CN118019187A - Remote control system and method for LED projection lamp - Google Patents
Remote control system and method for LED projection lamp Download PDFInfo
- Publication number
- CN118019187A CN118019187A CN202410284586.2A CN202410284586A CN118019187A CN 118019187 A CN118019187 A CN 118019187A CN 202410284586 A CN202410284586 A CN 202410284586A CN 118019187 A CN118019187 A CN 118019187A
- Authority
- CN
- China
- Prior art keywords
- effect control
- intention
- feature vector
- control intention
- light effect
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 230000000694 effects Effects 0.000 claims abstract description 139
- 239000013598 vector Substances 0.000 claims description 142
- 230000001795 light effect Effects 0.000 claims description 126
- 238000012545 processing Methods 0.000 claims description 26
- 230000009467 reduction Effects 0.000 claims description 19
- 230000004927 fusion Effects 0.000 claims description 13
- 230000011218 segmentation Effects 0.000 claims description 10
- 238000005457 optimization Methods 0.000 claims description 9
- 238000013507 mapping Methods 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 6
- 238000012937 correction Methods 0.000 claims description 5
- 230000007613 environmental effect Effects 0.000 abstract description 14
- 238000005516 engineering process Methods 0.000 abstract description 11
- 238000004458 analytical method Methods 0.000 abstract description 6
- 238000013135 deep learning Methods 0.000 abstract description 6
- 238000010586 diagram Methods 0.000 description 6
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 230000003993 interaction Effects 0.000 description 4
- 238000009792 diffusion process Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000004134 energy conservation Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
Landscapes
- Circuit Arrangement For Electric Light Sources In General (AREA)
Abstract
The application discloses a remote control system and a remote control method for an LED (light emitting diode) spotlight, which are characterized in that voice for controlling the LED spotlight effect is collected through a recording device, and a voice recognition and intention understanding technology based on deep learning is introduced at the rear end to perform semantic recognition and analysis on the voice for controlling the intention of the LED spotlight effect, so that the demand semantics of a user are accurately understood. Like this, can reduce the interference of environmental noise effectively, perfect the expression of user's intention to improve the precision of semantic identification, and optimize and generate more accurate lamp efficiency control command, through this kind of mode, can realize the remote control to the LED projecting lamp more accurately and intelligently, promoted user experience and lighting apparatus's intelligent level.
Description
Technical Field
The application relates to the technical field of intelligent LED projection lamp control, in particular to a remote control system and method of an LED projection lamp.
Background
The LED projection lamp is high-efficiency, energy-saving and environment-friendly lighting equipment and is widely applied to occasions such as buildings, landscapes and stages. With the development of intelligent home and internet of things, the control requirements of people on lighting equipment are also higher and higher. The traditional manual switch or remote controller control mode is not flexible and convenient to operate, and the pursuit of the individualized demands of users on the brightness, the color and the like of the light cannot be met. Thus, the remote control system and method becomes one of the research hotspots of LED floodlights.
Voice control, which is a remote, natural and convenient interaction mode, has been increasingly used in the smart home field. Parameters such as the switch, the brightness and the color of the LED projection lamp are controlled through voice instructions, so that more intelligent and convenient user experience can be provided. However, the accuracy of speech recognition has been a challenge, mainly due to factors including environmental noise and unclear user intent expression.
Accordingly, an optimized remote control system for LED floodlights is desired.
Disclosure of Invention
The application provides a remote control system and a remote control method for an LED (light emitting diode) spotlight, which are characterized in that voice for controlling the LED spotlight is collected through a recording device, and a voice recognition and intention understanding technology based on deep learning is introduced at the rear end to perform semantic recognition and analysis on the voice for controlling the intention of the LED spotlight, so that the demand semantics of a user are accurately understood. Like this, can reduce the interference of environmental noise effectively, perfect the expression of user's intention to improve the precision of semantic identification, and optimize and generate more accurate lamp efficiency control command, through this kind of mode, can realize the remote control to the LED projecting lamp more accurately and intelligently, promoted user experience and lighting apparatus's intelligent level.
The application also provides a remote control method of the LED projection lamp, which comprises the following steps:
Acquiring LED floodlight effect control voice acquired by recording equipment;
Performing noise reduction treatment on the LED spotlight effect control voice to obtain a denoised LED spotlight effect control voice;
performing voice recognition on the noise-reduced LED spotlight effect control voice to obtain a spotlight effect control intention voice recognition result;
performing word granularity embedded coding on the light effect control intention voice recognition result to obtain a sequence of light effect control intention word granularity embedded coding feature vectors;
Performing context global semantic coding on the sequence of the embedded coding feature vector of the lighting control intention word granularity to obtain the lighting control intention global semantic coding feature;
based on the global semantic coding features of the lighting effect control intention, generating complete lighting effect control intention expression text, and generating a remote lighting effect control instruction of the LED projection lamp.
In the above method for remotely controlling an LED projector, performing word granularity embedded coding on the recognition result of the light effect control intention voice to obtain a sequence of feature vectors of the word granularity embedded coding of the light effect control intention, including: and after word segmentation processing is carried out on the light effect control intention voice recognition result, a word granularity encoder based on a word embedding layer is used for obtaining a sequence of the light effect control intention word granularity embedded coding feature vector.
