A control method and system for range hood
By building a hand and pot lid recognition model, combining the camera and infrared fill light, real-time identification of cooking status and adjusting the power of the range hood, the problem of inability to cope with the change in the volume of oil smoke under different cooking methods in the existing technology is solved, and the pumping is achieved Intelligent adjustment of range hood power and better user experience.
Patent Information
- Application Number
- CN202311835624.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-12-27
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2043-12-27
AI Technical Summary
The existing range hoods cannot adjust the working gear in time according to the cooking methods, resulting in poor user experience and unable to effectively deal with changes in the amount of oil smoke under different cooking methods.
By building a hand recognition model and a pot lid recognition model, combining the camera and infrared fill light, the status of the hand and pot lid during cooking is recognized in real time, the difference in the fume concentration is calculated to automatically adjust the motor speed of the range hood to achieve power Intelligent adjustment.
It improves the humanization of the control of the range hood, dynamically adjusts the power of the range hood according to cooking techniques, and enhances the user experience and equipment usage efficiency.
Smart Images

Figure CN117927981B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of range hoods, and in particular to a control method and system for range hoods. Background Art
[0002] It is a kitchen appliance that purifies the kitchen environment. It is installed above the kitchen stove and can quickly extract the waste from stove combustion and the oil smoke generated during cooking that is harmful to the human body and discharge it outdoors. At the same time, it condenses and collects the oil smoke to reduce pollution, purify the air, and has anti-toxic and explosion-proof safety protection functions. As people's requirements for cooking experience and kitchen environment become higher and higher, range hoods have become an indispensable device in the kitchen.
[0003] However, existing range hoods have certain limitations in terms of wind speed control. For example, during cooking, the range hood always absorbs oil smoke at one working level. However, the amount of oil smoke generated varies with different cooking methods. Therefore, existing range hoods cannot change the working level in time according to the cooking method, giving users a bad experience. Summary of the invention
[0004] In view of the above-mentioned defects, the purpose of the present invention is to provide a control method and system for a range hood, which can adjust the power of the range hood according to cooking methods and improve the humanization of the range hood control.
[0005] To achieve this object, the present invention adopts the following technical solution: A control method for a range hood comprises the following steps:
[0006] Build hand recognition model and pot lid recognition model;
[0007] A recognition area is established in the direction of the stove, and the hand recognition model and the pot lid recognition model are called to recognize the recognition area;
[0008] When the hand and the pot cover are recognized at the same time, a photo of the current stove is obtained, marked as the first photo, and the current fume concentration is obtained based on the first photo and marked as the first concentration; after the first threshold time, a photo of the current stove is obtained again, marked as the second photo, and the current fume concentration is obtained based on the second photo and marked as the second concentration;
[0009] The difference between the first concentration and the second concentration is obtained. If the difference is a positive number and is greater than the fractional threshold, the motor speed of the range hood is reduced. If the difference is a negative number and the absolute value of the difference is greater than the fractional threshold, the motor speed of the range hood is increased.
[0010] Preferably, if the rotation speed of the range hood is manually reduced after the hand is recognized and within the second time threshold, a photo of the current stove is obtained and marked as the third photo;
[0011] The current third photo is compared with the photos in the collection of suspicious pot lid photos to determine whether there are identical features between the current third photo and the photos in the collection of suspicious pot lid photos. If there are identical features, the current third photo is sent to the labeling personnel through the cloud. If the labeling personnel determines that there is a pot lid in the current third photo, the same technical features are input into the pot lid recognition model for training. If there are no identical features, the third photo is stored in the collection of suspicious pot lid photos.
