Intelligent Assistance Method and System Based on Automobile Tail Light Information

Abstract

The present invention relates to the technical field of vehicle lamp recognition. To solve the problem that the taillights of a vehicle cannot be quickly and accurately recognized in a complex traffic scenario, an intelligent assistance method and system based on vehicle taillight information are provided. The method includes: collecting vehicle taillight pictures, where the vehicle taillight pictures include taillight pictures of each vehicle type and taillight pictures of the vehicle when braking, turning, and driving normally, and classifying and labeling the vehicle taillight pictures to construct a vehicle taillight picture information library; constructing a vehicle taillight detection model, inputting the classified and labeled vehicle taillight pictures in the vehicle taillight picture information library into the vehicle taillight detection model for training to obtain a vehicle taillight recognition model; obtaining the taillight information of the vehicle in front in real time through the acquisition device of the auxiliary vehicle, and performing real-time recognition on the taillight information of the vehicle in front according to the vehicle taillight recognition model; judging the driving state of the vehicle in front according to the recognition result to perform intelligent assistance on the driving of the auxiliary vehicle.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle lamp recognition, and particularly to an intelligent assistance method based on vehicle rear lamp information and an intelligent assistance system based on vehicle rear lamp information. Background Art

[0002] As an essential lighting tool for a vehicle to drive in different light environments and to warn the following vehicle, accurate recognition of vehicle rear lamps can reduce traffic accidents caused by drivers' negligence of changes in vehicle rear lamps. With the development of deep learning, the accuracy of target image detection by a computer is getting higher and higher, and it is gradually applied to vehicle lamp recognition.

[0003] In the prior art, most of the methods for recognizing vehicle lamps through deep learning are too complex. At the same time, due to the large differences in rear lamp pictures of different vehicles and different road conditions, the involved neural network structure cannot achieve rapid detection and has low accuracy, making it difficult to accurately recognize vehicle rear lamps in reality. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention provides an intelligent assistance method based on vehicle rear lamp information, which can quickly recognize vehicle rear lamps in a complex traffic scene with high accuracy.

[0005] The technical solution adopted by the present invention is as follows:

[0006] An intelligent assistance method based on vehicle rear lamp information includes the following steps: collecting vehicle rear lamp pictures, where the vehicle rear lamp pictures include rear lamp pictures of each vehicle type and rear lamp pictures of the vehicle when braking, turning, and driving normally, and classifying and labeling the vehicle rear lamp pictures to construct a vehicle rear lamp picture information library; constructing a vehicle rear lamp detection model, inputting the classified and labeled vehicle rear lamp pictures in the vehicle rear lamp picture information library into the vehicle rear lamp detection model for training to obtain a vehicle rear lamp recognition model; obtaining the front vehicle rear lamp information in real time through the acquisition device of the auxiliary vehicle, and performing real-time recognition on the front vehicle rear lamp information according to the vehicle rear lamp recognition model; judging the driving state of the front vehicle according to the recognition result to perform intelligent assistance on the driving of the auxiliary vehicle.

[0007] In an embodiment of the present invention, when a part of the front vehicle rear lamp is damaged or contaminated, the vehicle type of the front vehicle is recognized through a vehicle recognition model, and the front vehicle rear lamp is recognized according to the vehicle rear lamp picture information library, and the driving state of the front vehicle is judged according to the recognition result.

[0008] In an embodiment of the present invention, the real-time recognition of the front vehicle tail light information according to the vehicle tail light recognition model specifically includes: acquiring a picture of the rear of the front vehicle; performing information augmentation processing on the picture of the rear of the front vehicle to convert the picture of the rear of the front vehicle to a set size; obtaining a feature sequence of the picture of the rear of the front vehicle according to a preset neural network, where the feature sequence is the position of the tail lights, the radius of the circle containing a single light, and the middle straight line passing through the center of the circle containing a single light; performing a multiplication operation on the position of the tail lights and the middle straight line to obtain a position to be recognized, dividing each position of the tail lights according to the position to be recognized, and obtaining a picture of the front vehicle tail lights according to a preset algorithm; and performing real-time recognition on the picture of the front vehicle tail lights according to the vehicle tail light recognition model.

