阅读说明：本技术 不涉及个人隐私的人体姿态检测传感器及照明控制方法 (Human body posture detection sensor and lighting control method without personal privacy ) 是由 刘国良 巩晓雅 刘宏波 于 2021-08-16 设计创作，主要内容包括：本发明公开了一种不涉及个人隐私的人体姿态检测传感器及照明控制方法,其中传感器包括设于芯片内的图像传感单元、图像处理单元以及编码输出单元；所述图像传感单元用于将采集到的图像传输给所述图像处理单元,所述图像处理单元通过深度学习识别模型推理分析出图像中的场景信息和人体姿态信息；所述编码输出单元用于根据场景信息和人体姿态信息设定相应的控制编码进行输出。本发明不能输出图像数据,当用来控制照明设备时,对于某些隐私的场景和动作,如性爱动作、裸体姿态可以划分为与其它不涉及个人隐私动作的信息所属的分类类型中,从而从根本上避免了输出用户的隐私,但同时可以满足相应的照明效果。(The invention discloses a human posture detection sensor and an illumination control method which do not relate to personal privacy, wherein the sensor comprises an image sensing unit, an image processing unit and a coding output unit which are arranged in a chip; the image sensing unit is used for transmitting the acquired image to the image processing unit, and the image processing unit infers and analyzes scene information and human body posture information in the image through a deep learning recognition model; and the code output unit is used for setting corresponding control codes according to the scene information and the human body posture information and outputting the control codes. The invention can not output image data, when used for controlling the lighting equipment, for some privacy scenes and actions, such as sexual love actions and naked body gestures, can be divided into classification types which belong to other information which does not relate to personal privacy actions, thereby fundamentally avoiding the privacy of an output user, but simultaneously meeting the corresponding lighting effect.)

1. The human body posture detection sensor which does not relate to personal privacy is characterized by comprising an image sensing unit, an image processing unit and a coding output unit which are arranged in a chip; the image sensing unit is used for transmitting the acquired image to the image processing unit, and the image processing unit infers and analyzes scene information and human body posture information in the image through a deep learning recognition model; and the code output unit is used for setting corresponding control codes according to the scene information and the human body posture information and outputting the control codes.

2. The human body posture detection sensor not related to personal privacy of claim 1, wherein the image sensing unit comprises an array module, a row selection module and a signal processing module;

the array module is used for irradiating the pixel array by external light to generate a photoelectric effect and generate corresponding charges in the pixel unit;

the row selection module is used for gating the corresponding row pixel units;

and the signal processing module is used for converting the image signal into a digital image signal and outputting the digital image signal.

3. The human body posture detection sensor not related to personal privacy of claim 1, wherein the image processing unit comprises an acquisition module, a data processing module and a learning module;

the acquisition module is used for acquiring an image as a data set;

the data processing module is used for processing the data set to obtain a deep learning data set;

the learning module is used for creating a deep learning model, and training, verifying and testing the deep learning model by using the deep learning data set to obtain a deep learning identification model.

4. The personal pose detection sensor not related to personal privacy of claim 3, wherein the images captured by the capture module comprise images of human poses and gestures of a scene.

5. The human body posture detection sensor not related to personal privacy as claimed in claim 3, wherein the data processing module comprises a classification labeling sub-module and a data set dividing sub-module;

the classification and labeling submodule is used for classifying and labeling the data set;

the data set dividing submodule is used for dividing the data set into a training data set, a verification data set and a test data set according to the ratio of 6:2: 2.

6. The personal pose detection sensor not related to personal privacy of claim 3, wherein the learning module comprises a training sub-module, a verification sub-module and a testing sub-module;

the training submodule is used for training the deep learning model by utilizing a training data set;

the verification submodule is used for verifying the deep learning model by utilizing a verification data set;

and the testing submodule is used for testing the deep learning model by utilizing a testing data set so as to obtain an identification model.

