Road sign recognition method and apparatus, vehicle, and storage medium

Abstract

A road sign recognition method and apparatus, a vehicle, and a storage medium. The method comprises: acquiring a road environment image, which is captured in real time by an image collection apparatus with respect of a road environment where a vehicle is located, and vehicle position information of the vehicle that is collected by a positioning sensor in real time (S101); extracting a road sign to be recognized that is contained in the road environment image, and determining whether the road sign to be recognized meets a preset determination condition (S102); if the road sign to be recognized meets the preset determination condition, using, as a target recognition result of the road sign to be recognized, the meaning and type of a known road sign matching the road sign to be recognized (S103); and if the road sign to be recognized does not meet the preset determination condition, on the basis of the road sign to be recognized and the vehicle position information combined with a pre-trained vision-language model, determining the target recognition result of the road sign to be recognized (S104).

Description

Road sign recognition method and device, vehicle and storage medium

[0001] The present application claims priority to the Chinese patent application No. 202410862027.5, filed on June 28, 2024, with the Chinese Patent Office, the content of which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[0002] The present application relates to the field of artificial intelligence, for example to a road sign recognition method and device, vehicle and storage medium. BACKGROUND

[0003] Different countries have different road signs, and infrequent road signs and their meanings are confusing and easily confused, which can make driving in other countries very challenging. Autonomous and semi-autonomous vehicles can have difficulty recognizing unfamiliar road signs because they are not pre-trained with machine language algorithms. When these infrequent road signs appear, semi-autonomous vehicles should transfer control to the driver. However, the driver can also need to rely on tools such as Google for identification, and this reliance on external resources can introduce interference and increase the risk of accidents. In the case of fully autonomous vehicles without drivers, they must stop when encountering such unfamiliar road signs, which can cause traffic flow to be interrupted and other road users to be inconvenienced. SUMMARY

[0004] The embodiments of the present application provide a road sign recognition method, device, vehicle and storage medium, which realizes the recognition of infrequent road signs and provides protection for the safe driving of autonomous vehicles.

[0005] In a first aspect, the embodiments of the present application provide a road sign recognition method, which comprises:

[0006] obtaining a road environment image captured by an image acquisition device in real time from a road environment in which a vehicle is located and vehicle position information collected by a positioning sensor in real time from the vehicle;

[0007] extracting a road sign to be recognized contained in the road environment image, and determining whether the road sign to be recognized meets a preset determination condition, the preset determination condition being that the road sign to be recognized matches a known road sign and a confidence level is greater than or equal to a set confidence level threshold;

[0008] if the preset determination condition is met, taking the meaning and type of the known road sign matching the road sign to be recognized as a target recognition result of the road sign to be recognized;

[0009] If the preset judgment condition is not met, a target recognition result of the to-be-recognized road sign is determined according to the to-be-recognized road sign and the vehicle position information, in combination with a pre-trained visual language model.

[0010] In a second aspect, the embodiment provides a road sign recognition device, which comprises:

[0011] an image acquisition module configured to acquire a road environment image captured by an image collection device in real time and vehicle position information collected by a positioning sensor in real time;

[0012] a judgment module configured to extract a to-be-recognized road sign included in the road environment image, and judge whether the to-be-recognized road sign meets a preset judgment condition, the preset judgment condition being that the to-be-recognized road sign matches a known road sign and a confidence degree is greater than or equal to a set confidence degree threshold;

[0013] a first determination module configured to, if the preset judgment condition is met, take a meaning and a type of the known road sign matching the to-be-recognized road sign as a target recognition result of the to-be-recognized road sign;

[0014] a second determination module configured to, if the preset judgment condition is not met, determine the target recognition result of the to-be-recognized road sign according to the to-be-recognized road sign and the vehicle position information, in combination with a pre-trained visual language model.

[0015] In a third aspect, the embodiment provides a vehicle, which comprises:

[0016] a vehicle body;

[0017] an image collection device and a positioning sensor, which are respectively installed on the vehicle body;

[0018] a controller in communication connection with the image collection device and the positioning sensor, the controller comprising:

[0019] at least one processor; and

[0020] a memory in communication connection with the at least one processor; wherein

[0021] the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the road sign recognition method according to any one of the embodiments.

[0022] In a fourth aspect, the embodiment provides a computer readable storage medium, and the computer program is executed by the at least one processor to enable the at least one processor to perform the road sign recognition method in any embodiment of the present application.

[0023] The embodiment of the present application provides a road sign recognition method, device, vehicle and storage medium. The method comprises the following steps: firstly, acquiring a road environment image captured by an image acquisition device in real time and vehicle position information collected by a positioning sensor in real time; then, extracting a to-be-recognized road sign in the road environment image, and determining whether the to-be-recognized road sign meets a preset determination condition, wherein the preset determination condition is that the to-be-recognized road sign matches a known road sign and the confidence is greater than or equal to a set confidence threshold; if the preset determination condition is met, the meaning and type of the known road sign matching the to-be-recognized road sign are taken as a target recognition result of the to-be-recognized road sign; and if the preset determination condition is not met, the target recognition result of the to-be-recognized road sign is determined according to the to-be-recognized road sign and the vehicle position information, and in combination with a pre-trained visual language model. The above technical solution can capture the road environment image in real time during vehicle driving, detect the road sign, if the detected road sign is a known road sign, the meaning and type of the road sign can be directly acquired, if the detected road sign is an uncommon road sign, the to-be-recognized road sign can be processed based on the natural language processing function of the visual language model, and the to-be-recognized road sign is classified based on the processing result of the visual language model to obtain the meaning and type of the to-be-recognized road sign. The technical solution provides a control basis for an automatic driving vehicle or a semi-automatic driving vehicle, improves the driving safety, and has good practical value. BRIEF DESCRIPTION OF DRAWINGS

[0024] FIG. 1A is a flowchart of a road sign recognition method provided by an embodiment of the present application;

[0025] FIG. 1B is an example diagram of a plurality of uncommon road signs in an embodiment;

