Intelligent auxiliary method and system based on automobile tail light information

Abstract

The invention relates to the technical field of automobile lamp recognition, and provides an intelligent auxiliary method and system based on automobile tail lamp information in order to solve the problem that an automobile tail lamp cannot be quickly and accurately recognized in a complex traffic scene. The automobile tail lamp pictures comprise tail lamp pictures of each automobile type and tail lamp pictures of the automobile during braking, steering and normal driving, and classifying and marking the automobile tail lamp pictures to construct an automobile tail lamp picture information base; constructing an automobile tail lamp detection model, and inputting the classified and labeled automobile tail lamp pictures in the automobile tail lamp picture information base into the automobile tail lamp detection model for training so as to obtain an automobile tail lamp recognition model; front vehicle tail lamp information is obtained in real time through an acquisition device of the auxiliary vehicle, and the front vehicle tail lamp information is recognized in real time according to the vehicle tail lamp recognition model; the driving state of the front vehicle is judged according to the recognition result so that intelligent assistance can be conducted on driving of the auxiliary vehicle.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle light recognition, and in particular to an intelligent assistance method based on vehicle tail light information and an intelligent assistance system based on vehicle tail light information. Background Art

[0002] Headlights are essential lighting tools for cars when driving in different light environments, and they also serve as a warning to the following vehicles. Therefore, accurate recognition of vehicle taillights can reduce traffic accidents caused by drivers neglecting the changes in taillights. With the development of deep learning, the accuracy of target image detection by computers is getting higher and higher, and is gradually being applied to headlight recognition.

[0003] In the existing technology, most of the methods of identifying car lights through deep learning are too complicated. At the same time, due to the large differences in taillight images of different vehicles and under different road conditions, the neural network structure involved cannot achieve rapid detection and the accuracy is not high. In reality, it is difficult to accurately identify car taillights. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides an intelligent auxiliary method based on automobile taillight information, which can quickly identify automobile taillights in complex traffic scenes with high accuracy.

[0005] The technical solution adopted by the present invention is as follows: An intelligent assistance method based on automobile taillight information comprises the following steps: collecting automobile taillight pictures, the automobile taillight pictures including taillight pictures of each vehicle type and taillight pictures of the vehicle when braking, turning and driving normally, and classifying and labeling the automobile taillight pictures to build an automobile taillight picture information library; building an automobile taillight detection model, inputting the classified and labeled automobile taillight pictures in the automobile taillight picture information library into the automobile taillight detection model for training to obtain an automobile taillight recognition model; acquiring the preceding vehicle taillight information in real time through a collection device of an auxiliary vehicle, and recognizing the preceding vehicle taillight information in real time according to the automobile taillight recognition model; judging the driving state of the preceding vehicle according to the recognition result to provide intelligent assistance to the driving of the auxiliary vehicle.

[0006] In one embodiment of the present invention, when the taillights of the preceding vehicle are partially damaged or contaminated, the vehicle type of the preceding vehicle is identified by a vehicle identification model, and the taillights of the preceding vehicle are identified according to the automobile taillight image information library, and the driving status of the preceding vehicle is determined according to the identification result.

[0007] In one embodiment of the present invention, the real-time recognition of the taillight information of the preceding vehicle according to the automobile taillight recognition model specifically includes: obtaining a picture of the rear of the preceding vehicle; performing information amplification processing on the picture of the rear of the preceding vehicle to convert the picture of the rear of the preceding vehicle to a set size; obtaining a feature sequence of the picture of the rear of the preceding vehicle according to a preset neural network, the feature sequence being the position of the taillights, the radius of a single light ring, and a middle straight line passing through the center of the single light ring; multiplying the taillight position 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 picture of the taillights of the preceding vehicle according to a preset algorithm; and performing real-time recognition of the picture of the taillights of the preceding vehicle according to the automobile taillight recognition model.

[0008] In one 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 preceding vehicle, and the input end of the first convolutional neural network is a picture of arbitrary size. The first convolutional neural network restores the output feature sequence to the picture size of the input end through a transposed convolution layer, and saves the position data of the picture of the initial input end.