[0012] Preferably, determining whether the current third photo and the photos in the collection of suspicious pot cover photos have the same features is specifically as follows:
[0013] Respectively identify the pot body in the third photo and the collection of suspicious photos of the pot cover, and obtain information of multiple photos inside the pot body;
[0014] Binarization is performed on the photo information to obtain a first characteristic photo of the third photo and a plurality of second characteristic photos of the collection of suspicious pot cover photos;
[0015] Using the first characteristic photo as a template and multiple second characteristic photos as input, and calculating the similarity between the first characteristic photo and the second characteristic photo by similarity;
[0016] The similarity calculation formula is as follows:
[0017] ε(Q, T, c) = ∑SSIM(o, r)(c+r), where Q is the input, T is the template, c is the image position where the template meets the binarization threshold, r∈P, P represents the area centered on c, r is the offset position, and o is the image position where the input meets the binarization threshold;
[0018] If the similarity is greater than the threshold, the third photo and the photos in the collection of suspicious pot cover photos have the same features.
[0019] Preferably, the method for obtaining the current oil smoke concentration of the second photo or the first photo is as follows:
[0020] Perform dynamic background modeling on the video stream of the stove area to obtain the stove model;
[0021] Calling the stove model to remove the stove from the first photo or the second photo to identify the stove; obtaining a modified identification photo;
[0022] The grayscale value of the identification photo is adjusted, and the smoke concentration is determined according to the grayscale value in the identification photo.
[0023] Preferably, when acquiring the image below the range hood, an infrared fill light is used to periodically fill light the captured image area; the video stream of the stove area is captured while being filled light by the infrared fill light.
[0024] A control system for a range hood, wherein a control method for the range hood includes: a modeling module, a region building module, a recognition module and a control module;
[0025] The modeling module is used to build a hand recognition model and a pot lid recognition model;
[0026] The area construction module is used to construct a recognition area in the direction of the stove, and call the hand recognition model and the pot lid recognition model to recognize the recognition area;
[0027] The recognition module is used for obtaining a photo of the current stove when the hand and the pot cover are recognized at the same time, marking it as a first photo, and obtaining a current oil smoke concentration according to the first photo, marking it as a first concentration;
[0028] After the first threshold time, a photo of the current stove is obtained again, marked as a second photo, and the current oil smoke concentration is obtained according to the second photo, marked as a second concentration;
[0029] The control module is used to obtain the difference between the first concentration and the second concentration. If the difference is a positive number and greater than a fractional threshold, the motor speed of the range hood is reduced; if the difference is a negative number and greater than the fractional threshold, the motor speed of the range hood is increased.
[0030] Preferably, it further comprises a supplementary recording module, wherein after the hand is recognized and the rotation speed of the range hood is manually reduced within the second time threshold, the supplementary recording module is used to obtain a photo of the current stove and mark it as the third photo;
[0031] The current third photo is compared with the photos in the collection of suspicious pot lid photos to determine whether there are identical features between the current third photo and the photos in the collection of suspicious pot lid photos. If there are identical features, the current third photo is sent to the labeling personnel through the cloud. If the labeling personnel determines that there is a pot lid in the current third photo, the same technical features are input into the pot lid recognition model for training. If there are no identical features, the third photo is stored in the collection of suspicious pot lid photos.
[0032] Preferably, the supplementary recording module further includes a judgment module;
[0033] The judgment module is used to respectively identify the pot body in the third photo and the collection of the suspicious pot cover photos, and obtain information of multiple photos inside the pot body;
[0034] Binarization is performed on the photo information to obtain a first characteristic photo of the third photo and a plurality of second characteristic photos of the collection of suspicious pot cover photos;
[0035] Using the first characteristic photo as a template and multiple second characteristic photos as input, and calculating the similarity between the first characteristic photo and the second characteristic photo by similarity;
[0036] If the similarity is greater than the threshold, the third photo and the photos in the collection of suspicious pot cover photos have the same features.
[0037] Preferably, the identification module includes a model submodule, a elimination submodule and a determination submodule;
[0038] The model submodule is used to perform dynamic background modeling on the video stream of the stove area to obtain a stove model;
[0039] The elimination submodule is used to call the stove model to eliminate stoves identified in the first photo or the second photo to obtain a modified identification photo;
[0040] The determination submodule is used to adjust the gray value of the identification photo and determine the smoke concentration according to the gray value in the identification photo.