[0026] FIG. 2A is a flowchart of another road sign recognition method provided by an embodiment of the present application;

[0027] FIG. 2B is an example diagram of an uncommon road sign in a road sign recognition method provided by an embodiment of the present application;

[0028] FIG. 2C is an example diagram of a road sign recognition method executed in an application scenario provided by an embodiment of the present application;

[0029] FIG. 3 is a structural diagram of a road sign recognition device provided by an embodiment of the present application;

[0030] FIG. 4 is a structural schematic diagram of a vehicle according to an embodiment of the present application. DETAILED DESCRIPTION

[0031] In order to make the personnel in the technical field better understand the scheme of the present application, the technical scheme in the embodiments of the present application will be clearly and completely described below in combination with the drawings in the embodiments of the present application. The described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by the person of ordinary skill in the art without creative labor should belong to the scope of protection of the present application.

[0032] The terms "original", "target" and the like in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0033] Embodiment One

[0034] FIG. 1A is a flowchart of a road sign recognition method according to an embodiment of the present application. The method can be applied to the automatic recognition of unusual road signs. The method can be executed by a road sign recognition device, which can be realized in the form of hardware and / or software, and is generally integrated in a vehicle.

[0035] Different countries have different road signs, and some signs are not common. For example, FIG. IB is an example of a plurality of uncommon road signs in an embodiment, as shown in FIG. IB, these road signs represent a camel crossing sign, a no transport road sign, a no passing road sign, a steep cliff road sign, and a blind hill road sign. These road signs are uncommon in different regions, because they are not pre-trained to recognize them, so when an autonomous vehicle or a semi-autonomous vehicle encounters these road signs, it cannot be recognized by the machine language algorithm. When these uncommon signs appear, the semi-autonomous vehicle should hand over control to the driver. However, the driver may also be unsure of the meaning of the road sign and need to rely on search engines and other tools to identify it. This dependence on external resources can distract and increase the risk of accidents. For fully autonomous vehicles without drivers, when they encounter such unfamiliar road signs, they must stop, which can cause traffic disruption and inconvenience to other road users.

[0036] Based on this, the embodiment provides a road sign recognition method, which provides a solution based on a visual language model (VLM) that can recognize these unexpected uncommon road signs on the road. This capability is valuable for both autonomous and semi-autonomous vehicles, enabling them to react appropriately to unexpected road conditions.

[0037] As shown in FIG. 1A, the road sign recognition method provided by the embodiment can specifically include the following steps:

[0038] S101, acquiring a road environment image captured by an image acquisition device in real time and vehicle position information collected by a positioning sensor in real time.

[0039] The application scenario of the embodiment can be described as follows: when a vehicle is driving, the image of the current driving road environment is acquired in real time, and the road signs existing in the road environment are recognized, and the vehicle is controlled appropriately based on the recognized road signs. The vehicle is installed with an image acquisition device such as a camera, a video recording device, etc., and a positioning sensor such as a global positioning system (GPS) etc. During the driving of the vehicle, the image acquisition device captures the image of the road environment in real time, and the captured image is recorded as the road environment image. At the same time, the positioning sensor can locate the vehicle in real time, and the obtained positioning information is recorded as the vehicle position information. The road environment image and the vehicle position information are the basic data for subsequent road sign recognition.

[0040] S102, extracting the road sign to be recognized contained in the road environment image, and judging whether the road sign to be recognized satisfies a preset judgment condition.

[0041] The preset judgment condition is that the to-be-recognized road sign matches the known road sign and the confidence is greater than or equal to a set confidence threshold.

[0042] In this embodiment, after the road environment image is acquired, the road environment image can be detected based on the pre-trained target detection model, to determine potential objects existing in the road environment image, and to determine whether the potential objects contain road signs. If the road environment image does not contain road signs, the road sign recognition method does not need to be continuously executed. If the road environment image contains road signs, the to-be-recognized road sign is taken as a to-be-recognized road sign, and the to-be-recognized road sign is further recognized.

[0043] As described above, when it is determined that the road environment image contains road signs, the to-be-recognized road sign contained in the road environment image is extracted, and it is further determined whether the to-be-recognized road sign is a known road sign stored in advance. Specifically, the features of the to-be-recognized road sign and the features of the known road sign stored in advance are matched, and if the matching is successful, the confidence of the to-be-recognized road sign is further determined. That is, based on the matching of the features of the extracted to-be-recognized road sign and the features of the known road sign, the confidence level of the to-be-recognized road sign, that is, the confidence, is calculated. It can be considered that the more the to-be-recognized road sign matches the known road sign, the higher the confidence is; the lower the matching degree of the to-be-recognized road sign and the known road sign, the lower the confidence is. When the to-be-recognized road sign matches the known road sign, and the confidence is greater than or equal to a set confidence threshold, it is considered that the to-be-recognized road sign meets the preset judgment condition. If the features of the to-be-recognized road sign and the features of the known road sign stored in advance do not match, or the features of the to-be-recognized road sign and the features of the known road sign stored in advance match but the confidence is less than the preset confidence threshold, it is considered that the to-be-recognized road sign does not meet the preset judgment condition, and the to-be-recognized road sign needs to be further recognized.

[0044] S103, if the preset judgment condition is met, the meaning and type of the known road sign matching the to-be-recognized road sign are taken as the target recognition result of the to-be-recognized road sign.

[0045] Specifically, if the to-be-recognized road sign meets the preset judgment condition, the meaning and type of the known road sign matching the to-be-recognized road sign are taken as the target meaning and target type of the to-be-recognized road sign, and the target meaning and target type are taken as the target recognition result of the to-be-recognized road sign.

[0046] S104, if the preset judgment condition is not met, the target recognition result of the to-be-recognized road sign is determined according to the to-be-recognized road sign and the vehicle position information, in combination with the pre-trained visual language model.