[0009] In one embodiment of the present invention, the preset neural network includes a second convolutional neural network, and the second convolutional neural network is established in the following manner: the input end of the second convolutional neural network is set as the output end of the first convolutional neural network, and a bottom-to-top network structure is established; a bottom-to-top network structure corresponding to the top-to-bottom network structure is established, and a deconvolution process is performed; the next layer of the network is operated through a 1×1 convolution kernel according to a dimensionality reduction algorithm, the next layer and the last layer are element-by-element operated, and a 3×3 convolution smoothing process is performed; the previous step is repeated to establish the second convolutional neural network.

[0010] An intelligent assistance system based on automobile taillight information comprises: an acquisition module, the acquisition module is used to acquire automobile taillight pictures, the automobile taillight pictures include taillight pictures of each type of vehicle and taillight pictures of the vehicle when braking, turning and driving normally, and the automobile taillight pictures are classified and labeled to build an automobile taillight picture information library; a training module, the training module is used to build an automobile taillight detection model, and the classified and labeled automobile taillight pictures in the automobile taillight picture information library are input into the automobile taillight detection model for training to obtain an automobile taillight recognition model; an identification module, the identification module is used to acquire the front vehicle taillight information in real time through the acquisition device of the auxiliary vehicle, and recognize the front vehicle taillight information in real time according to the automobile taillight recognition model; an intelligent assistance module, the intelligent assistance module is used to judge the driving state of the front vehicle according to the recognition result to intelligently assist the driving of the auxiliary vehicle.

[0011] In one embodiment of the present invention, when the taillights of the preceding vehicle are partially damaged or contaminated, the vehicle type of the preceding vehicle is identified by a vehicle identification model, and the taillights of the preceding vehicle are identified according to the automobile taillight image information library, and the driving status of the preceding vehicle is determined according to the identification result.

[0012] In one embodiment of the present invention, the recognition module is specifically used to: obtain a picture of the rear end of the preceding vehicle; perform information amplification processing on the picture of the rear end of the preceding vehicle to convert the picture of the rear end of the preceding vehicle to a set size; obtain a feature sequence of the picture of the rear end of the preceding vehicle according to a preset neural network, the feature sequence being the position of the taillights, the radius of a single headlight ring, and a middle straight line passing through the center of the single headlight ring; multiplying the taillight position and the middle straight line to obtain a position to be identified, and dividing each taillight position according to the position to be identified, and obtaining a picture of the taillights of the preceding vehicle according to a preset algorithm; and performing real-time identification on the picture of the taillights of the preceding vehicle according to the car taillight recognition model.

[0013] In one 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 preceding vehicle, and the input end of the first convolutional neural network is a picture of arbitrary size. The first convolutional neural network restores the output feature sequence to the picture size of the input end through a transposed convolution layer, and saves the position data of the picture of the initial input end.

[0014] In one embodiment of the present invention, the preset neural network includes a second convolutional neural network, and the second convolutional neural network is established in the following manner: the input end of the second convolutional neural network is set as the output end of the first convolutional neural network, and a bottom-to-top network structure is established; a bottom-to-top network structure corresponding to the top-to-bottom network structure is established, and a deconvolution process is performed; the next layer of the network is operated through a 1×1 convolution kernel according to a dimensionality reduction algorithm, the next layer and the last layer are element-by-element operated, and a 3×3 convolution smoothing process is performed; the previous step is repeated to establish the second convolutional neural network.

[0015] Beneficial effects of the present invention: The present invention collects car taillight pictures and classifies and annotates the car taillight pictures to build a car taillight picture information library, then inputs the classified and annotated car taillight pictures in the car taillight picture information library into a car taillight detection model for training to obtain a car taillight recognition model, and obtains the leading car taillight information in real time through a collection device of an auxiliary vehicle, recognizes the leading car taillight information in real time according to the car taillight recognition model, and judges the driving state of the leading car according to the recognition result to intelligently assist the driving of the auxiliary vehicle, thereby being able to quickly recognize car taillights in complex traffic scenes with a high accuracy rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of an intelligent assistance method based on automobile taillight information according to an embodiment of the present invention; Figure 2 The figure is a block diagram of an intelligent assistance system based on automobile taillight information according to an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Figure 1 The figure is a flow chart of an intelligent assistance method based on automobile taillight information according to an embodiment of the present invention.

[0019] like Figure 1 As shown, the intelligent assistance method based on automobile taillight information of an embodiment of the present invention includes the following steps: S1, collects car taillight pictures, including taillight pictures of each type of vehicle and taillight pictures of the vehicle when braking, turning and driving normally, and classifies and annotates the car taillight pictures to build a car taillight picture information library.