7. The lighting control method of the human body posture detection sensor not related to the personal privacy, which is electrically connected with the lighting device, according to any one of claims 1 to 6, wherein the method comprises:

the human body posture detection sensor which does not relate to personal privacy collects scene information and human body posture information in space in real time;

carrying out reasoning analysis on the acquired scene information and human body posture information in the space by using the recognition model, and classifying;

generating a control code for controlling the lighting device according to the classification condition;

and the lighting equipment performs light adjustment according to the control code.

8. The lighting control method of the human body posture detection sensor not related to the personal privacy as claimed in claim 7, wherein in the step of performing inference analysis on the collected scene information and the human body posture information in the space by using the recognition model and classifying, when the information related to the personal privacy action is included, the information is classified into a classification type to which other information not related to the personal privacy action belongs.

9. The lighting control method of the human body posture detection sensor not related to the privacy of the person as claimed in claim 8, wherein the actions related to the privacy of the person include a sex action and a nude posture.

Technical Field

The invention relates to the field of intelligent sensing, in particular to a human body posture detection sensor and an illumination control method which do not relate to personal privacy.

Background

In an intelligent home system, intelligent sensing is a very important link. The intelligent home system comprises a plurality of aspects, such as environment perception, personnel perception and equipment state perception, and whether the states and parameters can be accurately perceived determines the performance index of the intelligent home system. At present, environmental perception, such as temperature and humidity, noise, illumination intensity and other parameters are accurately perceived, in addition, various civil and industrial communication networks are more and more stable at present, and the perception of the equipment state can be basically obtained in real time. In the aspect of personnel perception, the existing equipment and technical means can well perceive whether personnel are out of the field or whether the personnel move, but the distance from the intelligent perception is far away, the intelligent home system can accurately judge the states of the personnel, such as reading, eating, drinking tea, resting, sleeping and the like, only by realizing the accurate perception of scenes and human postures, and the states of light and other related equipment are adjusted according to the states of the personnel, such as reading the books by brighter light with higher color temperature, drinking the tea by soft light with lower color temperature, resting by dimming the light, sleeping by turning off the light and properly increasing the temperature and the like.

The current scene recognition based on the artificial intelligence method is widely applied to mobile phone photographing and has a good recognition effect. The human posture recognition sensor is divided into a wearable type and a non-wearable type, the wearable type sensor refers to a sensor carried by a human body, such as a gyroscope, an acceleration sensor and the like, and the sensors are not suitable for smart home occasions. The non-wearable sensors generally refer to millimeter wave radars, infrared sensors and image sensors, and the millimeter wave radars and the infrared sensors can basically and accurately detect whether people exist, but cannot identify accurate human body postures and gestures. The image sensor can accurately detect human body gestures, gestures and the like, but the image sensor can generate image output and is used for an intelligent home scene to enable a user to worry about privacy disclosure.

No exception is found in the specific commercially available image sensors that have image output interfaces, such as MIPI, DVP, SPI, etc., through which the user can read out the image for processing or display. However, the sensor capable of reading out the image at will poses a great threat to the privacy of the end user, some developers upload the image to their own server or public cloud for improving products or other purposes, and the transmitted user image data, like a timing bomb, can be illegally acquired by lawbreakers at any time, so that the privacy of the user is revealed, and the loss of spirit and materials is caused. To fundamentally solve such risks, the best method is to develop a sensor without image output, and even developers cannot acquire images of users, so that the privacy leakage problem worried by the users is fundamentally solved.

Disclosure of Invention

The invention aims to overcome the defects of the prior art and provide a human body posture detection sensor and a lighting control method which do not relate to personal privacy.

In order to achieve the purpose, the invention adopts the following technical scheme:

on one hand, the human body posture detection sensor which does not relate to personal privacy comprises an image sensing unit, an image processing unit and a coding output unit which are arranged in a chip; the image sensing unit is used for transmitting the acquired image to the image processing unit, and the image processing unit infers and analyzes scene information and human body posture information in the image through a deep learning recognition model; and the code output unit is used for setting corresponding control codes according to the scene information and the human body posture information and outputting the control codes.