[0047] In this embodiment, when it is determined that the to-be-recognized road sign does not meet the preset judgment condition, further recognition of the to-be-recognized road sign is needed. In this embodiment, the to-be-recognized road sign is recognized by using a visual language model to obtain the meaning and type of the to-be-recognized road sign as the target recognition result of the to-be-recognized road sign. For example, the type of the road sign can be a parking sign, a speed limit sign, etc., and the meaning can be no parking, or no more than 50 kilometers per hour, etc. The visual language model refers to an artificial intelligence model integrating text understanding and visual perception. These models aim to process and generate content related to text and images, achieving image captioning, visual question answering, and generating descriptions from visual input, etc. In this embodiment, the visual language model can be used to process and understand text information related to road signs. For example, processing the text description of the sign or combining the context information of natural language can enhance the overall understanding of the driving environment. For another example, understanding the natural language description of the road layout, weather conditions, or specific driving situations helps better recognize the sign.

[0048] The process of determining the target recognition result of the to-be-recognized road sign based on the pre-trained visual language model can be described as follows: extracting the text information, symbol information, and road sign color of the to-be-recognized road sign, interpreting the road sign information based on the visual language model to obtain the interpreted text of the to-be-recognized road sign; post-processing the interpreted text of the to-be-recognized road sign to obtain the post-processed interpreted text; and determining the target recognition result of the to-be-recognized road sign according to the post-processed interpreted text and the vehicle position information.

[0049] For example, continuing to refer to FIG. 1A, for the first road sign in FIG. 1A, the first road sign is processed by using the visual language model to obtain the following text interpretation: This picture shows a road sign with a silhouette of a camel on it, indicating that there may be camels crossing the road in this area. When you see this sign, you should be careful, slow down, and pay attention to any camels that may be crossing the road to avoid accidents. The presence of real camels in the background reinforces the reminder in this area. The road sign can be classified based on the interpreted text to determine its meaning and type.

[0050] The technical solution can enhance safety: accurate identification of unusual signs helps autonomous or semi-autonomous vehicles respond appropriately to unique traffic situations. Avoid traffic violations: identifying and understanding unusual road signs enables vehicles to comply with traffic rules and regulations, reducing the risk of traffic violations and related legal consequences. Efficient traffic flow: correct identification of unusual road signs enables vehicles to make informed decisions, contributing to smoother traffic flow, reducing congestion, and improving the overall traffic system. Prevent accidents: unusual road signs may indicate potential hazards or changes in road conditions, and timely identification of these signs enables vehicles to take preventive measures, reducing the likelihood of accidents. Adapt to different environments: different regions may have unique road signs that are unfamiliar to vehicles not accustomed to these areas. Identifying unusual signs enables vehicles to seamlessly adapt to different environments. Improved navigation: autonomous vehicles with the ability to identify unusual road signs can enhance navigation capabilities. This ensures accurate route planning and minimizes the risk of getting lost or taking the wrong route. Reliable identification of various road signs helps build public trust in autonomous and semi-autonomous vehicle technology, which is crucial for the widespread acceptance and adoption of these technologies. Reduce dependence on human intervention: due to the ability to identify unusual road signs, autonomous vehicles can operate more independently, reducing the frequency of human driver intervention. This contributes to a smoother and more seamless autonomous driving experience. Global compatibility: visual language models can be trained on different datasets, making them more adaptable to different regions and their unique road sign variations. This global compatibility benefits autonomous vehicles operating in different geographical locations. Reduce dependence on pre-programming: visual language models can reduce the need for extensive pre-programming of specific road sign features. Their ability to understand natural language allows for a more flexible and adaptable approach to identifying various signs. Improved generalization: visual language models have the potential to generalize their understanding of road signs, making them more robust in identifying variations or unexpected elements in different driving scenarios.

[0051] The embodiment of the present application provides a road sign identification method. During vehicle driving, road environment images are captured in real time for road sign detection. If the detected road sign is a known road sign, the meaning and type of the road sign can be directly obtained. If the detected road sign is an unusual road sign, the road sign to be identified can be processed based on the natural language processing function of the visual language model, and the road sign to be identified is classified based on the processing result of the visual language model to obtain the meaning and type of the road sign to be identified. The control basis provided for the autonomous vehicle or the semi-autonomous vehicle improves the safety of driving and has good practical value.

[0052] As an optional embodiment of the present application, based on the above embodiment, the optional embodiment can optimize that after determining the target recognition result of the to-be-recognized road sign, the target recognition result of the to-be-recognized road sign is sent to the control system of the vehicle to make the vehicle perform corresponding operations.

[0053] In the present embodiment, the vehicle is also equipped with a processing unit such as an electronic control unit that can support remote information processing of visual language models. The communication interface between different processing units can use a controller area network, Ethernet, etc. At the same time, the vehicle is equipped with processing units for parking, steering, and acceleration operations, which are used to control the vehicle.

[0054] Specifically, after obtaining the target recognition result of the to-be-recognized road sign, i.e., the target meaning and target type of the to-be-recognized road sign, the target meaning and target type of the to-be-recognized road sign can be sent to the control system of the vehicle to control the vehicle based on the meaning and type of the to-be-recognized road sign, so that the vehicle performs corresponding operations. For example, the decision made based on road sign recognition is integrated with other vehicle systems or electronic control units to perform appropriate operations. Such integration ensures coordination with other aspects of vehicle operation, makes decisions about taking appropriate action based on the recognized road sign, such as navigation, braking, steering, adjusting vehicle speed, changing lanes, or issuing warning signals to the driver.

[0055] The above technical solution adds the step of controlling the vehicle according to the recognition result of the road sign, which enhances safety, and accurate recognition of unusual signs helps autonomous or semi-autonomous vehicles respond appropriately to unique traffic situations. Avoiding traffic violations, recognizing and understanding unusual road signs enables the vehicle to comply with traffic rules and regulations. This reduces the risk of traffic violations and related legal consequences. Efficient traffic flow, correct identification of unusual road signs enables the vehicle to make informed decisions, contributing to smoother traffic flow. This efficiency helps reduce congestion and improve the overall traffic system. Accident prevention, unusual road signs may indicate potential hazards or changes in road conditions. Timely identification of these signs enables the vehicle to take preventive measures, thereby reducing the likelihood of accidents.