In the above remote control method for an LED projector, performing a context global semantic coding on the sequence of the light effect control intention word granularity embedded coding feature vector to obtain a light effect control intention global semantic coding feature, including: embedding the sequence of the light effect control intention word granularity into the coding feature vector through a context encoder based on a Transformer layer to obtain the light effect control intention global semantic coding feature vector as the light effect control intention global semantic coding feature.
In the above remote control method for an LED projector, the step of embedding the sequence of the light effect control intention word granularity into the coding feature vector to obtain the light effect control intention global semantic coding feature vector as the light effect control intention global semantic coding feature by a context encoder based on a transform layer includes: the sequence of embedding the light effect control intention word granularity into the coding feature vector is subjected to one-dimensional arrangement to obtain a global light effect control intention word granularity feature vector; calculating the product between the global light effect control intention word granularity feature vector and transpose vectors of the light effect control intention word granularity embedded coding feature vector in the sequence of the light effect control intention word granularity embedded coding feature vector to obtain a plurality of self-attention association matrixes; respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices; obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and weighting each light effect control intention word granularity embedded coding feature vector in the sequence of the light effect control intention word granularity embedded coding feature vector by taking each probability value in the plurality of probability values as a weight so as to obtain the light effect control intention global semantic coding feature vector.
In the above remote control method of the LED projector, generating a complete light effect control intention expression text based on the global semantic coding feature of the light effect control intention, and generating a remote light effect control instruction of the LED projector, including: the global semantic coding feature vector of the lighting effect control intention is processed through a control intention perfect expressive machine based on AIGC models to obtain a perfect lighting effect control intention expression text; and generating a remote lamp effect control instruction of the LED spotlight based on the complete lamp effect control intention expression text.
In the above remote control method for LED projector, the method for obtaining the complete lighting control intention expression text by passing the global semantic coding feature vector of the lighting control intention through a control intention complete expressive machine based on AIGC model comprises: performing feature distribution optimization on the global semantic coding feature vector of the lighting effect control intention to obtain an optimized global semantic coding feature vector of the lighting effect control intention; and the optimized lighting effect control intention global semantic coding feature vector passes through the AIGC model-based control intention perfect expressive machine to obtain the perfect lighting effect control intention expression text.
In the above remote control method for an LED projector, performing feature distribution optimization on the global semantic coding feature vector of the lighting effect control intention to obtain an optimized global semantic coding feature vector of the lighting effect control intention, including: and performing mapping fusion correction from each sequence of the global semantic coding feature vector of the lighting effect control intention and the embedded coding feature vector of the lighting effect control intention word granularity to a fused feature distribution domain to obtain the optimized global semantic coding feature vector of the lighting effect control intention.
The application also provides a remote control system of the LED projection lamp, which comprises:
The lamp effect control voice acquisition module is used for acquiring the LED spotlight effect control voice acquired by the recording equipment;
the noise reduction processing module is used for carrying out noise reduction processing on the LED spotlight effect control voice so as to obtain the LED spotlight effect control voice after noise reduction;
the voice recognition module is used for carrying out voice recognition on the LED spotlight effect control voice after noise reduction so as to obtain a spotlight effect control intention voice recognition result;
The word granularity embedded coding module is used for carrying out word granularity embedded coding on the light effect control intention voice recognition result so as to obtain a sequence of light effect control intention word granularity embedded coding feature vectors;
the context global semantic coding module is used for carrying out context global semantic coding on the sequence of the embedded coding feature vector of the lighting control intention word granularity so as to obtain the lighting control intention global semantic coding feature;
And the remote light effect control instruction generation module is used for generating complete light effect control intention expression text based on the global semantic coding features of the light effect control intention and generating a remote light effect control instruction of the LED spotlight.
In the remote control system of the LED floodlight, the word granularity embedding coding module is used for: and after word segmentation processing is carried out on the light effect control intention voice recognition result, a word granularity encoder based on a word embedding layer is used for obtaining a sequence of the light effect control intention word granularity embedded coding feature vector.
In the remote control system of the LED projector, the context global semantic coding module is configured to: embedding the sequence of the light effect control intention word granularity into the coding feature vector through a context encoder based on a Transformer layer to obtain the light effect control intention global semantic coding feature vector as the light effect control intention global semantic coding feature.
Compared with the prior art, the remote control system and the remote control method for the LED floodlight, provided by the application, have the advantages that the voice of LED floodlight effect control is collected through the recording equipment, and the voice recognition and intention understanding technology based on deep learning is introduced at the rear end to perform semantic recognition and analysis on the voice of the LED floodlight effect control intention, so that the demand semantics of a user are accurately understood. Like this, can reduce the interference of environmental noise effectively, perfect the expression of user's intention to improve the precision of semantic identification, and optimize and generate more accurate lamp efficiency control command, through this kind of mode, can realize the remote control to the LED projecting lamp more accurately and intelligently, promoted user experience and lighting apparatus's intelligent level.
Drawings
In order to more clearly illustrate the embodiments of the application or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. In the drawings:
Fig. 1 is a flowchart of a remote control method of an LED projector according to an embodiment of the present application.
Fig. 2 is a schematic diagram of a system architecture of a remote control method of an LED projector according to an embodiment of the present application.
Fig. 3 is a block diagram of a remote control system for an LED projector according to an embodiment of the present application.
Fig. 4 is an application scenario diagram of a remote control method of an LED projector according to an embodiment of the present application.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present application more apparent, the embodiments of the present application will be described in further detail with reference to the accompanying drawings. The exemplary embodiments of the present application and their descriptions herein are for the purpose of explaining the present application, but are not to be construed as limiting the application.