[0020] In one embodiment of the present invention, since there are differences in the taillights of cars of different brands and models, the collected car taillight pictures may include but are not limited to car taillight pictures of each model and each brand, and may also include car taillight pictures in different environments, such as foggy weather, rainy days, at night, etc.

[0021] S2, constructing a car taillight detection model, inputting the classified and labeled car taillight images in the car taillight image information database into the car taillight detection model for training to obtain a car taillight recognition model.

[0022] In one embodiment of the present invention, a car taillight detection model can be constructed through a YOLOv3 network, which is obtained by adding a structural pyramid in the first convolution layer and the second convolution layer. The structural pyramid is used to merge partial features of different parts, so that the obtained features contain partial features, and the partial features of different parts are combined with all features, thereby effectively improving the accuracy of detection.

[0023] S3, obtaining the preceding vehicle's taillight information in real time through the collecting device of the auxiliary vehicle, and identifying the preceding vehicle's taillight information in real time according to the automobile taillight recognition model.

[0024] In one embodiment of the present invention, the real-time recognition of the taillight information of the preceding vehicle is performed according to the automobile taillight recognition model, which may specifically include: obtaining a picture of the rear end of the preceding vehicle; performing information amplification processing on the picture of the rear end of the preceding vehicle to convert the picture of the rear end of the preceding vehicle to a set size; obtaining a feature sequence of the picture of the rear end of the preceding vehicle according to a preset neural network, the feature sequence being the position of the taillights, the radius of a single light ring, and a middle straight line passing through the center of the single light ring; multiplying 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 picture of the taillights of the preceding vehicle according to a preset algorithm; and performing real-time recognition of the picture of the taillights of the preceding vehicle according to the automobile taillight recognition model.

[0025] Specifically, first, the picture of the rear end of the front vehicle can be obtained through the car's own camera or sensor, and the picture of the rear end of the front vehicle can be pre-processed, which may include enlarging or reducing the picture to a certain size. During training, the picture can be processed for information amplification, that is, multiple pictures can be arbitrarily selected for rotation or cropping, and size changes are also added, and the size of the picture can be arbitrarily changed to a preset size.

[0026] Furthermore, the preset neural network may include a first convolutional neural network and a second convolutional neural network, wherein the first convolutional neural network can be used to classify the picture of the rear of the preceding vehicle, and the input end of the first convolutional neural network is a picture of arbitrary size, and the first convolutional neural network can restore the output feature sequence to the picture size of the input end through a transposed convolution layer, and save the position data of the picture of the initial input end.

[0027] The first convolutional neural network may restore the output feature sequence to the image size of the input end through the transposed convolution layer, which may include: obtaining the calibration position of the input image, and the calculation formula is:

[0028] Among them, aX is the horizontal coordinate of the pixel point of the input image, aY is the vertical coordinate of the pixel point of the input image, bX is the horizontal coordinate of the pixel point of the output image, bY is the vertical coordinate 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 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. The calculation formula is:

[0029] Among them, f[p, q] is the position of the pixel point in the pth row and qth column of the input image, and b[p, q] is the position of the pixel point in the pth row and qth column of the output image.

[0030] Furthermore, the method for establishing the second convolutional neural network specifically includes: setting the input end of the second convolutional neural network to the output end of the first convolutional neural network, and establishing a bottom-to-top network structure; establishing a bottom-to-top network structure corresponding to the top-to-bottom network structure, and performing deconvolution processing; operating the next layer of the network through a 1×1 convolution kernel according to the dimensionality reduction algorithm, performing element-by-element operations on the next layer and the last layer, and performing 3×3 convolution smoothing processing; repeating the previous step to establish the second convolutional neural network.

[0031] Furthermore, the contour line of the tail light of the front vehicle can be obtained according to a preset algorithm, that is, a point is randomly selected on the middle straight line passing through the center of a single headlight ring, and a cutting line and a vertical line are made, and then the midpoint of the intersection of the vertical line and the range passing through the center of the single headlight ring is obtained, and then moving to both sides from the midpoint.

[0032] S4, judging the driving state of the preceding vehicle according to the recognition result to provide intelligent assistance to the driving of the auxiliary vehicle.