[0041] Preferably, it also includes a fill light module, which uses an infrared fill light to periodically fill light the image collection area when the fill light module is used for the image below the range hood; the video stream of the stove area is collected by the infrared fill light.
[0042] One of the above technical solutions has the following advantages or beneficial effects: when the difference between the first concentration and the second concentration is a positive number and greater than the score threshold, the user is covering the pot lid at this time, and the stove fumes will suddenly decrease a lot, causing the first concentration to be greater than the second concentration. When the pot lid is covered, the power of the range hood can be reduced. When the difference between the first concentration and the second concentration is a negative number and greater than the score threshold, the user is picking up the pot lid at this time, and the stove fumes will suddenly increase a lot, causing the first concentration to be less than the second concentration. At this time, the power of the range hood can be increased. The intelligent power adjustment of the range hood is realized, and the humanization of the range hood is improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] Figure 1 is a flow chart of a method according to an embodiment of the present invention;
[0044] Figure 2 It is a schematic diagram of the system structure of an embodiment of the present invention. DETAILED DESCRIPTION
[0045] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals represent the same or similar elements or elements having the same or similar functions from beginning to end. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.
[0046] In the description of the embodiments of the present invention, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.
[0047] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, the meaning of "plurality" is two or more. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0048] like Figures 1-2 As shown, a control method for a range hood comprises the following steps:
[0049] Build hand recognition model and pot lid recognition model;
[0050] A recognition area is established in the direction of the stove, and the hand recognition model and the pot lid recognition model are called to recognize the recognition area;
[0051] When the hand and the pot cover are recognized at the same time, a photo of the current stove is obtained, marked as the first photo, and the current fume concentration is obtained based on the first photo and marked as the first concentration; after the first threshold time, a photo of the current stove is obtained again, marked as the second photo, and the current fume concentration is obtained based on the second photo and marked as the second concentration;
[0052] The difference between the first concentration and the second concentration is obtained. If the difference is a positive number and is greater than the fractional threshold, the motor speed of the range hood is reduced. If the difference is a negative number and the absolute value of the difference is greater than the fractional threshold, the motor speed of the range hood is increased.
[0053] When cooking, users may cook dishes by boiling or stewing. When cooking by these methods, the stove pot is usually covered with a pot lid. When the stove pot is covered, the oil smoke will be reduced by the pot lid. At this time, the power of the range hood needs to be appropriately reduced. When the pot lid is opened, the oil smoke will increase. At this time, the function of the range hood needs to be increased. In the existing technology, the power of the range hood is usually adjusted by identifying the temperature of the pot body, but the cooking method is not taken into consideration, resulting in a very inhuman adjustment effect.
[0054] In order to solve this problem, in the present invention, the user's hand information and the pot cover information used are first entered to construct a three-dimensional model of the hand model, and a hand recognition model and a pot cover recognition model are obtained respectively. A camera is provided on the range hood, and the camera is a fixed camera. The angle of the photo when acquiring the stove is fixed. For this purpose, a recognition area can be provided to only recognize the content of the recognition area to reduce unnecessary computing power consumption. The recognition area is an area containing the stove. When the range hood is used, when the hand recognition model and the pot cover recognition model recognize the pot cover and the hand at the same time, it means that the user may cover the pot cover or remove the pot cover. For this reason, when the hand and the pot cover are recognized, the first photo will be obtained; then after the threshold time, the second photo of the current stove is obtained again, wherein the threshold time is 2 to 4 seconds. The fume concentrations in the first photo and the second photo are obtained respectively through the first photo and the second photo, and are marked as the first concentration and the second concentration respectively.