Unless defined otherwise, all technical and scientific terms used in the embodiments of the application have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the present application is for the purpose of describing particular embodiments only and is not intended to limit the scope of the present application.
In describing embodiments of the present application, unless otherwise indicated and limited thereto, the term "connected" should be construed broadly, for example, it may be an electrical connection, or may be a communication between two elements, or may be a direct connection, or may be an indirect connection via an intermediate medium, and it will be understood by those skilled in the art that the specific meaning of the term may be interpreted according to circumstances.
It should be noted that, the term "first\second\third" related to the embodiment of the present application is merely to distinguish similar objects, and does not represent a specific order for the objects, it is to be understood that "first\second\third" may interchange a specific order or sequence where allowed. It is to be understood that the "first\second\third" distinguishing objects may be interchanged where appropriate such that embodiments of the application described herein may be practiced in sequences other than those illustrated or described herein.
The LED projection lamp is lighting equipment using an LED as a light source, has the advantages of high efficiency, energy conservation, environmental protection and the like, and gradually replaces the traditional lighting equipment. Along with the development of intelligent home and internet of things, the control mode of the LED projection lamp is also continuously innovated and improved so as to meet the personalized requirements of users on light brightness, color and the like. The remote control system and method provide more flexibility and convenience for the LED projector.
Through smart mobile phone App, the user can parameters such as remote control LED projecting lamp's switch, luminance, colour, realizes individualized light effect, and this kind of mode lets the user control light through the cell-phone at any time and any place, has greatly promoted the convenience. In combination with speech recognition technology, a user can control the LED floodlight through voice instructions, such as adjusting brightness, changing color and the like, and the control is more intelligent and convenient. The LED projection lamp can be automatically adjusted according to the conditions of ambient light, personnel activities and the like by using an ambient sensor or a human body sensor, so that intelligent self-adaptive illumination is realized. The user can be through predetermining the timing task, let LED projecting lamp automatic on at specific time, close or adjust luminance and colour, for example adjust the lighting effect according to sunrise sunset time, improve energy utilization efficiency. The LED projection lamp can be connected to an Internet of things platform, linkage with other intelligent devices is achieved, for example, linkage with a security system, a sound system and the like is achieved, and more intelligent light experience is provided.
The remote control system and the method of the LED projection lamp become research hot spots in the lighting industry, along with the rapid development of intelligent home and Internet of things technology, the control requirement of people on lighting equipment is more and more intelligent, and the LED projection lamp is hoped to be remotely controlled through a mobile phone, a flat plate or other intelligent equipment, so that personalized light effect and scene setting are realized. The remote control system and method can help the user to better manage the energy consumption, for example, through the functions of a timing switch, remote dimming and the like, and achieve the aims of energy conservation and environmental protection. In the occasions such as buildings, landscapes, stages and the like, the remote control system and the method can realize the instant adjustment and scene switching of the lamplight, and meet the lighting requirements of different occasions. The remote control system can collect the use data, optimize the lighting scheme through data analysis, improve lighting effect and user experience. The remote control system can improve the safety of lamplight, avoid potential safety hazards caused by improper operation, and provide a more convenient control mode.
In order to meet these demands, the lighting industry continuously researches and develops various remote control systems and methods, including remote control technologies based on Wi-Fi, bluetooth, zigbee and other communication protocols, and matched APP applications and cloud platform services, so as to realize remote intelligent control of the LED projector.
Voice control, which is a remote, natural and convenient interaction mode, has been widely used in the field of smart home. Parameters such as the switch, the brightness and the color of the LED projection lamp are controlled through voice instructions, so that more intelligent and convenient user experience can be provided. However, the accuracy of speech recognition has been a challenge, mainly due to factors including environmental noise and unclear user intent expression.
The noisy sound of the surrounding environment may affect the accuracy of the speech recognition system, making it difficult for the system to correctly recognize the instructions of the user. Sometimes, the voice command of the user may be unclear or ambiguous, which increases the recognition difficulty of the voice recognition system and reduces the accuracy. The need for speech recognition systems to be able to cope with multiple languages and dialects for users in different locales and cultural backgrounds also increases the complexity of recognition.
Although these challenges exist, with the continuous development of speech recognition technology, innovations such as noise suppression, contextual understanding, and personalized models, the application prospect of speech control in the smart home field is still quite broad.
In one embodiment of the present application, fig. 1 is a flowchart of a remote control method of an LED projector according to an embodiment of the present application. Fig. 2 is a schematic diagram of a system architecture of a remote control method of an LED projector according to an embodiment of the present application. As shown in fig. 1 and 2, a remote control method of an LED projector according to an embodiment of the present application includes: 110, acquiring LED floodlight effect control voice acquired by recording equipment; 120, performing noise reduction treatment on the LED spotlight effect control voice to obtain a denoised LED spotlight effect control voice; 130, performing voice recognition on the noise-reduced LED spotlight effect control voice to obtain a spotlight effect control intention voice recognition result; 140, performing word granularity embedded coding on the light effect control intention voice recognition result to obtain a sequence of light effect control intention word granularity embedded coding feature vectors; 150, performing context global semantic coding on the sequence of the embedded coding feature vectors of the lighting control intention word granularity to obtain lighting control intention global semantic coding features; 160, generating complete lighting effect control intention expression text based on the global semantic coding features of the lighting effect control intention, and generating a remote lighting effect control instruction of the LED projection lamp.