[0033] In one embodiment of the present invention, when the taillights of the preceding vehicle are partially damaged or contaminated, the vehicle type of the preceding vehicle can be identified through a vehicle recognition model, and the taillights of the preceding vehicle can be identified based on a car taillight image information library, and the driving status of the preceding vehicle can be determined based on the recognition results.

[0034] According to the intelligent assistance method based on automobile taillight information of an embodiment of the present invention, automobile taillight pictures are collected and classified and labeled to build an automobile taillight picture information library, and then the classified and labeled automobile taillight pictures in the automobile taillight picture information library are input into the automobile taillight detection model for training to obtain a automobile taillight recognition model, and the preceding vehicle's taillight information is obtained in real time through the collecting device of the auxiliary vehicle, the preceding vehicle's taillight information is recognized in real time according to the automobile taillight recognition model, and the driving status of the preceding vehicle is judged according to the recognition result to intelligently assist the driving of the auxiliary vehicle, thereby enabling the automobile taillights to be quickly recognized in complex traffic scenarios with a high accuracy rate.

[0035] In order to implement the intelligent assistance method based on automobile taillight information of the above embodiment, the present invention also proposes an intelligent assistance system based on automobile taillight information.

[0036] like Figure 2 As shown, the intelligent assistance system based on automobile taillight information of an embodiment of the present invention includes: an acquisition module 100, a training module 200, a recognition module 300 and an intelligent assistance module 400, wherein the acquisition module 100 is used to acquire automobile taillight pictures, the automobile taillight pictures include taillight pictures of each type of vehicle and taillight pictures of the vehicle when braking, turning and driving normally, and classify and annotate the automobile taillight pictures to build an automobile taillight picture information library; the training module 200 is used to build an automobile taillight detection model, and input the classified and annotated automobile taillight pictures in the automobile taillight picture information library into the automobile taillight detection model for training to obtain an automobile taillight recognition model; the recognition module 300 is used to obtain the front vehicle taillight information in real time through the acquisition device of the auxiliary vehicle, and recognize the front vehicle taillight information in real time according to the automobile taillight recognition model; the intelligent assistance module 400 is used to judge the driving state of the front vehicle according to the recognition result to intelligently assist the driving of the auxiliary vehicle.

[0037] In one embodiment of the present invention, since there are differences in the taillights of cars of different brands and models, the car taillight pictures collected by the collection module 100 may include but are not limited to car taillight pictures of each model and each brand, and may also include car taillight pictures in different environments, such as foggy weather, rainy days, at night, etc.

[0038] In one embodiment of the present invention, a car taillight detection model can be constructed through a YOLOv3 network, which is obtained by adding a structural pyramid in the first convolution layer and the second convolution layer. The structural pyramid is used to merge partial features of different parts, so that the obtained features contain partial features, and the partial features of different parts are combined with all features, thereby effectively improving the accuracy of detection.

[0039] In one embodiment of the present invention, the recognition module 300 can be specifically used to: obtain a picture of the rear end of the preceding vehicle; perform information amplification processing on the picture of the rear end of the preceding vehicle to convert the picture of the rear end of the preceding vehicle to a set size; obtain a feature sequence of the picture of the rear end of the preceding vehicle according to a preset neural network, the feature sequence being the position of the taillights, the radius of a single headlight ring, and a middle straight line passing through the center of the single headlight ring; multiplying the position of the taillights and the middle straight line to obtain a position to be identified, and dividing each taillight position according to the position to be identified, and obtaining a picture of the taillights of the preceding vehicle according to a preset algorithm; and performing real-time recognition of the picture of the taillights of the preceding vehicle according to a car taillight recognition model.

[0040] Specifically, first, the picture of the rear end of the front vehicle can be obtained through the car's own camera or sensor, and the picture of the rear end of the front vehicle can be pre-processed, which may include enlarging or reducing the picture to a certain size. During training, the picture can be processed for information amplification, that is, multiple pictures can be arbitrarily selected for rotation or cropping, and size changes are also added, and the size of the picture can be arbitrarily changed to a preset size.

[0041] Furthermore, the preset neural network may include a first convolutional neural network and a second convolutional neural network, wherein the first convolutional neural network can be used to classify the picture of the rear of the preceding vehicle, and the input end of the first convolutional neural network is a picture of arbitrary size, and the first convolutional neural network can restore the output feature sequence to the picture size of the input end through a transposed convolution layer, and save the position data of the picture of the initial input end.

[0042] The first convolutional neural network may restore the output feature sequence to the image size of the input end through the transposed convolution layer, which may include: obtaining the calibration position of the input image, and the calculation formula is:

[0043] Among them, aX is the horizontal coordinate of the pixel point of the input image, aY is the vertical coordinate of the pixel point of the input image, bX is the horizontal coordinate of the pixel point of the output image, bY is the vertical coordinate 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 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. The calculation formula is:

[0044] Among them, f[p, q] is the position of the pixel point in the pth row and qth column of the input image, and b[p, q] is the position of the pixel point in the pth row and qth column of the output image.

[0045] Furthermore, the method for establishing the second convolutional neural network specifically includes: setting the input end of the second convolutional neural network to the output end of the first convolutional neural network, and establishing a bottom-to-top network structure; establishing a bottom-to-top network structure corresponding to the top-to-bottom network structure, and performing deconvolution processing; operating the next layer of the network through a 1×1 convolution kernel according to the dimensionality reduction algorithm, performing element-by-element operations on the next layer and the last layer, and performing 3×3 convolution smoothing processing; repeating the previous step to establish the second convolutional neural network.

[0046] Furthermore, the contour line of the tail light of the front vehicle can be obtained according to a preset algorithm, that is, a point is randomly selected on the middle straight line passing through the center of a single headlight ring, and a cutting line and a vertical line are made, and then the midpoint of the intersection of the vertical line and the range passing through the center of the single headlight ring is obtained, and then moving to both sides from the midpoint.

[0047] In one embodiment of the present invention, when the taillights of the preceding vehicle are partially damaged or contaminated, the vehicle type of the preceding vehicle can be identified through a vehicle recognition model, and the taillights of the preceding vehicle can be identified based on a car taillight image information library, and the driving status of the preceding vehicle can be determined based on the recognition results.

[0048] In summary, the present invention collects car taillight pictures through the acquisition module, and classifies and labels the car taillight pictures to build a car taillight picture information library, and then inputs the classified and labeled car taillight pictures in the car taillight picture information library into the car taillight detection model for training through the training module to obtain a car taillight recognition model, and obtains the leading vehicle taillight information in real time through the acquisition device of the auxiliary vehicle, recognizes the leading vehicle taillight information in real time through the recognition module, and judges the driving status of the leading vehicle according to the recognition result through the intelligent auxiliary module to intelligently assist the driving of the auxiliary vehicle. As a result, the car taillights can be quickly recognized in complex traffic scenes with a high accuracy rate.

[0049] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. "Multiple" means two or more, unless otherwise clearly and specifically defined.

[0050] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0051] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may mean that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. Moreover, a first feature being "above", "above" or "above" a second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. A first feature being "below", "below" or "below" a second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.

[0052] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0053] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0054] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0055] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned 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 by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0056] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0057] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, 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 in the form of a software functional module. If 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.

[0058] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. An intelligent assistance method based on automobile taillight information, characterized in that: The following steps are involved: Collecting car taillight pictures, including taillight pictures of each type of vehicle and taillight pictures of the vehicle when braking, turning and driving normally, and classifying and labeling the car taillight pictures to build a car taillight picture information library; Constructing a car taillight detection model, inputting the classified and annotated car taillight pictures in the car taillight picture information database into the car taillight detection model for training, so as to obtain a car taillight recognition model; Acquire the preceding vehicle's taillight information in real time through the collecting device of the auxiliary vehicle, and recognize the preceding vehicle's taillight information in real time according to the automobile taillight recognition model; The driving status of the preceding vehicle is determined based on the recognition results to provide intelligent assistance to the driving of the auxiliary vehicle.

2. The intelligent assistance method based on automobile taillight information according to claim 1 is characterized in that: When the taillight of the preceding vehicle is partially damaged or contaminated, the vehicle type of the preceding vehicle is identified through the vehicle identification model, and the taillight of the preceding vehicle is identified according to the automobile taillight image information library, and the driving state of the preceding vehicle is determined according to the identification result.