[0055] When the difference between the first concentration and the second concentration is a positive number and greater than the score threshold, the user is covering the pot lid at this time, and the stove fumes will suddenly decrease a lot, causing the first concentration to be greater than the second concentration. When the pot lid is covered, the power of the range hood can be reduced. When the difference between the first concentration and the second concentration is a negative number and greater than the score threshold, the user is picking up the pot lid at this time, and the stove fumes will suddenly increase a lot, causing the first concentration to be less than the second concentration. At this time, the power of the range hood can be increased. The intelligent power adjustment of the range hood is realized, and the humanization of the range hood is improved.
[0056] Sometimes, the hand and the pot lid may be recognized at the same time in the recognition area, but the user does not use the pot lid, but may just move it. Therefore, when judging the concentration value, we cannot simply use the positive or negative difference between the first concentration and the second concentration to judge. We also need to add the score threshold judgment. Only when the difference between the two is greater than the score threshold, can we judge the change in the oil fume concentration.
[0057] It is worth mentioning that, in one embodiment of the present invention, the first photo and the second photo may be a collection of photos, i.e., multiple frames of photos acquired within a certain period of time, and the first concentration and the second concentration are the average values of the oil fume concentration in the multiple frames of photos. Because the change of oil fume concentration is relatively unstable, in order to avoid excessive changes in oil fume concentration at a single time point, thereby causing misjudgment, the average value of the oil fume concentration in multiple frames may be used as the first concentration or the second concentration.
[0058] Preferably, if the rotation speed of the range hood is manually reduced after the hand is recognized and within the second time threshold, a photo of the current stove is obtained and marked as the third photo;
[0059] The current third photo is compared with the photos in the collection of suspicious pot lid photos to determine whether there are identical features between the current third photo and the photos in the collection of suspicious pot lid photos. If there are identical features, the current third photo is sent to the labeling personnel through the cloud. If the labeling personnel determines that there is a pot lid in the current third photo, the same technical features are input into the pot lid recognition model for training. If there are no identical features, the third photo is stored in the collection of suspicious pot lid photos.
[0060] As the market updates products at a fast pace, the types and styles of pot covers will also change. Although the pot cover information has been entered before the range hood is used, the subsequent user may purchase a new pot cover, which may cause the pot cover recognition model to fail to recognize it. In order to improve the user experience.
[0061] After the hand is recognized and within the second time threshold, when the speed of the range hood is manually reduced, the user may use a new pot cover to cover the stove. When the cover is raised, the fume concentration will decrease. Since the pot cover cannot be recognized, the user performs manual adjustment. For this reason, in the present invention, when the above situation is encountered, a third photo will be obtained. At this time, the third photo will be compared with the photos in the collection of suspicious pot cover photos, and the collection of suspicious pot cover photos is the previous third photo. During the comparison process, it is determined whether the third photo and the photos in the collection of suspicious pot cover photos have the same features. If so, the third photo obtained this time is uploaded to the cloud. The marking personnel of the R&D party will observe the third picture. If there is a pot cover in the third picture, the marking personnel will input the same features and the pot cover in the third picture into the recognition model for training to achieve the update of pot cover recognition. If the same features do not exist, it may be a new pot lid, which is different from the previous pot lid, so the same features cannot be recognized. For this reason, the third photo is saved in the collection of suspicious pot lid photos. When the user uses the new pot lid to operate next time, the newly acquired third photo and the photos in the collection of suspicious pot lid photos can recognize the same features.