In the step 110, the LED spotlight effect control voice collected by the recording device is obtained, so that the recording device can capture clear voice signals for subsequent processing and recognition, high-quality voice collection can provide more reliable input for subsequent processing, and the accuracy of voice recognition is improved. In step 120, the noise reduction processing is performed on the LED spotlight effect control voice to obtain the noise-reduced LED spotlight effect control voice, and a suitable noise reduction algorithm and parameters are selected to effectively remove the environmental noise, meanwhile, key information of the voice signal is reserved as much as possible, and the noise reduction processing can improve the quality of the voice signal, so that the subsequent voice recognition accuracy is facilitated. In step 130, voice recognition is performed on the noise-reduced LED projector lamp effect control voice to obtain a lamp effect control intention voice recognition result, a voice recognition engine suitable for a scene is selected, accuracy and instantaneity of a voice recognition model are considered, and an accurate voice recognition result is a key for realizing intelligent household lamp effect control, so that the user intention can be ensured to be correctly understood. In the step 140, the word granularity embedded coding is performed on the light effect control intention voice recognition result to obtain a sequence of feature vectors of the light effect control intention word granularity embedded coding, a proper word embedded model is selected, the voice recognition result is converted into a continuous vector representation, semantic association among words is considered, and the word granularity embedded coding can convert the voice recognition result into a computer-processable vector representation, so that a basis is provided for subsequent semantic understanding and processing. In the step 150, the sequence of the light effect control intention word granularity embedded coding feature vector is subjected to context global semantic coding to obtain the light effect control intention global semantic coding feature, the word granularity embedded coding is converted into global semantic coding in consideration of context information, so that semantic information of the whole voice instruction is better captured, and the global semantic coding can provide richer semantic information, thereby being beneficial to accurately understanding the control intention of a user. In step 160, based on the global semantic coding feature of the lighting effect control intention, a complete lighting effect control intention expression text is generated, a remote lighting effect control instruction of the LED projector is generated, natural language expression is generated by using global semantic coding, meanwhile, the generated control instruction is ensured to be accurate and clear, and the LED projector can be ensured to be remotely controlled according to the intention of a user by generating the complete control intention expression text and the accurate control instruction.
Aiming at the technical problems, the technical concept of the application is to collect the LED spotlight effect control voice through the recording equipment, and introduce the voice recognition and intention understanding technology based on deep learning at the rear end to carry out the semantic recognition and analysis of the spotlight effect control intention voice so as to accurately understand the demand semantics of the user. Like this, can reduce the interference of environmental noise effectively, perfect the expression of user's intention to improve the precision of semantic identification, and optimize and generate more accurate lamp efficiency control command, through this kind of mode, can realize the remote control to the LED projecting lamp more accurately and intelligently, promoted user experience and lighting apparatus's intelligent level.
The deep learning voice recognition technology can better recognize and filter out the environmental noise, improves the accuracy of voice instructions of users, is beneficial to reducing the interference of the environmental noise on voice recognition, and improves the robustness of the whole system. The deep learning technology has higher accuracy and generalization capability in terms of voice recognition and intention understanding, and can more accurately understand the semantics of the user, thereby improving the accuracy of semantic recognition. Through the intention understanding technology, the voice instruction of the user can be more accurately understood, and the specific requirements and scene requirements for controlling the lighting effect are included, so that the actual requirements of the user are better met. By more accurately understanding the intention of the user, a more accurate and user-desired light effect control instruction can be generated, and the accuracy and reliability of remote control are improved. Through improving the accuracy of semantic recognition and generating accurate control instructions, a user can more easily realize remote control of the LED projection lamp, so that user experience and the intelligent level of lighting equipment are improved.
Specifically, in the technical scheme of the application, firstly, the LED spotlight effect control voice acquired by the recording equipment is acquired. It should be appreciated that in actually performing remote control of LED floodlights, the command to control the voice is often disturbed by noise from the surrounding environment, such as traffic noise, human noise, electrical noise, etc. These noise can reduce the signal-to-noise ratio of the speech signal, resulting in reduced accuracy of speech recognition. Therefore, in order to reduce the interference of environmental noise on the voice signal and improve the quality and definition of the voice signal, in the technical scheme of the application, the noise reduction processing is required to be performed on the LED spotlight effect control voice so as to obtain the LED spotlight effect control voice after noise reduction.
Then, consider that in a remote control system of an LED projector, a user controls parameters such as on-off, brightness, color, etc. of the light through voice instructions. Therefore, in order to further perform noise reduction processing, so as to improve the accuracy and stability of voice recognition, so that the remote control of the LED projector is more reliable and accurate, the voice input of the user needs to be recognized to acquire the specific control intention of the user. Specifically, firstly, voice recognition is performed on the noise-reduced LED spotlight effect control voice so as to obtain a spotlight effect control intention voice recognition result. Therefore, the voice signal can be converted into a text form for further analysis and processing, so that voice instruction semantics of a user can be understood more accurately, and further, a corresponding light effect control instruction is generated, and remote control of light is realized.