[0009] In an embodiment of the present invention, the preset neural network includes a first convolutional neural network, which is used to classify the picture of the rear of the front vehicle. The input end of the first convolutional neural network is a picture of any size. The first convolutional neural network restores the output feature sequence to the size of the picture at the input end through a transposed convolutional layer and saves the position data of the picture at the initial input end.

[0010] In an embodiment of the present invention, the preset neural network includes a second convolutional neural network. The establishment method of the second convolutional neural network is as follows: set the input end of the second convolutional neural network as the output end of the first convolutional neural network and establish a network structure from bottom to top; establish a corresponding network structure from bottom to top for the network structure from top to bottom and perform deconvolution processing; perform operations on the next layer of the network through a 1×1 convolutional kernel according to a dimensionality reduction algorithm, perform an element-by-element operation on the next layer and the last layer, and perform 3×3 convolutional smoothing processing; repeat the above operation to establish the second convolutional neural network.

[0011] An intelligent auxiliary system based on vehicle tail light information includes: a collection module, which is used to collect pictures of vehicle tail lights. The pictures of vehicle tail lights include pictures of the tail lights of each vehicle type and pictures of the tail lights of the vehicle when braking, turning, and driving normally, and classify and label the pictures of vehicle tail lights to construct a vehicle tail light picture information library; a training module, which is used to construct a vehicle tail light detection model, input the classified and labeled pictures of vehicle tail lights in the vehicle tail light picture information library into the vehicle tail light detection model for training to obtain a vehicle tail light recognition model; an identification module, which is used to acquire the front vehicle tail light information in real time through the collection device of the auxiliary vehicle and perform real-time recognition on the front vehicle tail light information according to the vehicle tail light recognition model; and an intelligent auxiliary module, which is used to judge the driving state of the front vehicle according to the recognition result to perform intelligent assistance on the driving of the auxiliary vehicle.

[0012] In an embodiment of the present invention, when the taillight part of the vehicle in front is damaged or contaminated, the vehicle type of the vehicle in front is identified through a vehicle recognition model, and the taillight of the vehicle in front is identified according to the vehicle taillight picture information library, and the driving state of the vehicle in front is judged according to the recognition result.

[0013] In an embodiment of the present invention, the recognition module is specifically configured to: obtain a picture of the rear part of the vehicle in front; perform information amplification processing on the picture of the rear part of the vehicle in front to convert the picture of the rear part of the vehicle in front to a set size; obtain a feature sequence of the picture of the rear part of the vehicle in front according to a preset neural network, where the feature sequence is the position of the taillight, the radius of the circle containing a single taillight, and the middle straight line passing through the center of the circle containing a single taillight; perform a multiplication process on the position of the taillight and the middle straight line to obtain a position to be recognized, and divide each taillight position according to the position to be recognized, and obtain a picture of the taillight of the vehicle in front according to a preset algorithm; perform real-time recognition on the picture of the taillight of the vehicle in front according to the vehicle taillight recognition model.

[0014] In an embodiment of the present invention, the preset neural network includes a first convolutional neural network, and the first convolutional neural network is used to classify the picture of the rear part of the vehicle in front, and the input end of the first convolutional neural network is a picture of any size. The first convolutional neural network restores the output feature sequence to the size of the picture at the input end through a transposed convolutional layer, and saves the position data of the picture at the initial input end.

[0015] In an embodiment of the present invention, the preset neural network includes a second convolutional neural network, and the second convolutional neural network is established as follows: set the input end of the second convolutional neural network as the output end of the first convolutional neural network, and establish a network structure from bottom to top; establish a corresponding network structure from bottom to top for the network structure from top to bottom, and perform deconvolution processing; perform operations on the next layer of the network through a 1×1 convolutional kernel according to a dimensionality reduction algorithm, perform element-by-element operations on the next layer and the last layer, and perform 3×3 convolutional smoothing processing; repeat the above operation to establish the second convolutional neural network.

