequent range hood oil fume recognition. In addition, the method uses the cumulative dirt frequency of multiple sets of historical statistics as the core judgment basis to carry out the identification of dirt status, abandoning the traditional single discrimination mode that only relies on a single instant of dirt data to complete the identification. Based on comprehensive judgment using multiple statistical data, the judgment criteria are made more rigorous and objective. This effectively reduces the interference caused by external uncertainties such as ambient light fluctuations and moisture obstruction, significantly reducing abnormal identification situations such as misjudgment and incorrect judgment due to dirt and grime. It further enhances anti-interference capabilities, reduces misidentification, and effectively improves recognition accuracy. It also prevents the continuous increase in lens dirt from directly causing blurred visual acquisition images and image distortion, which seriously interferes with the accuracy of oil fume recognition and judgment, leading to problems such as lag, misjudgment, and inaccurate adjustment of range hood settings, thus better improving the user experience.

[0082] In one embodiment of the present invention, a range hood 2000 is also provided.

[0083] Figure 7 This is a schematic diagram of the structure of a range hood according to an embodiment of the present invention. Figure 7 As shown, in one embodiment of the present invention, the range hood 2000 includes: a range hood body 2100; a camera 2200 disposed on the range hood body 2100 for capturing images of oil fumes generated by the stove; and a camera dirt recognition device 1000 for the range hood as described in the above embodiment of the present invention.

[0084] Figure 8This is a schematic diagram of the structure of a range hood according to another embodiment of the present invention. Figure 8 As shown, in another embodiment of the present invention, the range hood 2000 includes: a range hood body 2100; a camera 2200 disposed on the range hood body 2100 for capturing images of oil fumes generated by the stove; and a controller 2300 for executing the camera dirt recognition method of the range hood as described in the above embodiment of the present invention.

[0085] It should be noted that the specific implementation method of the range hood 2000 in this embodiment of the invention for camera dirt recognition is similar to the specific implementation method described in the above embodiment of the invention for camera dirt recognition. For details, please refer to the description of the range hood camera dirt recognition method section. To reduce redundancy, it will not be repeated here.

[0086] Furthermore, other components and functions of the range hood 2000 according to the above embodiments of the present invention are known to those skilled in the art, and will not be described in detail in order to reduce redundancy.

[0087] Specifically, the range hood 2000 according to an embodiment of the present invention includes: a range hood body 2100; a camera 2200, disposed on the range hood body 2100, for capturing images of oil fumes generated by the cooktop; and a camera dirt recognition device 1000 for the range hood as described in the above embodiment of the present invention; or, a controller 2300, the controller 2300 being used to execute the camera dirt recognition method for the range hood as described in the above embodiment of the present invention, wherein during the operation of the cooktop, n oil fume images captured by the camera in n sampling periods are first acquired, and then the dirt corresponding to each of the n oil fume images is acquired. The method first assesses the level of dirtiness, then obtains the first number of the first target dirtiness level representing the presence of dirtiness based on each dirtiness level, records the number of times dirtiness occurs based on the first number, and finally determines whether the camera is dirty based on the recorded number of dirtiness occurrences. By collecting multiple sets of images, the data becomes more comprehensive, which is beneficial to the accuracy and reliability of the results and ensures the accuracy of subsequent range hood recognition of oil fumes. In addition, this method uses the cumulative dirtiness frequency of multiple sets of historical statistics as the core judgment criterion to identify the dirtiness status, abandoning the traditional single discrimination mode that relies solely on a single instant of dirtiness data to complete the identification and judgment. Based on comprehensive judgment using multiple statistical data, the judgment criteria are made more rigorous and objective. This effectively reduces the interference caused by external uncertainties such as ambient light fluctuations and moisture obstruction, significantly reducing abnormal identification situations such as misjudgment and incorrect judgment due to dirt and grime. It further enhances anti-interference capabilities, reduces misidentification, and effectively improves recognition accuracy. It also prevents the continuous increase in lens dirt from directly causing blurred visual acquisition images and image distortion, which seriously interferes with the accuracy of oil fume recognition and judgment, leading to problems such as lag, misjudgment, and inaccurate adjustment of range hood settings, thus better improving the user experience.

[0088] In one embodiment of the present invention, a computer-readable storage medium is also provided.

[0089] In one embodiment of the present invention, a dirt detection program for a range hood camera is stored on a computer-readable storage medium. When the dirt detection program for a range hood camera is executed by a processor, it implements the dirt detection method for a range hood camera as described in the above embodiment of the present invention.

[0090] It should be noted that when the computer-readable storage medium of the present invention executes the dirt recognition program of the range hood camera through the processor, its specific implementation is similar to the specific implementation described in the above-described method of the range hood camera dirt recognition method. For details, please refer to the description of the method of the range hood camera dirt recognition. To reduce redundancy, it will not be repeated here.