Next, considering that the control requirement and intention semantics of the user exist in the light effect control intention voice recognition result, and the minimum semantic unit is word granularity. Therefore, in order to more fully and accurately perform semantic recognition and analysis on the light effect control intention voice recognition result so as to more accurately perform light effect control, in the technical scheme of the application, a sequence of embedding the light effect control intention word granularity into the coding feature vector is needed to be obtained through a word granularity encoder based on a word embedding layer after word segmentation processing is performed on the light effect control intention voice recognition result.
In a specific embodiment of the present application, performing word granularity embedded coding on the recognition result of the light effect control intention voice to obtain a sequence of feature vectors of the light effect control intention word granularity embedded coding, including: and after word segmentation processing is carried out on the light effect control intention voice recognition result, a word granularity encoder based on a word embedding layer is used for obtaining a sequence of the light effect control intention word granularity embedded coding feature vector.
It should be understood that the word granularity encoder based on the word embedding layer can convert the word-segmented light effect control intention voice recognition result into a sequence of continuous light effect control intention word granularity embedded coding feature vectors, so that semantic similarity and relevance among words can be captured better, and word meaning in voice instructions can be understood. The word granularity encoder based on the word embedding layer can convert the high-dimensional word representation into a lower-dimensional continuous vector representation, thus facilitating subsequent calculation and processing, and reducing the complexity of data. The sequence of the light effect control intention word granularity embedded coding feature vector can better express semantic information in a voice instruction, and is favorable for accurately understanding and processing the intention of a user in subsequent steps.
Further, in order to better understand the lighting effect control intention, the relationship and the context information between the individual words in the voice recognition result of the lighting effect control intention need to be considered. The transform layer based context encoder is a powerful model that can capture global semantic information in the light effect control intent speech recognition results through a self-attention mechanism. That is, it can contextually encode each word in the input sequence such that the representation of each word can be affected by the entire sequence. Based on the above, in the technical scheme of the application, the sequence of the light effect control intention word granularity embedded coding feature vector is further processed by a context encoder based on a transform layer to obtain the light effect control intention global semantic coding feature vector. By inputting the sequence of the embedded coding feature vector of the light effect control intention word granularity into the context encoder based on the transform layer, the semantic information of the whole voice instruction can be captured, so that the control intention of a user can be understood more comprehensively and accurately, and the accuracy and the intelligent level of the light effect control are improved.
In a specific embodiment of the present application, performing context global semantic coding on the sequence of the light effect control intention word granularity embedded coding feature vector to obtain a light effect control intention global semantic coding feature, including: embedding the sequence of the light effect control intention word granularity into the coding feature vector through a context encoder based on a Transformer layer to obtain the light effect control intention global semantic coding feature vector as the light effect control intention global semantic coding feature.
Further, in a specific embodiment of the present application, embedding the sequence of the lighting control intention word granularity into the coding feature vector through a context encoder based on a transform layer to obtain a lighting control intention global semantic coding feature vector as the lighting control intention global semantic coding feature, including: the sequence of embedding the light effect control intention word granularity into the coding feature vector is subjected to one-dimensional arrangement to obtain a global light effect control intention word granularity feature vector; calculating the product between the global light effect control intention word granularity feature vector and transpose vectors of the light effect control intention word granularity embedded coding feature vector in the sequence of the light effect control intention word granularity embedded coding feature vector to obtain a plurality of self-attention association matrixes; respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices; obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and weighting each light effect control intention word granularity embedded coding feature vector in the sequence of the light effect control intention word granularity embedded coding feature vector by taking each probability value in the plurality of probability values as a weight so as to obtain the light effect control intention global semantic coding feature vector.
And then, the global semantic coding feature vector of the lighting effect control intention is passed through a control intention perfect expressive machine based on AIGC model to obtain perfect lighting effect control intention expression text. That is, the lighting control intent global context semantic feature information is utilized to generate more complete and accurate expression text. The expression text can more accurately describe the intention and the requirement of a user, including the specific operation of the lamplight, the setting of the brightness, the color and other parameters. And further, generating a remote light effect control instruction of the LED projector based on the complete light effect control intention expression text. Therefore, the accuracy of understanding and expressing the intention of the user can be improved, so that more intelligent and accurate lamp efficiency control is realized, the remote control of the LED projection lamp can be realized more accurately and intelligently, and the user experience and the intelligent level of the lighting equipment are improved.
In a specific embodiment of the present application, based on the global semantic coding feature of the lighting control intention, generating a complete lighting control intention expression text, and generating a remote lighting control instruction of the LED projector, including: the global semantic coding feature vector of the lighting effect control intention is processed through a control intention perfect expressive machine based on AIGC models to obtain a perfect lighting effect control intention expression text; and generating a remote lamp effect control instruction of the LED spotlight based on the complete lamp effect control intention expression text.
Further, in a specific embodiment of the present application, passing the light effect control intention global semantic coding feature vector through a AIGC model-based control intention perfect expressive machine to obtain the perfect light effect control intention expression text includes: performing feature distribution optimization on the global semantic coding feature vector of the lighting effect control intention to obtain an optimized global semantic coding feature vector of the lighting effect control intention; and the optimized lighting effect control intention global semantic coding feature vector passes through the AIGC model-based control intention perfect expressive machine to obtain the perfect lighting effect control intention expression text.
Further, in a specific embodiment of the present application, performing feature distribution optimization on the global semantic coding feature vector of the lighting effect control intention to obtain an optimized global semantic coding feature vector of the lighting effect control intention, including: and performing mapping fusion correction from each sequence of the global semantic coding feature vector of the lighting effect control intention and the embedded coding feature vector of the lighting effect control intention word granularity to a fused feature distribution domain to obtain the optimized global semantic coding feature vector of the lighting effect control intention.