[0016] Advantages of the present invention:

[0017] The present invention collects pictures of automobile tail lights, classifies and labels the pictures of automobile tail lights to construct an information library of pictures of automobile tail lights, then inputs the classified and labeled pictures of automobile tail lights in the information library of pictures of automobile tail lights into an automobile tail light detection model for training to obtain an automobile tail light recognition model, and obtains the information of the front vehicle's tail lights in real time through the acquisition device of the auxiliary vehicle, recognizes the information of the front vehicle's tail lights in real time according to the automobile tail light recognition model, and judges the driving state of the front vehicle according to the recognition result to perform intelligent assistance on the driving of the auxiliary vehicle. Thus, it can quickly recognize automobile tail lights in complex traffic scenarios with high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flowchart of an intelligent assistance method based on automobile tail light information according to an embodiment of the present invention;

[0019] Figure 2 is a block diagram of an intelligent assistance system based on automobile tail light information according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0021] Figure 1 is a flowchart of an intelligent assistance method based on automobile tail light information according to an embodiment of the present invention.

[0022] As Figure 1 shown, the intelligent assistance method based on automobile tail light information according to an embodiment of the present invention includes the following steps:

[0023] S1, collect pictures of automobile tail lights, where the pictures of automobile tail lights include pictures of tail lights of each vehicle type and pictures of tail lights of the vehicle when braking, turning, and driving normally, and classify and label the pictures of automobile tail lights to construct an information library of pictures of automobile tail lights.

[0024] In an embodiment of the present invention, since there are differences in the tail lights of different brands and models of automobiles, the collected pictures of automobile tail lights may include, but are not limited to, pictures of tail lights of each model and each brand of automobile, and may also include pictures of automobile tail lights in different environments, such as foggy weather, rainy days, and nights.

[0025] S2. Build a car taillight detection model. Input the classified and labeled car taillight pictures in the car taillight picture information library into the car taillight detection model for training to obtain a car taillight recognition model.

[0026] In an embodiment of the present invention, the car taillight detection model can be built through the YOLOv3 network, which is obtained by adding a feature pyramid to the first layer of convolution and the second layer of convolution. The feature pyramid combines partial features of different parts, so that the obtained features contain partial features, and combines the partial features and all features of different parts, thereby effectively improving the detection accuracy.

[0027] S3. Real-time obtain the information of the taillights of the vehicle in front through the acquisition device of the auxiliary vehicle, and perform real-time recognition on the information of the taillights of the vehicle in front according to the car taillight recognition model.

[0028] In an embodiment of the present invention, performing real-time recognition on the information of the taillights of the vehicle in front according to the car taillight recognition model may specifically include: obtaining a picture of the rear of the vehicle in front; performing information amplification processing on the picture of the rear of the vehicle in front to convert the picture of the rear of the vehicle in front to a set size; obtaining a feature sequence of the picture of the rear of the vehicle in front according to a preset neural network, and the feature sequence is the position of the rear taillight, the radius of the circle containing a single taillight, and the middle straight line passing through the center of the circle containing the center of a single taillight; performing a multiplication process on the position of the rear taillight and the middle straight line to obtain a position to be recognized, dividing each position of the rear taillight according to the position to be recognized, and obtaining a picture of the taillights of the vehicle in front according to a preset algorithm; performing real-time recognition on the picture of the taillights of the vehicle in front according to the car taillight recognition model.

[0029] Specifically, first, a picture of the rear of the vehicle in front can be obtained through the shooting device or sensor of the vehicle itself, and preprocessing can be performed on the picture of the rear of the vehicle in front, which may include enlarging or reducing the picture to a certain size. During training, information amplification processing can be performed on the picture, that is, multiple pictures can be arbitrarily selected for rotation or cropping, and at the same time, size changes are also added, and the size of the picture can be arbitrarily changed to the size of the preset size.

[0030] Furthermore, the preset neural network may include a first convolutional neural network and a second convolutional neural network. Among them, the first convolutional neural network can be used to classify the picture of the rear of the vehicle in front, and the input end of the first convolutional neural network is a picture of arbitrary size. The first convolutional neural network can restore the output feature sequence to the size of the picture at the input end through a transposed convolutional layer and save the position data of the picture at the initial input end.