In some embodiments, the image sensing unit includes an array module, a row selection module, and a signal processing module;

the array module is used for irradiating the pixel array by external light to generate a photoelectric effect and generate corresponding charges in the pixel unit;

the row selection module is used for gating the corresponding row pixel units;

and the signal processing module is used for converting the image signal into a digital image signal and outputting the digital image signal.

In some embodiments, the image processing unit comprises an acquisition module, a data processing module, and a learning module;

the acquisition module is used for acquiring an image as a data set;

the data processing module is used for processing the data set to obtain a deep learning data set;

the learning module is used for creating a deep learning model, and training, verifying and testing the deep learning model by using the deep learning data set to obtain a deep learning identification model.

In some embodiments, the images captured by the capture module include human pose and gesture images of a scene.

In some embodiments, the data processing module comprises a classification labeling sub-module and a data set partitioning sub-module;

the classification and labeling submodule is used for classifying and labeling the data set;

the data set dividing submodule is used for dividing the data set into a training data set, a verification data set and a test data set according to the ratio of 6:2: 2.

In some embodiments, the learning module includes a training sub-module, a validation sub-module, and a testing sub-module;

the training submodule is used for training the deep learning model by utilizing a training data set;

the verification submodule is used for verifying the deep learning model by utilizing a verification data set;

and the testing submodule is used for testing the deep learning model by utilizing a testing data set so as to obtain an identification model.

On the other hand, a lighting control method based on the above-mentioned human body posture detection sensor not involving privacy, the human body posture detection sensor not involving privacy being electrically connected to a lighting apparatus, is characterized by comprising:

the human body posture detection sensor which does not relate to personal privacy collects scene information and human body posture information in space in real time;

carrying out reasoning analysis on the acquired scene information and human body posture information in the space by using the recognition model, and classifying;

generating a control code for controlling the lighting device according to the classification condition;

and the lighting equipment performs light adjustment according to the control code.

In some embodiments, in the step of performing inference analysis on the collected scene information and human body posture information in the space by using the recognition model and classifying, when the information related to the personal privacy action is included, the information is classified into a classification type corresponding to other information which is not related to the personal privacy action.

In some embodiments, the actions relating to personal privacy include sex actions and nude gestures.

Compared with the prior art, the invention has the beneficial effects that: the invention encapsulates the image sensing unit, the image processing unit and the coding output unit on a chip, the chip has no image output interface and pins and can not output images, the images acquired by the image sensing unit can only be provided for the image processing unit in the same chip for processing, and the coding output unit outputs the scene information, the human body posture information and the gesture information to developers in a digital or character form for use according to the processing result of the image processing unit. When the method is used for controlling the lighting equipment, certain privacy scenes and actions, such as sex actions and naked posture, can be classified into classification types of other information which does not relate to personal privacy actions, so that the privacy of an output user is fundamentally avoided, the privacy leakage problem worried by the user is solved, and meanwhile, the corresponding lighting effect can be met.

The foregoing description is only an overview of the technical solutions of the present invention, and in order to make the technical means of the present invention more clearly understood, the present invention may be implemented according to the content of the description, and in order to make the above and other objects, features, and advantages of the present invention more apparent, the following detailed description will be given of preferred embodiments.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

FIG. 1 is a block diagram of a structural package of a particular embodiment of a human gesture detection sensor of the present invention that does not involve personal privacy;

FIG. 2 is a schematic block diagram of an image sensing unit in an embodiment of a human posture detection sensor of the present invention not related to privacy of a person;

FIG. 3 is a schematic block diagram of an image processing unit in an embodiment of a human gesture detection sensor of the present invention not related to privacy of a person;

FIG. 4 is a schematic block diagram of a data processing module in an embodiment of a human gesture detection sensor of the present invention not related to personal privacy;

FIG. 5 is a schematic block diagram of a learning module in an embodiment of a human gesture detection sensor of the present invention that is not related to privacy.