In the technical scheme of the application, the sequence of the light effect control intention word granularity embedded coding feature vector expresses text semantic features embedded based on word segmentation semantic coding of the light effect control intention voice recognition result, so that after the sequence of the light effect control intention word granularity embedded coding feature vector passes through a context encoder based on a Transformer layer, global text semantic features associated based on word segmentation embedded semantic context can be enhanced, and the expression effect of the light effect control intention global semantic coding feature vector is enhanced.
But this also causes the text semantic feature representation of the light effect control intent global semantic coding feature vector to deviate from the word segmentation semantic based text semantic feature representation of the light effect control intent granularity embedded coding feature vector sequence, thereby affecting the correspondence of the text representation of the refined light effect control intent expression text of the light effect control intent global semantic coding feature vector to control instructions obtained by a control intent refined expressior based on AIGC model. Accordingly, the present application contemplates improving the textual semantic feature representation of the lighting control intent global semantic coding feature vector by further fusing the lighting control intent global semantic coding feature vector with the sequence of lighting control intent word granularity embedded coding feature vectors.
In addition, in order to promote the fusion effect, the mapping fusion correction from each of the global semantic coding feature vector of the lighting effect control intention and the sequence of the embedded coding feature vector of the lighting effect control intention word granularity to the fused feature distribution domain is carried out by taking into consideration the feature distribution information representation difference caused by the global-local text semantic feature distribution difference between the global semantic coding feature vector of the lighting effect control intention and the sequence of the embedded coding feature vector of the lighting effect control intention word granularity, specifically expressed as: performing mapping fusion correction from each sequence of the global semantic coding feature vector of the lighting effect control intention and the embedded coding feature vector of the lighting effect control intention word granularity to a fused feature distribution domain by using the following optimization formula to obtain the global semantic coding feature vector of the lighting effect control intention after optimization; wherein, the optimization formula is: ; wherein/> Is the global semantic coding feature vector of the lighting effect control intention,/>Is a cascading characteristic vector obtained by cascading the sequence of the embedded coding characteristic vector of the granularity of the lighting effect control intention word, and the characteristic vector/>And/>Having the same length,/>And/>Feature vector/>, respectivelyMean and standard deviation of corresponding feature sets,/>And/>Feature vector/>, respectivelyMean and standard deviation of corresponding feature sets,/>Representing the position-by-position evolution of the feature vector, and/>Is a logarithm based on 2,/>Representing addition by position,/>Representing multiplication by location,/>And representing the optimized lighting effect control intention global semantic coding feature vector.
Here, in order to promote the mapping effect of the sequence of the global semantic coding feature vector of the lighting effect control intention and the embedded coding feature vector of the lighting effect control intention word granularity under the feature fusion scene to the fusion feature distribution domain, on the basis that the traditional weighted fusion mode has limitation on deducing the semantic space evolution diffusion mode based on feature superposition, the fusion effect of the global semantic coding feature vector of the lighting effect control intention and the sequence of the embedded coding feature vector of the lighting effect control intention word granularity is promoted by adopting a mode of combining the low-order superposition fusion mode and the high-order superposition fusion mode of the space and simulating the evolution center and the evolution track through the statistical feature interaction relation of the global semantic coding feature vector of the lighting effect control intention word granularity under the action of different evolution diffusion speed fields so as to reconstruct the semantic space diffusion under the fusion scene based on asynchronous evolution, thereby effectively promoting the projection effect in the same high-dimensional feature space. Therefore, the text semantic feature expression effect of the lamp effect control intention global semantic coding feature vector is improved, and accordingly the correspondence of the text expression of the complete lamp effect control intention expression text obtained by the lamp effect control intention global semantic coding feature vector through the control intention complete expressive machine based on AIGC models relative to a control instruction is improved. Like this, can reduce the interference of environmental noise effectively, perfect the expression of the lamp effect control pronunciation intention of user input to improve the precision of semantic identification, and optimize and generate more accurate lamp effect control command, through this kind of mode, can realize the remote control to the LED projecting lamp more accurately and intelligently, promoted user experience and lighting apparatus's intelligent level.
In summary, the remote control method of the LED projection lamp based on the embodiment of the application is explained, which can effectively reduce the interference of environmental noise, perfect the expression of user intention, thereby improving the accuracy of semantic recognition, and optimally generating more accurate lamp effect control instructions.
Fig. 3 is a block diagram of a remote control system for an LED projector according to an embodiment of the present application. As shown in fig. 3, the remote control system 200 of the LED projector includes: the lamp efficiency control voice acquisition module 210 is used for acquiring the LED spotlight efficiency control voice acquired by the recording device; the noise reduction processing module 220 is configured to perform noise reduction processing on the LED light projector lamp effect control voice to obtain a noise-reduced LED light projector lamp effect control voice; the voice recognition module 230 is configured to perform voice recognition on the noise-reduced LED projector lamp effect control voice to obtain a lamp effect control intention voice recognition result; the word granularity embedded coding module 240 is configured to perform word granularity embedded coding on the recognition result of the lighting control intention voice to obtain a sequence of feature vectors of the lighting control intention word granularity embedded coding; the context global semantic coding module 250 is configured to perform context global semantic coding on the sequence of the light effect control intention word granularity embedded coding feature vector to obtain a light effect control intention global semantic coding feature; the remote lighting control instruction generating module 260 is configured to generate complete lighting control intention expression text based on the global semantic coding feature of the lighting control intention, and generate a remote lighting control instruction of the LED projector.