[0031] Among them, the first convolutional neural network restores the output feature sequence to the size of the picture at the input end through a transposed convolutional layer may include: obtaining the calibration position of the input picture, and the calculation formula is:

[0032]

[0033] Among them, aX is the abscissa of the pixel point of the input image, aY is the ordinate of the pixel point of the input image, bX is the abscissa of the pixel point of the output image, bY is the ordinate of the pixel point of the output image, aM is the height of the input image, bM is the height of the output image, aN is the width of the input image, and bN is the width of the output image; Calculate the deviation degrees of the input image on the X-axis and Y-axis respectively, that is, c = aX - X, d = aY - Y, and calculate (p, q) in the transposed convolution process according to the deviation degrees. The calculation formula is:

[0034]

[0035] Among them, f[p, q] is the position of the pixel point at the p-th row and q-th column of the input image, and b[p, q] is the position of the pixel point at the p-th row and q-th column of the output image.

[0036] Furthermore, the establishment method of the second convolutional neural network specifically includes: setting the input end of the second convolutional neural network as the output end of the first convolutional neural network, and establishing a bottom-up network structure; establishing a bottom-up network structure corresponding to the top-down network structure, and performing deconvolution processing; operating on the next layer of the network through a 1×1 convolutional kernel according to the dimensionality reduction algorithm, performing element-wise operations on the next layer and the last layer, and performing 3×3 convolutional smoothing processing; repeating the above operation to establish the second convolutional neural network.

[0037] Even further, the contour line of the rear vehicle's tail light can be obtained according to a preset algorithm, that is, arbitrarily select a point on the middle straight line passing through the center of the single taillight ring, and make a cutting line and a vertical line, and then obtain the middle point at the intersection of the vertical line and the range passing through the center of the single taillight ring, and move from the middle point to both sides respectively.

[0038] S4. Judge the driving state of the vehicle in front according to the recognition result to perform intelligent assistance for the driving of the auxiliary vehicle.

[0039] In an embodiment of the present invention, when a part of the rear vehicle's taillight is damaged or contaminated, the vehicle type of the vehicle in front can be recognized through the vehicle recognition model, and the rear vehicle's taillight can be recognized according to the automobile taillight picture information library, and the driving state of the vehicle in front can be judged according to the recognition result.

[0040] The intelligent assistance method based on automotive tail light information according to an embodiment of the present invention collects pictures of automotive tail lights, classifies and labels the pictures of automotive tail lights to construct an information library of pictures of automotive tail lights, then inputs the classified and labeled pictures of automotive tail lights in the information library of pictures of automotive tail lights into an automotive tail light detection model for training to obtain an automotive tail light recognition model, and obtains the information of the front vehicle's tail lights in real time through the acquisition device of the auxiliary vehicle, recognizes the information of the front vehicle's tail lights in real time according to the automotive tail light recognition model, and judges the driving state of the front vehicle according to the recognition result to perform intelligent assistance on the driving of the auxiliary vehicle. Thus, it is possible to quickly recognize automotive tail lights in a complex traffic scenario with a relatively high accuracy rate.

[0041] To implement the intelligent assistance method based on automotive tail light information in the above embodiment, the present invention also proposes an intelligent assistance system based on automotive tail light information.

[0042] As Figure 2 shown, the intelligent assistance system based on automotive tail light information according to an embodiment of the present invention includes: an acquisition module 100, a training module 200, an identification module 300, and an intelligent assistance module 400. The acquisition module 100 is used to collect pictures of automotive tail lights. The pictures of automotive tail lights include pictures of tail lights of each vehicle type and pictures of tail lights of a vehicle when braking, turning, and driving normally, and classify and label the pictures of automotive tail lights to construct an information library of pictures of automotive tail lights; the training module 200 is used to construct an automotive tail light detection model, input the classified and labeled pictures of automotive tail lights in the information library of pictures of automotive tail lights into the automotive tail light detection model for training to obtain an automotive tail light recognition model; the identification module 300 is used to obtain the information of the front vehicle's tail lights in real time through the acquisition device of the auxiliary vehicle, and recognize the information of the front vehicle's tail lights in real time according to the automotive tail light recognition model; the intelligent assistance module 400 is used to judge the driving state of the front vehicle according to the recognition result to perform intelligent assistance on the driving of the auxiliary vehicle.