Detailed Description

The technical solutions of the present invention will be described clearly and completely with reference to specific embodiments of the present invention, and it should be understood that the described embodiments are a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It will be understood that the terms "comprises" and/or "comprising," when used in this specification and claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the specification of the present invention and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

It should be further understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

FIG. 1 shows a schematic diagram of one embodiment of a human gesture detection sensor of the present invention that does not involve personal privacy.

The human body posture detection sensor which does not relate to personal privacy comprises an image sensing unit 1, an image processing unit 2 and a code output unit 3 which are arranged in a chip (the sensor can be in a chip packaging form and can also be in other packaging forms which cannot be disassembled); the image sensing unit 1 is used for transmitting the acquired image to the image processing unit 2, and the image processing unit 2 infers and analyzes scene information and human body posture information in the image through a deep learning recognition model; and the code output unit 3 is used for setting corresponding control codes according to the scene information and the human body posture information and outputting the control codes.

The sensor is free of an image output interface and a PIN PIN, such as any form of interface and image data PIN PIN of MIPI, DVP, SPI and the like. The image data collected by the image sensing unit 1 can only be transmitted inside the chip and can only be transmitted to the image processing unit 2 for processing, and the output result after the processing by the image processing unit 2 can only output the numeric or letter codes of various scenes and human postures through the code output unit 3, such as 0000 for reading, 0001 for sleeping, 0002 for drinking tea, and the like. Therefore, the sensor does not transmit any image data to a user or a developer, only outputs a code and transmits the code to the developer for use through an external PIN PIN, and internal image data cannot be obtained even if the sensor or the developer is forcibly disassembled, so that the problem that image data concerning privacy worried by the user is leaked is fundamentally solved, for example, the image data of sexual love of similar users or the image data of naked bodies can be prevented from being leaked.

In addition, in some embodiments, the functions that can be realized by the code output unit 3 can also be integrated in the image processing unit 2, so that it is not necessary to separately partition one code output unit 3.

Fig. 2 shows a schematic block diagram of the image sensing unit 1 in some embodiments. The image sensing unit 1 includes an array module 11, a row selection module 12, and a signal processing module 13; wherein the content of the first and second substances,

the array module 11 is used for irradiating the pixel array by external light to generate a photoelectric effect and generate corresponding charges in the pixel unit;

a row selection module 12 for gating the corresponding row of pixel cells;

and a signal processing module 13, configured to convert the image signal into a digital image signal and output the digital image signal.

Fig. 3 shows a schematic block diagram of the image processing unit 2 in some embodiments, the image processing unit 2 comprising an acquisition module 21, a data processing module 22 and a learning module 23; wherein the content of the first and second substances,

the acquisition module 21 is configured to acquire an image as a data set, where the acquired image includes a human body posture and a gesture image including a scene;

and the data processing module 22 is configured to process the data set to obtain a deep learning data set.

In some embodiments, referring to fig. 4, the data processing module 22 includes a classification labeling sub-module 221 and a data set partitioning sub-module 222; wherein the content of the first and second substances,

the classification labeling submodule 221 is configured to classify and label the data set;

and the data set dividing submodule 222 is used for dividing the data set into a training data set, a verification data set and a test data set according to the ratio of 6:2:2, and using the training data set, the verification data set and the test data set as deep learning data sets.

In some embodiments, each image is first classified, the corresponding scene and body pose of the image are determined, and the image is labeled using the scene and body pose, such as 0000 for reading, 0001 for sleeping, 0002 for drinking tea, and so on.

And the learning module 23 is configured to create a deep learning model, and train, verify and test the deep learning model by using the deep learning data set to obtain a deep learning identification model.

By transplanting the deep learning identification model into the chip, the inference analysis of the on-site acquired image can be realized.