In the remote control system of the LED projector, the word granularity embeds a coding module for: and after word segmentation processing is carried out on the light effect control intention voice recognition result, a word granularity encoder based on a word embedding layer is used for obtaining a sequence of the light effect control intention word granularity embedded coding feature vector.
In the remote control system of the LED projector, the context global semantic coding module is configured to: embedding the sequence of the light effect control intention word granularity into the coding feature vector through a context encoder based on a Transformer layer to obtain the light effect control intention global semantic coding feature vector as the light effect control intention global semantic coding feature.
It will be appreciated by those skilled in the art that the specific operation of the respective steps in the above-described remote control system for an LED projector has been described in detail in the above description of the remote control method for an LED projector with reference to fig. 1 to 2, and thus, repetitive description thereof will be omitted.
As described above, the remote control system 200 of the LED projector according to the embodiment of the application may be implemented in various terminal devices, such as a server or the like for remote control of the LED projector. In one example, the remote control system 200 of the LED projector according to an embodiment of the application may be integrated into the terminal device as one software module and/or hardware module. For example, the remote control system 200 of the LED projector may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the remote control system 200 of the LED projector may also be one of a number of hardware modules of the terminal device.
Alternatively, in another example, the remote control system 200 of the LED projector and the terminal device may be separate devices, and the remote control system 200 of the LED projector may be connected to the terminal device through a wired and/or wireless network and transmit the interaction information in a agreed data format.
Fig. 4 is an application scenario diagram of a remote control method of an LED projector according to an embodiment of the present application. As shown in fig. 4, in the application scenario, first, an LED projector effect control voice (e.g., C as illustrated in fig. 4) collected by a recording apparatus is acquired; then, the obtained LED floodlight effect control voice is input into a server (for example, S as illustrated in FIG. 4) provided with a remote control algorithm of the LED floodlight, wherein the server can process the LED floodlight effect control voice based on the remote control algorithm of the LED floodlight to generate perfect lamp effect control intention expression text and generate a remote lamp effect control instruction of the LED floodlight.
The foregoing description of the embodiments has been provided for the purpose of illustrating the general principles of the application, and is not meant to limit the scope of the application, but to limit the application to the particular embodiments, and any modifications, equivalents, improvements, etc. that fall within the spirit and principles of the application are intended to be included within the scope of the application.
Claims (10)
1. The remote control method of the LED projection lamp is characterized by comprising the following steps of:
Acquiring LED floodlight effect control voice acquired by recording equipment;
Performing noise reduction treatment on the LED spotlight effect control voice to obtain a denoised LED spotlight effect control voice;
performing voice recognition on the noise-reduced LED spotlight effect control voice to obtain a spotlight effect control intention voice recognition result;
performing word granularity embedded coding on the light effect control intention voice recognition result to obtain a sequence of light effect control intention word granularity embedded coding feature vectors;
Performing context global semantic coding on the sequence of the embedded coding feature vector of the lighting control intention word granularity to obtain the lighting control intention global semantic coding feature;
based on the global semantic coding features of the lighting effect control intention, generating complete lighting effect control intention expression text, and generating a remote lighting effect control instruction of the LED projection lamp.
2. The method for remotely controlling an LED floodlight according to claim 1, wherein performing word granularity embedded encoding on the recognition result of the lighting control intention voice to obtain a sequence of feature vectors of the lighting control intention word granularity embedded encoding, comprises: and after word segmentation processing is carried out on the light effect control intention voice recognition result, a word granularity encoder based on a word embedding layer is used for obtaining a sequence of the light effect control intention word granularity embedded coding feature vector.
3. The method of claim 2, wherein performing a contextual global semantic coding on the sequence of lighting effect control intent word granularity embedded coding feature vectors to obtain lighting effect control intent global semantic coding features, comprises: embedding the sequence of the light effect control intention word granularity into the coding feature vector through a context encoder based on a Transformer layer to obtain the light effect control intention global semantic coding feature vector as the light effect control intention global semantic coding feature.
4. The method of claim 3, wherein embedding the sequence of the lighting control intent word granularity into the encoded feature vector through a transform layer-based context encoder to obtain the lighting control intent global semantic encoded feature vector as the lighting control intent global semantic encoded feature, comprising:
the sequence of embedding the light effect control intention word granularity into the coding feature vector is subjected to one-dimensional arrangement to obtain a global light effect control intention word granularity feature vector;
Calculating the product between the global light effect control intention word granularity feature vector and transpose vectors of the light effect control intention word granularity embedded coding feature vector in the sequence of the light effect control intention word granularity embedded coding feature vector to obtain a plurality of self-attention association matrixes;
Respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices;
obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and
And weighting each light effect control intention word granularity embedded coding feature vector in the sequence of the light effect control intention word granularity embedded coding feature vector by taking each probability value in the plurality of probability values as a weight so as to obtain the light effect control intention global semantic coding feature vector.