[0056] As an optional embodiment of the present application, based on the above embodiment, the optional embodiment can optimize that before extracting the to-be-recognized road sign contained in the road environment image, it further includes:

[0057] a1) pre-processing the road environment image to obtain a processed road environment image.

[0058] In this embodiment, after collecting the road environment image in real time, the road environment image is first processed to determine whether the road environment image contains road signs, so as to determine whether the identification of road signs is needed. It should be noted that the image collection device continuously collects road environment images, and the collection process of the road environment images is continuously performed. While the road environment image at the current time is processed, the road environment image at the next time can be collected.

[0059] This step is used to pre-process the road environment image, so that the road environment image is suitable for subsequent road sign identification, and a pre-processed road environment image is obtained.

[0060] As a specific implementation, the pre-processing of the road environment image to obtain the pre-processed road environment image can include feature enhancement, normalized color, and size adjustment.

[0061] Specifically, the original image data is pre-processed, including feature enhancement, normalized color, and size adjustment, to enhance image quality, remove noise, extract relevant features, and achieve subsequent efficient processing.

[0062] b1) performing target detection on the processed road environment image based on a preset target detection algorithm to obtain potential objects contained in the processed road environment image.

[0063] The target detection algorithm can be understood as a pre-trained detection algorithm, and the input of the algorithm is an image and the output is the target contained in the image. In this embodiment, the target contained in the road environment image is referred to as a potential object. Specifically, the target detection algorithm is applied to identify the potential objects in the road environment image, including road signs.

[0064] c1) If the potential objects contain road signs, the road signs are determined as the road signs to be identified.

[0065] Specifically, if the potential objects contained in the road environment image contain road signs, the road signs are determined as the road signs to be identified, and information about the detected road signs, such as their positions and sizes, will be extracted. The road signs to be identified need to be further identified.

[0066] d1) If the potential objects do not contain road signs, it is determined that the road environment image does not need to be identified for road signs.

[0067] Specifically, if the potential objects contained in the road environment image do not contain road signs, the road environment image does not need to be further identified for road signs.

[0068] The technical solution increases the step of judging whether the road environment image contains a road sign, thereby providing a basis for whether to perform road sign recognition.

[0069] Embodiment Two

[0070] FIG. 2A is a flowchart of another road sign recognition method provided by Embodiment Two of the present application. Embodiment Two is a further optimization of Embodiment One. In Embodiment Two, the step of "determining a target recognition result of the road sign to be recognized according to the road sign to be recognized and the vehicle position information, in combination with a pre-trained visual language model" is further optimized, and the step of "judging whether the road sign to be recognized meets a preset judgment condition" is further optimized.

[0071] As shown in FIG. 2A, Embodiment Two provides a road sign recognition method, which specifically includes the following steps:

[0072] S201: Obtain a road environment image captured by an image capture device in real time and vehicle position information collected by a positioning sensor in real time.

[0073] In this embodiment, the image capture device is used as an input device to capture real-time images of the road environment, including road signs, and provide continuous updates of the road scene. The vehicle position information collected by the positioning sensor is used to supplement the image data.

[0074] S202: Extract a road sign to be recognized from the road environment image.

[0075] Specifically, a road sign contained in the road environment image is extracted, which is referred to as a road sign to be recognized.

[0076] S203: Extract feature information of the road sign to be recognized.

[0077] Specifically, feature information of the road sign to be recognized is extracted.

[0078] S204: Compare the feature information of the road sign to be recognized with feature information of a known road sign to obtain a comparison result and a confidence level of the comparison result.

[0079] In this embodiment, a known road sign is pre-stored. The feature information of the road sign to be recognized is compared with the feature information of the known road sign to obtain a comparison result. The confidence level of the comparison result is calculated based on the matching of the feature information of the road sign to be recognized and the feature information of the known road sign.

[0080] S205, if the comparison result is that the feature information of the to-be-identified road sign matches the feature information of the known road sign, and the confidence is greater than or equal to the set confidence threshold, it is determined that the to-be-identified road sign meets the preset judgment condition. Continue to execute step S207.

[0081] Specifically, the feature information of the to-be-identified road sign is compared with the feature information of the known road sign. If the feature information of the to-be-identified road sign matches the feature information of the known road sign, and the confidence is greater than or equal to the set confidence threshold, it is determined that the to-be-identified road sign meets the preset judgment condition.

[0082] S206, otherwise, it is determined that the to-be-identified road sign does not meet the preset judgment condition. Continue to execute step S208.

[0083] Specifically, the feature information of the to-be-identified road sign is compared with the feature information of the known road sign. If the feature information of the to-be-identified road sign does not match the feature information of the known road sign, or the feature information of the to-be-identified road sign matches the feature information of the known road sign but the confidence is less than the set confidence threshold, it is determined that the to-be-identified road sign does not meet the preset judgment condition.

[0084] S207, if the preset judgment condition is met, the meaning and type of the known road sign matched with the to-be-identified road sign are taken as the target recognition result of the to-be-identified road sign.

[0085] Specifically, if the comparison result meets the preset judgment condition, the meaning and type of the known road sign matched with the to-be-identified road sign are taken as the target meaning and target type of the to-be-identified road sign, and the target meaning and target type are taken as the target recognition result.

[0086] S208, if the preset judgment condition is not met, the road sign information of the to-be-identified road sign is extracted.

[0087] The road sign information at least includes text information, symbol information, and road sign color.

[0088] Specifically, if the comparison result does not meet the preset judgment condition, the to-be-identified road sign needs to be further identified based on the visual language model. This step is used to extract the road sign information of the to-be-identified road sign, which at least includes text information, symbol information, and road sign color.