[0043] In an embodiment of the present invention, since there are differences in the tail lights of different brands of automobiles and vehicle models, the pictures of automotive tail lights collected by the acquisition module 100 may include, but are not limited to, pictures of tail lights of each vehicle model and each brand of automobile, and may also include pictures of automotive tail lights in different environments, such as foggy weather, rainy days, nights, etc.

[0044] In an embodiment of the present invention, an automotive tail light detection model can be constructed through the YOLOv3 network, which is obtained by adding a structure pyramid in the first-layer convolution and the second-layer convolution. The structure pyramid combines partial features of different parts, so that the obtained features contain partial features, and combines the partial features and all features of different parts. Thus, the accuracy of detection can be effectively improved.

[0045] In an embodiment of the present invention, the recognition module 300 is specifically configured to: obtain a rear view of the vehicle in front; perform information amplification processing on the rear view of the vehicle in front to convert the rear view of the vehicle in front to a set size; obtain a feature sequence of the rear view of the vehicle in front according to a preset neural network, where the feature sequence is the position of the rear lights, the radius of a single light ring, and a middle straight line passing through the center of the single light ring; perform a multiplication process on the position of the rear lights and the middle straight line to obtain a position to be recognized, divide each position of the rear lights according to the position to be recognized, and obtain a rear light picture of the vehicle in front according to a preset algorithm; perform real-time recognition on the rear light picture of the vehicle in front according to a vehicle rear light recognition model.

[0046] Specifically, first, a picture of the rear of the vehicle in front can be obtained through the shooting device or sensor of the vehicle itself, and preprocessing can be performed on the picture of the rear of the vehicle in front, which may include enlarging or reducing the picture to a certain size. During training, information amplification processing can be performed on the picture, that is, multiple pictures can be randomly selected for rotation or cropping, and at the same time, size changes are also added, and the size of the picture can be arbitrarily changed to the size of the preset size.

[0047] Further, the preset neural network may include a first convolutional neural network and a second convolutional neural network. Among them, the first convolutional neural network is used to classify the picture of the rear of the vehicle in front, and the input end of the first convolutional neural network is a picture of any size. The first convolutional neural network can restore the output feature sequence to the size of the input picture through a transposed convolutional layer and save the position data of the initial input picture.

[0048] Among them, the first convolutional neural network restores the output feature sequence to the size of the input picture through a transposed convolutional layer, which may include: obtaining the calibration position of the input picture, and the calculation formula is:

[0049]

[0050] Among them, aX is the abscissa of the pixel point of the input picture, aY is the ordinate of the pixel point of the input picture, bX is the abscissa of the pixel point of the output picture, bY is the ordinate of the pixel point of the output picture, aM is the height of the input picture, bM is the height of the output picture, aN is the width of the input picture, and bN is the width of the output picture; calculate the deviation degrees of the input picture on the X-axis and Y-axis respectively, that is, c = aX - X, d = aY - Y, and calculate (p, q) in the transposed convolutional process according to the deviation degrees. The calculation formula is:

[0051]

[0052] Among them, f[p, q] is the position of the pixel point at the p-th row and q-th column of the input picture, and b[p, q] is the position of the pixel point at the p-th row and q-th column of the output picture.

[0053] Further, the specific method for establishing the second convolutional neural network includes: setting the input end of the second convolutional neural network as the output end of the first convolutional neural network, and establishing a bottom-up network structure; establishing a bottom-up network structure corresponding to the top-down network structure and performing deconvolution processing; operating on the next layer of the network through a 1×1 convolutional kernel according to the dimensionality reduction algorithm, performing element-wise operations on the next layer and the last layer, and performing 3×3 convolutional smoothing processing; repeating the above operation to establish the second convolutional neural network.

[0054] Even further, the contour line of the rear vehicle tail light can be obtained according to a preset algorithm, that is, an arbitrary point is selected on the middle straight line passing through the center of the single taillight ring, a cutting line and a vertical line are made, and then the middle point at the intersection of the vertical line and the range passing through the center of the single taillight ring is obtained, and the middle point moves towards both sides respectively.

[0055] In an embodiment of the present invention, when the rear vehicle tail light is partially damaged or contaminated, the vehicle type of the front vehicle can be identified through the vehicle recognition model, the rear vehicle tail light of the front vehicle can be identified according to the vehicle rear tail light picture information library, and the driving state of the front vehicle can be judged according to the recognition result.