3. The intelligent assistance method based on automobile taillight information according to claim 2 is characterized in that: The real-time recognition of the preceding vehicle's taillight information according to the vehicle taillight recognition model specifically includes: Get the rear image of the front vehicle; Performing information amplification processing on the image of the rear end of the preceding vehicle to convert the image of the rear end of the preceding vehicle to a set size; Acquire a feature sequence of the image of the rear of the preceding vehicle according to a preset neural network, wherein the feature sequence is the position of the taillight, the radius of a circle containing a single light, and a middle straight line passing through the center of the circle containing the single light; The taillight position is multiplied by the middle straight line to obtain a position to be identified, and each taillight position is divided according to the position to be identified, and a front vehicle taillight image is obtained according to a preset algorithm; The preceding vehicle taillight image is recognized in real time according to the vehicle taillight recognition model.

4. The intelligent assistance method based on automobile taillight information according to claim 3 is characterized in that: The preset neural network includes a first convolutional neural network, which is used to classify the picture of the rear of the preceding vehicle, and the input end of the first convolutional neural network is a picture of arbitrary size. The first convolutional neural network restores the output feature sequence to the picture size of the input end through a transposed convolution layer, and saves the position data of the picture of the initial input end.

5. The intelligent assistance method based on automobile taillight information according to claim 4 is characterized in that: The preset neural network includes a second convolutional neural network, and the second convolutional neural network is established in the following manner: The input end of the second convolutional neural network is set as the output end of the first convolutional neural network, and a bottom-up network structure is established; Establish a bottom-up network structure corresponding to the top-down network structure and perform deconvolution processing; According to the dimensionality reduction algorithm, the next layer of the network is operated through a 1×1 convolution kernel, the next layer and the last layer are operated element by element, and a 3×3 convolution smoothing process is performed; Repeat the previous step to establish the second convolutional neural network.

6. An intelligent assistance system based on automobile taillight information, characterized in that: include: A collection module, the collection module is used to collect car taillight pictures, the car taillight pictures include taillight pictures of each type of vehicle and taillight pictures of the vehicle when braking, turning and driving normally, and classify and annotate the car taillight pictures to build a car taillight picture information library; A training module, wherein the training module is used to construct a car taillight detection model, and input the classified and annotated 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; A recognition module, the recognition module is used to obtain the preceding vehicle's taillight information in real time through the acquisition device of the auxiliary vehicle, and to recognize the preceding vehicle's taillight information in real time according to the automobile taillight recognition model; An intelligent assistance module is used to determine the driving state of the preceding vehicle according to the recognition result so as to provide intelligent assistance to the driving of the auxiliary vehicle.

7. The intelligent assistance system based on automobile taillight information according to claim 6 is characterized in that: When the taillight of the preceding vehicle is partially damaged or contaminated, the vehicle type of the preceding vehicle is identified through the vehicle identification model, and the taillight of the preceding vehicle is identified according to the automobile taillight image information library, and the driving state of the preceding vehicle is determined according to the identification result.

8. The intelligent assistance system based on automobile taillight information according to claim 7 is characterized in that: The identification module is specifically used for: Get the rear image of the front vehicle; Performing information amplification processing on the image of the rear end of the preceding vehicle to convert the image of the rear end of the preceding vehicle to a set size; Acquire a feature sequence of the image of the rear of the preceding vehicle according to a preset neural network, wherein the feature sequence is the position of the taillight, the radius of a circle containing a single light, and a middle straight line passing through the center of the circle containing the single light; The taillight position is multiplied by the middle straight line to obtain a position to be identified, and each taillight position is divided according to the position to be identified, and a front vehicle taillight image is obtained according to a preset algorithm; The preceding vehicle taillight image is recognized in real time according to the vehicle taillight recognition model.

9. The intelligent assistance system based on automobile taillight information according to claim 8, characterized in that: The preset neural network includes a first convolutional neural network, which is used to classify the picture of the rear of the preceding vehicle, and the input end of the first convolutional neural network is a picture of arbitrary size. The first convolutional neural network restores the output feature sequence to the picture size of the input end through a transposed convolution layer, and saves the position data of the picture of the initial input end.

10. The intelligent assistance system based on automobile taillight information according to claim 9, characterized in that: The preset neural network includes a second convolutional neural network, and the second convolutional neural network is established in the following manner: The input end of the second convolutional neural network is set as the output end of the first convolutional neural network, and a bottom-up network structure is established; Establish a bottom-up network structure corresponding to the top-down network structure and perform deconvolution processing; According to the dimensionality reduction algorithm, the next layer of the network is operated through a 1×1 convolution kernel, the next layer and the last layer are operated element by element, and a 3×3 convolution smoothing process is performed; Repeat the previous step to establish the second convolutional neural network.

Images