[0062] Preferably, determining whether the current third photo and the photos in the collection of suspicious pot cover photos have the same features is specifically as follows:
[0063] Respectively identify the pot body in the third photo and the collection of suspicious photos of the pot cover, and obtain information of multiple photos inside the pot body;
[0064] Binarization is performed on the photo information to obtain a first characteristic photo of the third photo and a plurality of second characteristic photos of the collection of suspicious pot cover photos;
[0065] Using the first characteristic photo as a template and multiple second characteristic photos as input, and calculating the similarity between the first characteristic photo and the second characteristic photo by similarity;
[0066] The similarity calculation formula is as follows:
[0067] ε(Q, T, c) = ∑SSIM(o, r)(c + r), where Q is the input, T is the template, c is the image position where the template meets the binarization threshold, r∈P, P represents the area centered on c, r is the offset position, and o is the image position where the input meets the binarization threshold;
[0068] If the similarity is greater than the threshold, the third photo and the photos in the collection of suspicious pot cover photos have the same features.
[0069] When cooking, users may use the same pot body but different pot lids. In order to avoid using the pot body as the same feature in the third picture and the collection of suspicious pot lid photos during identification, when identifying the same features, first identify the pot body in the third photo and the collection of suspicious pot lid photos, process the edge of the pot body by masking, then obtain the photo information inside the pot body, and then binarize the photo. At this time, the edge of the pot lid, the handle or the air outlet will turn white or increase its gray value due to binarization. Then calculate the similarity between the first feature photo and the second feature photo. In the similarity technology, the calculation is performed based on the image position that meets the binarization threshold, which can avoid extra computing power waste and improve recognition efficiency.
[0070] Preferably, the method for obtaining the current oil smoke concentration of the second photo or the first photo is as follows:
[0071] Perform dynamic background modeling on the video stream of the stove area to obtain the stove model;
[0072] Calling the stove model to remove the stove from the first photo or the second photo to identify the stove; obtaining a modified identification photo;
[0073] The grayscale value of the identification photo is adjusted, and the smoke concentration is determined according to the grayscale value in the identification photo.
[0074] Since the stove is used as the background for oil smoke recognition, it will affect the recognition result. Therefore, in the present invention, the stove model is constructed to remove the stoves in the first photo and the second photo. In the online recognition, the pot body is also recognized. Similarly, the pot body can be removed in this step to reduce the recognition amount. Then the gray value of the recognition photo is adjusted. The specific adjustment method can refer to the patent application number: CN108548199A. After the gray value is adjusted, different oil smoke concentrations can present different gray values. The oil smoke concentration can be defined according to the gray value. Specifically, when the oil smoke concentration is greater, its gray value in the recognition photo is greater.
[0075] Preferably, when acquiring the image below the range hood, an infrared fill light is used to periodically fill light the captured image area; the video stream of the stove area is captured while being filled light by the infrared fill light.
[0076] Infrared fill light can make the smoke feature more obvious and easier to detect visually. At the same time, during the recognition process of the pot and the stove, the captured video stream or the third picture can be kept at a certain brightness to avoid misjudgment in the subsequent binarization or gray value calculation.
[0077] A control system for a range hood, wherein a control method for the range hood includes: a modeling module, a region building module, a recognition module and a control module;
[0078] The modeling module is used to build a hand recognition model and a pot lid recognition model;
[0079] The area construction module is used to construct a recognition area in the direction of the stove, and call the hand recognition model and the pot lid recognition model to recognize the recognition area;
[0080] The recognition module is used for obtaining a photo of the current stove when the hand and the pot cover are recognized at the same time, marking it as a first photo, and obtaining a current oil smoke concentration according to the first photo, marking it as a first concentration;
[0081] After the first threshold time, a photo of the current stove is obtained again, marked as a second photo, and the current oil smoke concentration is obtained according to the second photo, marked as a second concentration;
[0082] The control module is used to obtain the difference between the first concentration and the second concentration. If the difference is a positive number and greater than a fractional threshold, the motor speed of the range hood is reduced; if the difference is a negative number and greater than the fractional threshold, the motor speed of the range hood is increased.
[0083] Preferably, it further comprises a supplementary recording module, wherein after the hand is recognized and the rotation speed of the range hood is manually reduced within the second time threshold, the supplementary recording module is used to obtain a photo of the current stove and mark it as the third photo;
[0084] The current third photo is compared with the photos in the collection of suspicious pot lid photos to determine whether there are identical features between the current third photo and the photos in the collection of suspicious pot lid photos. If there are identical features, the current third photo is sent to the labeling personnel through the cloud. If the labeling personnel determines that there is a pot lid in the current third photo, the same technical features are input into the pot lid recognition model for training. If there are no identical features, the third photo is stored in the collection of suspicious pot lid photos.
[0085] Preferably, the supplementary recording module further includes a judgment module;
[0086] The judgment module is used to respectively identify the pot body in the third photo and the collection of the suspicious pot cover photos, and obtain information of multiple photos inside the pot body;
[0087] Binarization is performed on the photo information to obtain a first characteristic photo of the third photo and a plurality of second characteristic photos of the collection of suspicious pot cover photos;
[0088] Using the first characteristic photo as a template and multiple second characteristic photos as input, and calculating the similarity between the first characteristic photo and the second characteristic photo by similarity;
[0089] If the similarity is greater than the threshold, the third photo and the photos in the collection of suspicious pot cover photos have the same features.
[0090] Preferably, the identification module includes a model submodule, a elimination submodule and a determination submodule;
[0091] The model submodule is used to perform dynamic background modeling on the video stream of the stove area to obtain a stove model;
[0092] The elimination submodule is used to call the stove model to eliminate stoves identified in the first photo or the second photo to obtain a modified identification photo;
[0093] The determination submodule is used to adjust the gray value of the identification photo and determine the smoke concentration according to the gray value in the identification photo.
[0094] Preferably, it also includes a fill light module, which uses an infrared fill light to periodically fill light the image collection area when the fill light module is used for the image below the range hood; the video stream of the stove area is collected by the infrared fill light.
[0095] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.
[0096] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.
Claims
1. A control method for a range hood, characterized in that: The following steps are involved: Build hand recognition model and pot lid recognition model; A recognition area is established in the direction of the stove, and the hand recognition model and the pot lid recognition model are called to recognize the recognition area; When the hand and the pot cover are recognized at the same time, a photo of the current stove is obtained, marked as the first photo, and the current fume concentration is obtained based on the first photo and marked as the first concentration; after the first threshold time, a photo of the current stove is obtained again, marked as the second photo, and the current fume concentration is obtained based on the second photo and marked as the second concentration; The difference between the first concentration and the second concentration is obtained. If the difference is a positive number and is greater than the fractional threshold, the motor speed of the range hood is reduced. If the difference is a negative number and the absolute value of the difference is greater than the fractional threshold, the motor speed of the range hood is increased.
2. The control method of a range hood according to claim 1, characterized in that: If the rotation speed of the range hood is manually reduced after the hand is recognized and within the second time threshold, a photo of the current stove is obtained and marked as the third photo; The current third photo is compared with the photos in the collection of suspicious pot lid photos to determine whether there are identical features between the current third photo and the photos in the collection of suspicious pot lid photos. If there are identical features, the current third photo is sent to the labeling personnel through the cloud. If the labeling personnel determines that there is a pot lid in the current third photo, the same technical features are input into the pot lid recognition model for training. If there are no identical features, the third photo is stored in the collection of suspicious pot lid photos.
3. The control method of a range hood according to claim 2, characterized in that: The specific steps for determining whether the third photo has the same features as the photos in the collection of suspicious pot lid photos are as follows: Respectively identify the pot body in the third photo and the collection of suspicious photos of the pot cover, and obtain information of multiple photos inside the pot body; Binarization is performed on the photo information to obtain a first characteristic photo of the third photo and a plurality of second characteristic photos of the collection of suspicious pot cover photos; Using the first characteristic photo as a template and multiple second characteristic photos as input, and calculating the similarity between the first characteristic photo and the second characteristic photo by similarity; The similarity calculation formula is as follows: ε(Q, T, c) = ∑SSIM(o, r)(c+r), where Q is the input, T is the template, c is the image position where the template meets the binarization threshold, r∈P, P represents the area centered on c, r is the offset position, and o is the image position where the input meets the binarization threshold; If the similarity is greater than the threshold, the third photo and the photos in the collection of suspicious pot cover photos have the same features.
4. The control method of a range hood according to claim 3, characterized in that: The method for obtaining the current oil smoke concentration of the second photo or the first photo is as follows: Perform dynamic background modeling on the video stream of the stove area to obtain the stove model; Calling the stove model to identify and remove stoves from the first photo or the second photo; Obtaining a modified identification photo; The grayscale value of the identification photo is adjusted, and the smoke concentration is determined according to the grayscale value in the identification photo.
5. The control method of a range hood according to claim 4, characterized in that: When acquiring the image below the range hood, an infrared fill light is used to perform phased fill light on the captured image area; the video stream of the stove area is collected while being filled with light by the infrared fill light.
6. A control system for a range hood, using the control method for a range hood according to any one of claims 1 to 5, characterized in that: include: Modeling module, region building module, identification module and control module; The modeling module is used to build a hand recognition model and a pot lid recognition model; The area construction module is used to construct a recognition area in the direction of the stove, and call the hand recognition model and the pot lid recognition model to recognize the recognition area; The recognition module is used for obtaining a photo of the current stove when the hand and the pot cover are recognized at the same time, marking it as a first photo, and obtaining a current oil smoke concentration according to the first photo, marking it as a first concentration; After the first threshold time, a photo of the current stove is obtained again, marked as a second photo, and the current oil smoke concentration is obtained according to the second photo, marked as a second concentration; The control module is used to obtain the difference between the first concentration and the second concentration. If the difference is a positive number and greater than a fractional threshold, the motor speed of the range hood is reduced; if the difference is a negative number and greater than the fractional threshold, the motor speed of the range hood is increased.
7. A range hood control system according to claim 6, characterized in that: The system further includes a supplementary recording module, wherein after the hand is recognized and the rotation speed of the range hood is manually reduced within a second time threshold, the supplementary recording module is used to obtain a photo of the current stove and mark it as a third photo; The current third photo is compared with the photos in the collection of suspicious pot lid photos to determine whether there are identical features between the current third photo and the photos in the collection of suspicious pot lid photos. If there are identical features, the current third photo is sent to the labeling personnel through the cloud. If the labeling personnel determines that there is a pot lid in the current third photo, the same technical features are input into the pot lid recognition model for training. If there are no identical features, the third photo is stored in the collection of suspicious pot lid photos.
8. The control system of a range hood according to claim 7, characterized in that: The supplementary recording module also includes a judgment module; The judgment module is used to respectively identify the pot body in the third photo and the collection of the suspicious pot cover photos, and obtain information of multiple photos inside the pot body; Binarization is performed on the photo information to obtain a first characteristic photo of the third photo and a plurality of second characteristic photos of the collection of suspicious pot cover photos; Using the first characteristic photo as a template and multiple second characteristic photos as input, and calculating the similarity between the first characteristic photo and the second characteristic photo by similarity; If the similarity is greater than the threshold, the third photo and the photos in the collection of suspicious pot cover photos have the same features.
9. The control system of a range hood according to claim 8, characterized in that: The identification module includes a model submodule, a elimination submodule and a determination submodule; The model submodule is used to perform dynamic background modeling on the video stream of the stove area to obtain a stove model; The elimination submodule is used to call the stove model to identify and eliminate stoves in the first photo or the second photo; Obtaining a modified identification photo; The determination submodule is used to adjust the gray value of the identification photo and determine the smoke concentration according to the gray value in the identification photo.
10. The control system of a range hood according to claim 8, characterized in that: It also includes a fill light module, which uses an infrared fill light to perform phased fill light on the image collection area when the image is below the range hood; the video stream of the stove area is collected by the infrared fill light.