[0056] In summary, the present invention collects pictures of vehicle rear tail lights through the collection module, classifies and labels the pictures of vehicle rear tail lights to construct a vehicle rear tail light picture information library, and then inputs the classified and labeled pictures of vehicle rear tail lights in the vehicle rear tail light picture information library into the vehicle rear tail light detection model through the training module for training to obtain a vehicle rear tail light recognition model, and the information of the front vehicle rear tail light is obtained in real time through the collection device of the auxiliary vehicle, the information of the front vehicle rear tail light is recognized in real time through the recognition module, and the driving state of the front vehicle is judged according to the recognition result through the intelligent auxiliary module to perform intelligent assistance on the driving of the auxiliary vehicle. Thus, the vehicle rear tail light can be quickly recognized in a complex traffic scene with high accuracy.

[0057] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. The meaning of "a plurality" is two or more unless otherwise specifically defined.

[0058] In the present invention, unless otherwise clearly defined or limited, terms such as "installed", "connected", "coupled", "fixed", etc. shall be construed broadly. For example, it may be a fixed connection, a detachable connection, or integrated; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the internal communication of two components or the interaction relationship between two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0059] In the present invention, unless otherwise clearly defined or limited, the first feature being "on" or "under" the second feature may mean that the first and second features are in direct contact, or the first and second features are indirectly in contact through an intermediate medium. Moreover, the first feature being "above", "over" and "on top of" the second feature may mean that the first feature is directly above or obliquely above the second feature, or merely indicates that the first feature has a higher horizontal height than the second feature. The first feature being "under", "below" and "beneath" the second feature may mean that the first feature is directly below or obliquely below the second feature, or merely indicates that the first feature has a lower horizontal height than the second feature.

[0060] In the description of this specification, the description with reference to terms such as "an embodiment", "some embodiments", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0061] Any process or method description shown in the flowchart or described in other ways herein can be understood to represent a module, segment, or portion of code including one or more executable instructions for implementing a specific logical function or process. The scope of the preferred embodiments of the present invention includes additional implementations, where the functions may be executed in a substantially simultaneous manner or in a reverse order according to the functions involved, rather than in the order shown or discussed, which should be understood by those skilled in the art to which the embodiments of the present invention pertain.

[0062] The logic and / or steps represented in the flowchart or otherwise described herein can, for example, be considered a definitional sequence list of executable instructions for implementing logical functions, and can be embodied specifically in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection portion having one or more wirings (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable medium on which the program can be printed, as the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpretation, or otherwise processing as appropriate, and then storing it in a computer memory.

[0063] It should be understood that various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0064] Those of ordinary skill in the art of this technology can understand that all or part of the steps carried by the method of the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.

[0065] In addition, in each embodiment of the present invention, each functional unit may be integrated into a processing module, may exist independently as individual units physically, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware, or may be implemented in the form of a software functional module. When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0066] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. An intelligent assistance method based on automobile tail light information, characterized in that, Including the following steps: Collecting car taillight pictures, where the car taillight pictures include taillight pictures of each vehicle type and taillight pictures of the vehicle when braking, turning, and driving normally, and classifying and labeling the car taillight pictures to construct a car taillight picture information library; Constructing a car taillight detection model, inputting the classified and labeled car taillight pictures in the car taillight picture information library into the car taillight detection model for training to obtain a car taillight recognition model; Obtaining the taillight information of the vehicle in front in real time through the acquisition device of the auxiliary vehicle, and performing real-time recognition on the taillight information of the vehicle in front according to the car taillight recognition model; Judging the driving state of the vehicle in front according to the recognition result to perform intelligent assistance on the driving of the auxiliary vehicle, The real-time recognition of the taillight information of the vehicle in front according to the car taillight recognition model specifically includes: obtaining a picture of the rear of the vehicle in front; performing information amplification processing on the picture of the rear of the vehicle in front to convert the picture of the rear of the vehicle in front to a set size; obtaining a feature sequence of the picture of the rear of the vehicle in front according to a preset neural network, where the feature sequence is the position of the taillights, the radius of a single taillight ring, and the middle straight line passing through the center of the single taillight ring; performing a multiplication operation on the position of the taillights and the middle straight line to obtain a position to be recognized, dividing each taillight position according to the position to be recognized, and obtaining a car taillight picture according to a preset algorithm; performing real-time recognition on the car taillight picture according to the car taillight recognition model, The preset neural network includes a first convolutional neural network, which is used to classify the picture of the rear of the vehicle in front, and the input end of the first convolutional neural network is a picture of any size. The first convolutional neural network restores the output feature sequence to the picture size of the input end through a transposed convolutional layer and saves the position data of the initial input picture, The preset neural network includes a second convolutional neural network, and the establishment method of the second convolutional neural network is: setting the input end of the second convolutional neural network as the output end of the first convolutional neural network and establishing a network structure from bottom to top; establishing a corresponding network structure from bottom to top for the network structure from top to bottom and performing deconvolution processing; Operating on the next layer of the network through a 1×1 convolutional kernel according to a dimensionality reduction algorithm, performing an element-by-element operation on the next layer and the last layer, and performing 3×3 convolutional smoothing processing; repeating the above operation to establish the second convolutional neural network.

2. The intelligent assistance method based on automobile tail lamp information according to claim 1, wherein, When the taillights of the vehicle in front are partially damaged or contaminated, identifying the vehicle type of the vehicle in front through a vehicle recognition model, recognizing the taillights of the vehicle in front according to the car taillight picture information library, and judging the driving state of the vehicle in front according to the recognition result.

3. An intelligent auxiliary system based on automobile tail lamp information, characterized in that, Including: An acquisition module, which is used to collect car taillight pictures, where the car taillight pictures include taillight pictures of each vehicle type and taillight pictures of the vehicle when braking, turning, and driving normally, and classifying and labeling the car taillight pictures to construct a car taillight picture information library; A training module, which is used to build a car tail light detection model, input the classified and labeled car tail light pictures in the car tail light picture information library into the car tail light detection model for training to obtain a car tail light recognition model; A recognition module, which is used to obtain the front vehicle's tail light information in real time through the acquisition device of the auxiliary vehicle, and perform real-time recognition on the front vehicle's tail light information according to the car tail light recognition model; An intelligent assistance module, which is used to judge the driving state of the front vehicle according to the recognition result to provide intelligent assistance for the driving of the auxiliary vehicle; Specifically, the recognition module is used to: obtain a picture of the front vehicle's rear; perform information amplification processing on the picture of the front vehicle's rear to convert the picture of the front vehicle's rear to a set size; obtain a feature sequence of the picture of the front vehicle's rear according to a preset neural network, and the feature sequence is the position of the tail light, the radius of the circle containing a single lamp, and the middle straight line passing through the center of the circle containing a single lamp; perform a multiplication operation on the position of the tail light and the middle straight line to obtain a position to be recognized, divide each tail light position according to the position to be recognized, and obtain a picture of the front vehicle's tail light according to a preset algorithm; perform real-time recognition on the picture of the front vehicle's tail light according to the car tail light recognition model; The preset neural network includes a first convolutional neural network, which is used to classify the picture of the front vehicle's rear, and the input end of the first convolutional neural network is a picture of any size. The first convolutional neural network restores the output feature sequence to the size of the picture at the input end through a transposed convolutional layer and saves the position data of the picture at the initial input end; The preset neural network includes a second convolutional neural network, and the establishment method of the second convolutional neural network is: set the input end of the second convolutional neural network as the output end of the first convolutional neural network, and establish a network structure from bottom to top; establish a corresponding network structure from bottom to top for the network structure from top to bottom, and perform deconvolution processing; Perform operations on the next layer of the network through a 1×1 convolutional kernel according to the dimensionality reduction algorithm, perform element-by-element operations on the next layer and the last layer, and perform 3×3 convolutional smoothing processing; repeat the above operation to establish the second convolutional neural network.

4. The intelligent auxiliary system based on automobile tail lamp information according to claim 3, characterized in that, When the rear tail light of the front vehicle is partially damaged or contaminated, identify the vehicle type of the front vehicle through the vehicle recognition model, identify the rear tail light of the front vehicle according to the car tail light picture information library, and judge the driving state of the front vehicle according to the recognition result.

Images