In some embodiments, referring to fig. 5, learning module 23 includes a training submodule 231, a validation submodule 232, and a testing submodule 233; wherein the content of the first and second substances,

a training submodule 231 for training the deep learning model by using the training data set;

the verification submodule 232 is used for verifying the deep learning model by using the verification data set;

and the testing sub-module 233 is used for testing the deep learning model by using the testing data set to obtain the recognition model.

In some embodiments, a deep learning model is trained using a training dataset, a hyper-parameter of the deep learning model is adjusted using a validation dataset, and a performance of the deep learning model is evaluated using a test dataset.

It should be noted that the data acquisition, data processing and learning processes are completed on a server and a workstation with stronger processing capability, and the inference analysis of the on-site acquired images is realized by transplanting the trained deep learning identification model into a chip.

Based on the human body posture detection sensor which does not relate to the personal privacy, the human body posture detection sensor can be applied to some specific fields, the most suitable is the illumination field, and the light regulation and control of the illumination equipment are realized through the human body posture detection sensor which does not relate to the personal privacy. The method of regulation is described in the following specific examples.

An illumination control method based on the human body posture detection sensor which does not relate to personal privacy is characterized in that the human body posture detection sensor which does not relate to personal privacy is electrically connected with an illumination device, and the method comprises the following steps:

and S10, the human body posture detection sensor which does not relate to personal privacy collects scene information and human body posture information in the space in real time.

Because light needs to be regulated and controlled in real time during actual use, a human body posture detection sensor which does not relate to personal privacy needs to acquire scene information and human body posture information in space in real time. The scene information and the human body posture information in the space are combined to provide basic data for the adjustment of the light.

And S20, carrying out inference analysis on the acquired scene information and human body posture information in the space by using the recognition model, and classifying.

Due to the fact that the brightness of the light needs to be adjusted according to different scene information and human body posture information, for example, a user needs a light effect when reading books, needs a light effect when resting, and needs a light effect when drinking tea. Of course, in order to better protect the privacy of the individual on the premise of ensuring the corresponding lighting effect, when the scene information and the body posture information include information of the privacy action of the individual (for example, a sexual action and a nude posture), the scene information and the body posture information are classified into a classification type with other information which does not relate to the privacy action of the individual. For example, when a sexual action is included in the scene information, the situation is classified as the situation when the user is at rest, i.e., the light effect in this case is the same as the light effect when the user is at rest.

S30, generating a control code for controlling the lighting equipment according to the classification condition;

since the human body posture detection sensor which does not relate to personal privacy does not output image data, but only outputs the image data in a code form, after scene information and human body posture information are classified, a code for controlling light needs to be generated according to the classification condition, for example, a 0000 code is generated, which represents the reading state of a user, and then the light effect of the lighting device is adjusted to the brightness degree suitable for reading; generating a 0001 code, which represents a state of rest of the user, and adjusting the lighting effect of the lighting equipment to a brightness degree suitable for the rest; and generating 0002 codes, which represents the state of drinking tea by the user, and adjusting the lighting effect of the lighting device to the light and shade degree suitable for drinking tea at the moment.

And S40, the lighting equipment adjusts the light according to the control code.

And the lighting equipment adjusts the light and shade degree of the corresponding lamp according to the received control code. Because the lighting equipment receives the control code directly, the lighting equipment can be directly used without further data conversion and processing, the requirement on lighting equipment hardware is low, the reflection is more sensitive, and the switching of the lighting effect is smoother.

It should be noted that, in practical use, the human body posture detection sensor not related to the privacy may be disposed separately from the lighting device, or may be integrated into the lighting device for use, which may be determined according to lighting products, such as a table lamp, for example, the human body posture detection sensor not related to the privacy may be integrated onto the table lamp, and if the lighting system has a wide coverage area, such as a bedroom or a living room, the human body posture detection sensor not related to the privacy may be disposed separately from the lighting system, as long as the human body posture detection sensor not related to the privacy is electrically connected to the lighting system.

While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.