[0089] S209, the road sign information is interpreted based on the visual language model to obtain the interpretation text of the to-be-identified road sign.

[0090] In this embodiment, the visual language model reasons on the input data, using its learned knowledge to identify and interpret the textual content of the road sign. The visual language model is the core component responsible for understanding and interpreting the textual information present in the road sign, which processes the road sign information of the road sign to be identified extracted from the image to derive the meaning and context. The visual language model uses machine learning frameworks such as TensorFlow, PyTorch, or Hugging Face Transformer to interpret the textual information from the road sign. Specifically, based on the interpretation of the road sign information by the visual language model, the interpretation text of the road sign to be identified is obtained.

[0091] S210, post-processing the interpretation text of the road sign to be identified to obtain a post-processed interpretation text.

[0092] Specifically, the interpretation text of the road sign to be identified is processed by additional processing, which is recorded as post-processing, to optimize the results, eliminate noise and improve accuracy.

[0093] As a specific implementation, the post-processing of the interpretation text of the road sign to be identified to obtain a post-processed interpretation text includes: filtering, error correction, consistency check, time correlation check, map cross-reference check, proximity check and priority check on the interpretation text of the road sign to be identified to obtain a post-processed interpretation text.

[0094] Specifically, the post-processing includes: filtering out errors: deleting any false or irrelevant information that the visual language model may misinterpret. Correcting misclassification: adjusting the results to correct any errors in text recognition or interpretation. Consistency check: ensuring that the recognized text fits the expected context according to the location and type of road signs common in the area. Time correlation: adjusting the interpretation according to the time of day, which can affect the relevance of certain signs (such as school zones). Cross-reference with map: verifying the detected signs against a digital map to confirm their existence and relevance at a given location. Proximity check: ensuring that the detected signs are relevant to the current location and route of the vehicle. Priority check: ranking the importance of the detected signs to determine which signs require immediate action.

[0095] S211, determining the target recognition result of the road sign to be identified according to the post-processed interpretation text and the vehicle location information.

[0096] In this embodiment, the post-processing explanation text is classified to determine the type of the road sign and its related meaning. At the same time, the vehicle position data and the digital map can also be referred to. Based on the post-processing explanation text of the road sign to be recognized, combined with the vehicle position information, i.e. combined with the context information of natural language, the overall understanding of the driving environment can be enhanced. For example, understanding the road layout, weather conditions or specific driving situations described in natural language helps to better recognize the signs.

[0097] For example, FIG. 2B is an example of an infrequent road sign in a road sign recognition method provided by Embodiment Two of the present application. As shown in FIG. 2B, it is taken as a road sign to be recognized, its text is excerpted: “50”; shape: circle; color: white background, black text and red border; and the GPS data indicating the position of the highway is obtained. Then based on the pre-trained visual language model inference: the model associates “50” with the regulatory speed symbol. It is confirmed that the circular “50” with white background, black text and red border is a standard regulatory sign. Final classification: type: regulatory sign, meaning: the maximum allowed speed is 50 km / h. After feeding the meaning and type of the road sign to the vehicle control unit, the vehicle control unit decides to limit the vehicle speed to 50 km / h.

[0098] As a specific implementation, according to the post-processing explanation text and the vehicle position information, the target recognition result of the road sign to be recognized is determined, including:

[0099] a2) According to the post-processing explanation text and the vehicle position information, the initial meaning and the initial type of the road sign to be recognized are determined as the initial recognition result and the confidence of the initial recognition result is determined.

[0100] Specifically, the post-processing explanation text is classified to determine the type of the road sign and its related meaning. At the same time, the vehicle position data and the digital map can also be referred to. Based on the post-processing explanation text of the road sign to be recognized, combined with the vehicle position information, i.e. combined with the context information of natural language, the overall understanding of the driving environment can be enhanced. For example, understanding the road layout, weather conditions or specific driving situations described in natural language helps to better recognize the signs. The initial meaning and the initial type of the road sign to be recognized are taken as the initial recognition result, and the confidence of the initial recognition result is calculated.

[0101] Further, the confidence of the initial recognition result is determined, including:

[0102] a21) Obtain the confidence of each processing stage in the road environment image processing process and the pre-allocated weight of each processing stage.

[0103] In this embodiment, the confidence is calculated in the visual language model analysis stage. It evaluates the accuracy and reliability of the detected and interpreted road signs from target detection algorithm, bounding box accuracy, text analysis, context verification, noise reduction, consistency check, etc. The final confidence is obtained by assigning weights to each stage according to the importance of each stage, and then calculating the weighted average. This step is used to obtain the confidence of each processing stage in the road environment image processing process and the pre-allocated weight of each processing node.

[0104] a22) Multiply the confidence of each processing stage by the corresponding pre-allocated weight and sum them up to obtain the confidence of the initial recognition result.

[0105] This step is used to calculate the weighted average to obtain the final confidence. For example, assume that the confidence scores of different stages are as follows: the confidence of the target detection algorithm stage is 0.85, the confidence of the text extraction stage is 0.90, the confidence of the text analysis stage (visual language model analysis) is 0.80, and the confidence of the context verification stage (GPS) is 0.75. Weights can be assigned to each stage according to the importance of each stage, for example: target detection: 30%, text extraction: 30%, text analysis: 20%, context verification: 20%, then the weighted average calculation of the final confidence = (0.85 x 0.30) + (0.90 x 0.30) + (0.80 x 0.20) + (0.75 x 0.20) = 0.835.

[0106] The above technical solution specifically realizes the visual language model detection stage and the determination step of the confidence of the initial recognition result.

[0107] b2) If the confidence of the initial recognition result is less than the set confidence threshold, it is determined that the to-be-recognized road sign cannot be recognized.

[0108] The confidence threshold can be set according to actual experience, and the threshold is set to ensure that the classification confidence is high enough. For example, the confidence threshold can be set to 80%. Specifically, if the confidence of the initial recognition result is less than the set confidence threshold, it is determined that the method provided in this embodiment cannot accurately determine the to-be-recognized road sign, i.e., it is determined that the road sign cannot be recognized.

[0109] For example, if the confidence is lower than the threshold (40%), the system may warn the driver and hand over control of the vehicle to the driver for semi-autonomous vehicles. In the case of an autonomous vehicle, the vehicle is pulled to an emergency stop. A confidence level higher than 80% is considered high and can continue to classify road signs and make decisions.

[0110] c2) if the confidence of the initial recognition result is greater than or equal to a set confidence threshold, the initial recognition result is taken as the target recognition result.

[0111] Specifically, if the confidence of the initial recognition result is greater than or equal to a set confidence threshold, it is considered that the recognition result is relatively accurate, and the initial recognition result can be taken as the target recognition result.

[0112] The above technical solution specifically implements the determination step of the confidence of the initial recognition result, further discriminates the accuracy of the initial recognition result, and guarantees the accuracy of the to-be-recognized road sign.

[0113] As an optional embodiment of the present application, on the basis of the above embodiment, after it is determined that the to-be-recognized road sign cannot be recognized, the method further includes: controlling the vehicle to stop urgently and issuing an alarm prompt.

[0114] Specifically, if it is determined that the to-be-recognized road sign cannot be recognized, the vehicle needs to be controlled to stop urgently, and an alarm prompt is issued. The above technical solution increases the control step of urgent stopping, and guarantees the safety of the vehicle when the road sign is unclear.

[0115] In order to more clearly describe the execution process of the road sign recognition method provided by the present application, the execution process of the road sign recognition method provided by the present application is described by taking the recognition of a road sign in a certain actual application scenario as an example. Exemplarily, FIG. 2C is a flowchart of the execution of the road sign recognition method in a certain application scenario provided by the second embodiment of the present application, as shown in FIG. 2C, the execution steps of the road sign recognition method specifically include:

[0116] S1, acquiring a road environment image captured by an image acquisition device in real time and vehicle position information collected by a positioning sensor in real time.

[0117] S2, performing feature enhancement, normalized color and size adjustment processing on the road environment image to obtain a preprocessed road environment image.

[0118] S3, performing target detection on the processed road environment image based on a preset target detection algorithm to obtain potential objects contained in the processed road environment image.

[0119] S4, if the potential objects contain a road sign, the road sign is determined as a to-be-recognized road sign. Jump to step S6.

[0120] S5, if the potential objects do not contain a road sign, it is determined that the road sign recognition on the road environment image is not needed.

[0121] S6, extracting the to-be-recognized road sign contained in the road environment image.

[0122] S7, extract feature information of the to-be-recognized road sign.

[0123] S8, compare the feature information of the to-be-recognized road sign with the feature information of the known road sign, obtain a comparison result and a confidence level of the comparison result.

[0124] S9, if the comparison result is that the feature information of the to-be-recognized road sign matches the feature information of the known road sign, and the confidence level is greater than or equal to a set confidence threshold, it is determined that the to-be-recognized road sign meets a preset judgment condition. Jump to step S11.

[0125] S10, otherwise, it is determined that the to-be-recognized road sign does not meet the preset judgment condition. Jump to step S12.

[0126] S11, if the preset judgment condition is met, the meaning and type of the known road sign matching the to-be-recognized road sign are taken as the target recognition result of the to-be-recognized road sign. Jump to S18.

[0127] S12, if the preset judgment condition is not met, extract road sign information of the to-be-recognized road sign.

[0128] S13, interpret the road sign information based on a visual language model to obtain an interpreted text of the to-be-recognized road sign.

[0129] S14, filter, correct, consistency check, time correlation check, map cross-reference check, proximity check and priority check on the interpreted text of the to-be-recognized road sign to obtain a post-processed interpreted text.

[0130] S15, according to the post-processed interpreted text and vehicle position information, determine an initial meaning and an initial type of the to-be-recognized road sign as an initial recognition result and determine a confidence level of the initial recognition result.

[0131] S16, if the confidence level of the initial recognition result is less than a set confidence threshold, it is determined that the to-be-recognized road sign cannot be recognized. Jump to S19.

[0132] S17, if the confidence level of the initial recognition result is greater than or equal to the set confidence threshold, the initial recognition result is taken as the target recognition result. Jump to S18.

[0133] S18, send the target recognition result of the to-be-recognized road sign to the control system of the vehicle to make the vehicle perform corresponding operations.

[0134] S19, control the vehicle to stop urgently and issue an alarm prompt.

[0135] Embodiment three

[0136] Fig. 3 is a structural schematic diagram of a road sign recognition device provided by Embodiment Three of the present application, which can be applied to the case of automatically recognizing infrequent road signs. The road sign recognition device can be realized in the form of hardware and / or software and is generally integrated in a vehicle. As shown in Fig. 3, the system comprises an image acquisition module 31, a judgment module 32, a first determination module 33 and a second determination module 34. Wherein,

[0137] The image acquisition module 31 is configured to acquire a road environment image captured by an image acquisition device in real time and vehicle position information collected by a positioning sensor in real time.

[0138] The judgment module 32 is configured to extract a to-be-recognized road sign contained in the road environment image and judge whether the to-be-recognized road sign satisfies a preset judgment condition, the preset judgment condition being that the to-be-recognized road sign matches a known road sign and a confidence degree is greater than or equal to a set confidence threshold.

[0139] The first determination module 33 is configured to, if the preset judgment condition is satisfied, take the meaning and type of the known road sign matching the to-be-recognized road sign as a target recognition result of the to-be-recognized road sign.

[0140] The second determination module 34 is configured to, if the preset judgment condition is not satisfied, determine the target recognition result of the to-be-recognized road sign according to the to-be-recognized road sign and the vehicle position information and in combination with a pre-trained visual language model.

[0141] The above technical solution can capture a road environment image in real time during vehicle driving for road sign detection. If the detected road sign is a known road sign, the meaning and type of the road sign can be directly acquired. If the detected road sign is an infrequent road sign, the to-be-recognized road sign can be processed based on the natural language processing function of the visual language model, and the to-be-recognized road sign is classified based on the processing result of the visual language model to obtain the meaning and type of the to-be-recognized road sign. The control basis is provided for an autonomous vehicle or a semi-autonomous vehicle, the safety of driving is improved, and the technical solution has good practical value.

[0142] Optionally, the second determination module 34 comprises:

[0143] An information extraction unit is configured to extract road sign information of the to-be-recognized road sign, the road sign information at least comprising text information, symbol information and road sign color;

[0144] A text determination unit is configured to interpret the road sign information based on the visual language model to obtain an interpretation text of the to-be-recognized road sign.

[0145] The post-processing unit is configured to post-process the interpretation text of the road sign to be recognized to obtain a post-processed interpretation text.

[0146] The result determination unit is configured to determine a target recognition result of the road sign to be recognized according to the post-processed interpretation text and the vehicle position information.

[0147] Optionally, the post-processing unit is configured to:

[0148] filter, correct, consistency check, time correlation check, map cross-reference check, proximity check and priority check the interpretation text of the road sign to be recognized to obtain the post-processed interpretation text.

[0149] Optionally, the result determination unit is configured to:

[0150] determine an initial meaning and an initial type of the road sign to be recognized as an initial recognition result according to the post-processed interpretation text and the vehicle position information, and determine a confidence degree of the initial recognition result;

[0151] if the confidence degree of the initial recognition result is less than a set confidence threshold, determine that the road sign to be recognized cannot be recognized;

[0152] if the confidence degree of the initial recognition result is greater than or equal to the set confidence threshold, determine the initial recognition result as the target recognition result.

[0153] Optionally, the confidence degree of the initial recognition result is determined by:

[0154] obtaining a confidence degree of each processing stage in a road environment image processing process and a pre-allocated weight of each processing stage;

[0155] multiplying the confidence degree of each processing stage by the corresponding pre-allocated weight and summing to obtain the confidence degree of the initial recognition result.

[0156] Optionally, the judging module 32 is configured to:

[0157] extract feature information of the road sign to be recognized

[0158] compare the feature information of the road sign to be recognized with feature information of a known road sign to obtain a comparison result and a confidence degree of the comparison result;

[0159] if the comparison result is that the feature information of the road sign to be recognized matches the feature information of the known road sign and the confidence degree is greater than or equal to a set confidence threshold, determine that the road sign to be recognized meets a preset judging condition;

[0160] otherwise, determine that the road sign to be recognized does not meet the preset judging condition.

[0161] Optionally, the apparatus further comprises a target detection module, configured to:

[0162] performing preprocessing on the road environment image to obtain a preprocessed road environment image;

[0163] performing target detection on the preprocessed road environment image based on a preset target detection algorithm to obtain potential objects contained in the preprocessed road environment image;

[0164] if the potential objects contain the road sign, determining the road sign as the to-be-identified road sign;

[0165] if the potential objects do not contain the road sign, determining that the road sign identification on the road environment image is not needed.

[0166] Optionally, the preprocessing on the road environment image to obtain the preprocessed road environment image comprises:

[0167] performing feature enhancement, normalized color and size adjustment on the road environment image to obtain the preprocessed road environment image.

[0168] Optionally, the apparatus further comprises a result sending module, configured to:

[0169] send the target identification result of the to-be-identified road sign to a control system of the vehicle to make the vehicle perform corresponding operation.

[0170] Optionally, the apparatus further comprises an emergency stop control module, configured to:

[0171] control the vehicle to stop urgently and send an alarm prompt.

[0172] The road sign identification apparatus provided in the embodiments of the present application can execute the road sign identification method provided in any of the embodiments of the present application, and has the corresponding function modules and beneficial effects of the execution method.

[0173] Embodiment Four

[0174] FIG. 4 is a structural schematic diagram of a vehicle provided in Embodiment Four of the present application. The components shown herein, their connections and relationships, and their functions are taken as examples.

[0175] As shown in FIG. 4, the vehicle includes a vehicle body, an image acquisition device 20, a positioning sensor 30, and a controller 40. The image acquisition device 20 and the positioning sensor 30 are mounted on the vehicle body. The controller 40 is communicatively connected with the image acquisition device 20 and the positioning sensor 30. The controller 40 includes at least one processor 41 and a memory, such as a read-only memory (ROM) 42, a random access memory (RAM) 43, and the like, which is communicatively connected with the at least one processor 41. The memory stores a computer program that can be executed by the at least one processor 41. The processor 41 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 42 or loaded from the storage unit 48 into the random access memory (RAM) 43. Various programs and data required for the operation of the electronic device 40 can also be stored in the RAM 43. The processor 41, the ROM 42, and the RAM 43 are connected with each other through a bus 44. An input / output (I / O) interface 45 is also connected to the bus 44.

[0176] Various components in the electronic device 40 are connected to the I / O interface 45, including an input unit 46, such as a keyboard, a mouse, and the like, an output unit 47, such as various types of displays, a speaker, and the like, a storage unit 48, such as a magnetic disk, an optical disk, and the like, and a communication unit 49, such as a network card, a modem, a wireless communication transceiver, and the like. The communication unit 49 allows the electronic device 40 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0177] The processor 41 can be various general and / or special purpose processing components having processing and computing capabilities. The processor 41 can include a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, and the like. The processor 41 performs various methods and processes described above, such as the road sign recognition method.

[0178] In some embodiments, the road sign recognition method can be implemented as a computer program tangibly embodied in a computer readable storage medium, such as the storage unit 48. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 40 via the ROM 42 and / or the communication unit 49. When the computer program is loaded into the RAM 43 and executed by the processor 41, one or more steps of the road sign recognition method described above can be performed. Alternatively, in other embodiments, the processor 41 can be configured to perform the road sign recognition method by any other appropriate means, such as by means of firmware.

[0179] The various embodiments of the systems and techniques described above can be implemented in digital electronic circuitry, integrated circuitry, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system on a chip systems (SOCs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0180] Computer programs used to implement the processes of the present application can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer program

[0181] In the context of the present application, a computer-readable storage medium can be a tangible medium that can contain or store computer programs for use by or in connection with an instruction execution system, apparatus, or device. Computer-readable storage media can include electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium will include one or more lines of a program of instructions in a transitory signal, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0182] To provide for interaction with a user, the systems and techniques described here can be implemented on a vehicle having a display device (e.g., a cathode ray tube (CRT) or liquid crystal display (LCD) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the vehicle. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0183] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0184] The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. A server can be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system, and solves the defects of large management difficulty and weak business scalability in traditional physical host and VPS service.

Claims

1. A road sign recognition method, comprising: The system acquires real-time images of the road environment captured by the image acquisition device and real-time vehicle location information collected by the positioning sensor. Extract the road signs to be identified contained in the road environment image, and determine whether the road signs to be identified meet the preset judgment conditions. The preset judgment conditions are that the road signs to be identified match the known road signs and the confidence level is greater than or equal to the set confidence level threshold. If the preset judgment condition is met, the meaning and type of the known road sign that matches the road sign to be identified will be used as the target identification result of the road sign to be identified. If the preset judgment conditions are not met, the target recognition result of the road sign to be recognized is determined based on the road sign to be recognized and the vehicle location information, combined with the pre-trained visual language model.

2. The method according to claim 1, wherein, The step of determining the target recognition result of the road sign to be recognized based on the road sign to be recognized and the vehicle location information, combined with a pre-trained visual language model, includes: Extract the road sign information of the road sign to be identified, wherein the road sign information includes at least text information, symbol information, and road sign color; The road sign information is interpreted based on the visual language model to obtain the explanatory text of the road sign to be identified; The explanatory text of the road sign to be identified is post-processed to obtain the post-processed explanatory text; Based on the post-processed interpretation text and the vehicle location information, the target recognition result of the road sign to be recognized is determined.

3. The method according to claim 2, wherein, The post-processing of the explanatory text of the road sign to be identified to obtain post-processed explanatory text includes: The explanatory text of the road sign to be identified is filtered, corrected, checked for consistency, temporal relevance, cross-referenced on maps, proxies, and prioritized to obtain the post-processed explanatory text.

4. The method according to claim 2, wherein, The step of determining the target recognition result of the road sign to be recognized based on the post-processed interpretation text and the vehicle location information includes: Based on the post-processed interpretation text and the vehicle location information, the initial meaning and initial type of the road sign to be identified are determined as the initial identification result, and the confidence level of the initial identification result is determined. If the confidence level of the initial recognition result is less than the set confidence level threshold, then it is determined that the road sign to be recognized cannot be identified; If the confidence level of the initial identification result is greater than or equal to the set confidence level threshold, then the initial identification result is taken as the target identification result.

5. The method according to claim 4, wherein, Determining the confidence level of the initial identification result includes: Obtain the confidence level of each processing stage in the road environment image processing process and the pre-assigned weight of each processing stage; The confidence level of the initial identification result is obtained by multiplying the confidence level of each processing stage by the corresponding pre-assigned weight and summing the results.

6. The method according to claim 1, wherein, The step of determining whether the road sign to be identified meets the preset judgment conditions includes: Extract the feature information of the road sign to be identified The feature information of the road sign to be identified is compared with the feature information of known road signs to obtain the comparison result and the confidence level of the comparison result; If the comparison result shows that the feature information of the road sign to be identified matches the feature information of the known road sign, and the confidence level is greater than or equal to the set confidence level threshold, then the road sign to be identified is determined to meet the preset judgment condition. Otherwise, it is determined that the road sign to be identified does not meet the preset judgment conditions.

7. The method according to claim 1, further comprising, before extracting the road signs to be identified contained in the road environment image: The road environment image is preprocessed to obtain a processed road environment image; The processed road environment image is subjected to target detection based on a preset target detection algorithm to obtain potential objects contained in the processed road environment image; If the potential object contains a road sign, then the road sign is identified as the road sign to be identified; If the potential objects do not contain road signs, then it is determined that there is no need to perform road sign recognition on the road environment image.

8. The method according to claim 7, wherein, The step of preprocessing the road environment image to obtain a preprocessed road environment image includes: The road environment image is subjected to feature enhancement, color normalization, and resizing to obtain a preprocessed road environment image.

9. The method according to claim 1, wherein after determining the target recognition result of the road sign to be recognized, the method further comprises: The target recognition result of the road sign to be recognized is sent to the vehicle's control system so that the vehicle can perform the corresponding operation.

10. The method of claim 4, further comprising, after determining that the road sign to be identified cannot be identified: Control the vehicle to stop urgently and issue an alarm.

11. A road sign recognition device, comprising: The image acquisition module is configured to acquire real-time road environment images captured by the image acquisition device and real-time vehicle position information collected by the positioning sensor. The judgment module is configured to extract the road signs to be identified contained in the road environment image and determine whether the road signs to be identified meet the preset judgment conditions. The preset judgment conditions are that the road signs to be identified match known road signs and the confidence level is greater than or equal to a set confidence level threshold. The first determining module is configured to, if the preset judgment condition is met, take the meaning and type of a known road sign that matches the road sign to be identified as the target identification result of the road sign to be identified; The second determining module is configured to determine the target recognition result of the road sign to be recognized based on the road sign to be recognized and the vehicle location information, combined with a pre-trained visual language model, if the preset judgment conditions are not met.

12. A vehicle comprising: Vehicle body; An image acquisition device and a positioning sensor are respectively installed on the vehicle body; The controller is communicatively connected to the image acquisition device and the positioning sensor, and the controller includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the road sign recognition method according to any one of claims 1-10.

13. A computer-readable storage medium storing computer instructions that, when executed by a processor, implement the road sign recognition method according to any one of claims 1-10.

Images