[0091] Specifically, according to embodiments of the present invention, a dirt recognition program for a range hood camera is stored on a computer-readable storage medium. When the dirt recognition program for a range hood camera is executed by a processor, it implements the dirt recognition method for a range hood camera as described in the above embodiments of the present invention. During the operation of the stove, n oil fume images collected by the camera in n sampling cycles are first acquired. Then, the dirt level corresponding to each of the n oil fume images is acquired. Next, based on each dirt level, a first number of first target dirt levels representing the presence of dirt is acquired. Then, the number of times dirt is recorded based on the first number. Finally, the camera is judged to be dirty based on the recorded number of times dirt is collected. In this way, by acquiring multiple sets of images, the data is made more substantial, which is conducive to the accuracy and reliability of the results to ensure the accuracy of subsequent oil fume recognition by the range hood. In addition, the method uses the cumulative dirt frequency of multiple sets of historical statistics as the core judgment basis to carry out the identification of dirt status, abandoning the traditional single discrimination mode that relies only on a single instant of dirt data to complete the identification. Based on comprehensive judgment using multiple statistical data, the judgment criteria are made more rigorous and objective. This effectively reduces the interference caused by external uncertainties such as ambient light fluctuations and moisture obstruction, significantly reducing abnormal identification situations such as misjudgment and incorrect judgment due to dirt and grime. It further enhances anti-interference capabilities, reduces misidentification, and effectively improves recognition accuracy. It also prevents the continuous increase in lens dirt from directly causing blurred visual acquisition images and image distortion, which seriously interferes with the accuracy of oil fume recognition and judgment, leading to problems such as lag, misjudgment, and inaccurate adjustment of range hood settings, thus better improving the user experience.

[0092] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example.

[0093] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for identifying dirt and grime in a range hood's camera, characterized in that, The camera is used to capture images of oil fumes produced by the stove, and the method for identifying dirt and grime in the range hood's camera includes: During the operation of the stove, n images of cooking fumes are acquired by the camera in n sampling cycles; Determine the level of dirtiness corresponding to each of the n images of oil fume; Obtain the first number of the first target dirt level that represents the presence of dirt among each of the dirt levels; Based on the first quantity and the number n of the oil fume images, a first dirt coverage rate is calculated to characterize the frequency of dirt occurrence; when the first dirt coverage rate is greater than or equal to a first preset coverage threshold, a dirt occurrence is recorded; wherein, the first preset coverage threshold is used to distinguish between real dirt scenes and instantaneous interference scenes: in the instantaneous interference scene, the first dirt coverage rate is less than the first preset coverage threshold; in the real dirt scene, the first dirt coverage rate is greater than or equal to the first preset coverage threshold. Determining whether the camera is dirty based on the recorded number of times it gets dirty includes: determining that the camera is dirty if the cumulative number of times it gets dirty reaches or exceeds a first preset threshold; or, determining that the camera is dirty if the number of times it gets dirty within a preset time period reaches or exceeds a second preset threshold. The method further includes: when the range hood is turned on, acquiring m oil fume images collected by the camera in m sampling cycles; determining the dirt level corresponding to each of the m oil fume images; acquiring a second number of each dirt level that represents the degree of dirt reaching a second target dirt level; and determining whether the camera is dirty based on the second number.

2. The method for identifying dirt and grime in a range hood using a camera according to claim 1, characterized in that, The step of determining whether the camera is dirty based on the second quantity includes: A second dirt coverage rate, used to characterize the frequency of dirt occurrence, is calculated based on the second quantity and the number m of the oil fume images. When the second dirt coverage rate is greater than or equal to the second preset coverage threshold, the camera is determined to be dirty.

3. The method for identifying dirt and grime in a range hood using a camera according to claim 1 or 2, characterized in that, After determining that the camera is dirty, the following steps are also included: The control alarm device issues an alarm message indicating that the camera is dirty; and / or Control the fan of the range hood to run at the current fan speed setting until the recorded number of times the dust has accumulated is cleared.

4. The method for identifying dirt and grime in a range hood using a camera according to claim 1, characterized in that, Determining the dirt level corresponding to each of the n oil fume images includes: The dirt level corresponding to each oil fume image is determined based on the pre-calibrated two-dimensional mapping relationship between oil fume images and dirt levels.

5. A dirt detection device for a range hood camera, characterized in that, The camera is used to capture images of the oil fumes produced by the stove, and the dirt recognition device of the range hood camera includes: The first acquisition module is used to acquire n oil fume images collected by the camera in n sampling cycles during the operation of the stove; The determination module is used to determine the dirt level corresponding to each of the n oil fume images; The second acquisition module is used to acquire the first number of the first target dirt level that represents the presence of dirt in each of the dirt levels; The recording module is used to calculate a first dirt coverage rate to characterize the frequency of dirt occurrence based on the first quantity and the number n of the oil fume images; when the first dirt coverage rate is greater than or equal to a first preset coverage threshold, the number of dirt occurrences is recorded; wherein, the first preset coverage threshold is used to distinguish between real dirt scenes and instantaneous interference scenes: in the instantaneous interference scene, the first dirt coverage rate is less than the first preset coverage threshold; in the real dirt scene, the first dirt coverage rate is greater than or equal to the first preset coverage threshold. The processing module is used to determine whether the camera is dirty based on the number of times the dirt is recorded, including: determining that the camera is dirty if the cumulative number of times the dirt is recorded reaches or exceeds a first preset threshold; or, determining that the camera is dirty if the number of times the dirt is recorded within a preset time reaches or exceeds a second preset threshold. The first acquisition module is further configured to: acquire m oil fume images collected by the camera in m sampling cycles when the range hood is turned on; the determination module is further configured to: determine the dirt level corresponding to each of the m oil fume images; the second acquisition module is further configured to: acquire a second number of each dirt level that represents a degree of dirt reaching a second target dirt level; the processing module is further configured to: determine whether the camera is dirty based on the second number.

6. A range hood, characterized in that, include: Range hood body; A camera, mounted on the range hood body, is used to capture images of the oil fumes produced by the stove. The camera dirt recognition device for a range hood as described in claim 5.

7. A range hood, characterized in that, include: A controller for performing the camera dirt recognition method for a range hood as described in any one of claims 1-4.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a dirt detection program for the camera of the range hood, which, when executed by a processor, implements the dirt detection method for the camera of the range hood as described in any one of claims 1-4.

Images