5. The method of claim 4, wherein generating complete light effect control intent expression text based on the light effect control intent global semantic coding feature, and generating a remote light effect control instruction for the LED light projector, comprises:
The global semantic coding feature vector of the lighting effect control intention is processed through a control intention perfect expressive machine based on AIGC models to obtain a perfect lighting effect control intention expression text;
And generating a remote lamp effect control instruction of the LED spotlight based on the complete lamp effect control intention expression text.
6. The method of claim 5, wherein passing the light effect control intent global semantic coding feature vector through a AIGC model-based control intent perfect expressive machine to obtain the perfect light effect control intent expression text, comprises:
Performing feature distribution optimization on the global semantic coding feature vector of the lighting effect control intention to obtain an optimized global semantic coding feature vector of the lighting effect control intention;
And the optimized lighting effect control intention global semantic coding feature vector passes through the AIGC model-based control intention perfect expressive machine to obtain the perfect lighting effect control intention expression text.
7. The method of claim 6, wherein performing feature distribution optimization on the lighting control intent global semantic coding feature vector to obtain an optimized lighting control intent global semantic coding feature vector, comprises: and performing mapping fusion correction from each sequence of the global semantic coding feature vector of the lighting effect control intention and the embedded coding feature vector of the lighting effect control intention word granularity to a fused feature distribution domain to obtain the optimized global semantic coding feature vector of the lighting effect control intention.
8. A remote control system for an LED projector, comprising:
The lamp effect control voice acquisition module is used for acquiring the LED spotlight effect control voice acquired by the recording equipment;
the noise reduction processing module is used for carrying out noise reduction processing on the LED spotlight effect control voice so as to obtain the LED spotlight effect control voice after noise reduction;
the voice recognition module is used for carrying out voice recognition on the LED spotlight effect control voice after noise reduction so as to obtain a spotlight effect control intention voice recognition result;
The word granularity embedded coding module is used for carrying out word granularity embedded coding on the light effect control intention voice recognition result so as to obtain a sequence of light effect control intention word granularity embedded coding feature vectors;
the context global semantic coding module is used for carrying out context global semantic coding on the sequence of the embedded coding feature vector of the lighting control intention word granularity so as to obtain the lighting control intention global semantic coding feature;
And the remote light effect control instruction generation module is used for generating complete light effect control intention expression text based on the global semantic coding features of the light effect control intention and generating a remote light effect control instruction of the LED spotlight.
9. The remote control system of an LED projector of claim 8, wherein the word granularity embeds an encoding module for: and after word segmentation processing is carried out on the light effect control intention voice recognition result, a word granularity encoder based on a word embedding layer is used for obtaining a sequence of the light effect control intention word granularity embedded coding feature vector.
10. The remote control system of an LED projector of claim 9, wherein the contextual global semantic coding module is configured to: embedding the sequence of the light effect control intention word granularity into the coding feature vector through a context encoder based on a Transformer layer to obtain the light effect control intention global semantic coding feature vector as the light effect control intention global semantic coding feature.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410284586.2A CN118019187A (en) | 2024-03-13 | 2024-03-13 | Remote control system and method for LED projection lamp |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410284586.2A CN118019187A (en) | 2024-03-13 | 2024-03-13 | Remote control system and method for LED projection lamp |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118019187A true CN118019187A (en) | 2024-05-10 |
Family
ID=90960154
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410284586.2A Pending CN118019187A (en) | 2024-03-13 | 2024-03-13 | Remote control system and method for LED projection lamp |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118019187A (en) |
-
2024
- 2024-03-13 CN CN202410284586.2A patent/CN118019187A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104049962B (en) | Foreign language voice low coverage or the method for long-range control electromechanical equipment operation | |
| CN110060677A (en) | Voice remote controller control method, device and computer readable storage medium | |
| CN107170449A (en) | Intelligent domestic system and its control method | |
| CN105512113A (en) | Communication type voice translation system and translation method | |
| JP2010511958A (en) | Gesture / voice integrated recognition system and method | |
| CN110992932B (en) | Self-learning voice control method, system and storage medium | |
| CN109542233A (en) | A kind of lamp control system based on dynamic gesture and recognition of face | |
| CN109346083A (en) | A kind of intelligent sound exchange method and device, relevant device and storage medium | |
| CN105788596A (en) | Speech recognition television control method and system | |
| CN115480923B (en) | Multi-mode intelligent classroom edge computing control system | |
| CN114743554A (en) | Intelligent household interaction method and device based on Internet of things | |
| CN117082700A (en) | LED lamplight interaction control system | |
| CN113611306A (en) | Intelligent household voice control method and system based on user habits and storage medium | |
| CN209328511U (en) | A kind of portable AI interactive voice control system | |
| CN118019187A (en) | Remote control system and method for LED projection lamp | |
| CN117193524A (en) | Man-machine interaction system and method based on multi-mode feature fusion | |
| CN112218414A (en) | Method and system for adjusting brightness of self-adaptive equipment | |
| CN116685028A (en) | Intelligent control system for digital human scene lamplight in virtual environment | |
| WO2018023523A1 (en) | Motion and emotion recognizing home control system | |
| CN115064160B (en) | Voice wake-up method and device | |
| CN116647960A (en) | Stage lighting regulation and control system based on thing networking | |
| CN107180629B (en) | Voice acquisition and recognition method and system | |
| Bharathi et al. | Signtalk: Sign language to text and speech conversion | |
| CN111489745A (en) | Chinese speech recognition system applied to artificial intelligence | |
| CN104062910A (en) | Command generation device, and equipment intelligent control method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination |