A traffic image segmentation control method and system

Abstract

The application discloses a kind of traffic image segmentation control method and system, it is related to image processing field, wherein the method comprises: the camera of interactive target vehicle is obtained first traffic image;According to wavelet filtering algorithm, first traffic image is carried out denoising processing, and second traffic image is obtained;Based on the second traffic image, ORB feature recognition is carried out, and ORB feature recognition result is obtained;Based on the ORB feature recognition result, clustering is carried out, and ORB feature clustering result is obtained;Based on the ORB feature clustering result, the second traffic image is carried out background isolation, and third traffic image is obtained, and, third traffic image has ORB feature mark;Traffic image segmentation channel is built;Based on the ORB feature mark, according to the traffic image segmentation channel, third traffic image is carried out segmentation control, and target traffic image segmentation result is obtained.The technical problem that present technique is low in accuracy for traffic image segmentation is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of image processing, in particular, to a traffic image segmentation control method and system. BACKGROUND

[0002] Traffic scene perception is of great significance to vehicle obstacle detection, passability estimation, path planning and the like. However, traffic scenes have the characteristics of complexity, variability and uncertainty, which makes the perception of traffic scenes more and more difficult. Traffic image segmentation is one of the important directions of traffic scene perception. There is a technical problem of low accuracy of traffic image segmentation in the prior art. SUMMARY

[0003] The present application provides a traffic image segmentation control method and system. The technical problem of low accuracy of traffic image segmentation in the prior art is solved. The technical effect of improving the accuracy of traffic image segmentation and providing reliable and effective reference data for the driving safety of vehicles is achieved.

[0004] In view of the above problems, the present application provides a traffic image segmentation control method and system.

[0005] In the first aspect, the present application provides a traffic image segmentation control method, wherein the method is applied to a traffic image segmentation control system, and the method comprises: obtaining a target vehicle; interacting with a vehicle-mounted camera of the target vehicle to obtain a first traffic image; performing noise reduction processing on the first traffic image according to a wavelet filtering algorithm to obtain a second traffic image; performing ORB feature recognition based on the second traffic image to obtain an ORB feature recognition result; performing clustering based on the ORB feature recognition result to obtain an ORB feature clustering result; performing background isolation on the second traffic image based on the ORB feature clustering result to obtain a third traffic image, and the third traffic image has an ORB feature identifier; building a traffic image segmentation channel; performing segmentation control on the third traffic image based on the ORB feature identifier and the traffic image segmentation channel to obtain a target traffic image segmentation result.

[0006] Preferably, the noise reduction processing on the first traffic image according to the wavelet filtering algorithm to obtain the second traffic image comprises:

[0007] wavelet decomposing the image signal of the first traffic image to obtain image signal wavelet coefficients;

[0008] quantifying the image signal wavelet coefficients according to a threshold value to obtain a wavelet selection threshold value;

[0009] The wavelet coefficient of the image signal is intercepted based on the wavelet selection threshold, noise signals less than or equal to the wavelet selection threshold are set to zero, and effective signal information greater than the wavelet selection threshold is obtained.

[0010] The effective signal information is filtered and reconstructed to obtain the second traffic image.

[0011] Preferably, the ORB feature recognition is performed based on the second traffic image to obtain an ORB feature recognition result, including:

[0012] According to a feature detection algorithm, key point recognition is performed on the second traffic image to obtain an image key point recognition result.

[0013] Based on a vector creation algorithm, a feature vector is constructed based on the image key point recognition result to obtain a feature vector construction result.

[0014] Based on the feature vector construction result, the ORB feature recognition result is generated.

[0015] Preferably, according to a feature detection algorithm, key point recognition is performed on the second traffic image to obtain an image key point recognition result, including:

[0016] Based on the second traffic image, pixel recognition is performed to obtain an image pixel recognition result.

[0017] The image pixel recognition result is analyzed for neighborhood pixel change degree to obtain a neighborhood pixel change degree analysis result.

[0018] A preset neighborhood pixel change degree is obtained.

[0019] It is determined whether the neighborhood pixel change degree analysis result is greater than the preset neighborhood pixel change degree.

[0020] If the neighborhood pixel change degree analysis result is greater than the preset neighborhood pixel change degree, the image key point recognition result is generated.

[0021] Preferably, based on the ORB feature clustering result, the second traffic image is isolated from the background to obtain a third traffic image, including:

[0022] Based on the ORB feature clustering result, a plurality of center ORB features are obtained.

[0023] Based on the plurality of center ORB features, a center ORB feature curve is obtained by curve fitting.

[0024] Based on a preset ORB feature, the center ORB feature curve is positioned to obtain a background center ORB feature.

[0025] obtain a background ORB feature clustering result according to the ORB feature clustering result;

[0026] perform background isolation on the second traffic image based on the background ORB feature clustering result, and obtain the third traffic image.

[0027] Preferably, the traffic image segmentation channel is built, and the traffic image segmentation channel comprises:

[0028] obtain a traffic image segmentation record set;

[0029] divide the traffic image segmentation record set based on a preset division operator, and obtain a training data sequence, a test data sequence and a verification data sequence;

[0030] train, test and verify based on the full connection neural network according to the training data sequence, the test data sequence and the verification data sequence, and generate the traffic image segmentation channel.

[0031] In a second aspect, the present application further provides a traffic image segmentation control system, wherein the system comprises: a vehicle obtaining module, the vehicle obtaining module being used to obtain a target vehicle; a camera interacting module, the camera interacting module being used to interact a vehicle-mounted camera of the target vehicle, and obtain a first traffic image; a noise reduction processing module, the noise reduction processing module being used to perform noise reduction processing on the first traffic image according to a wavelet filtering algorithm, and obtain a second traffic image; an ORB feature recognition module, the ORB feature recognition module being used to perform ORB feature recognition based on the second traffic image, and obtain an ORB feature recognition result; an ORB feature clustering module, the ORB feature clustering module being used to perform clustering based on the ORB feature recognition result, and obtain an ORB feature clustering result; an image background isolation module, the image background isolation module being used to perform background isolation on the second traffic image based on the ORB feature clustering result, and obtain a third traffic image, and the third traffic image has an ORB feature identifier; a building module, the building module being used to build a traffic image segmentation channel; and a segmentation control module, the segmentation control module being used to perform segmentation control on the third traffic image based on the ORB feature identifier and according to the traffic image segmentation channel, and obtain a target traffic image segmentation result.

[0032] In a third aspect, the present application further provides an electronic device, comprising: a memory, used to store executable instructions; and a processor, used to execute the executable instructions stored in the memory, and realize a traffic image segmentation control method provided by the present application.

[0033] In a fourth aspect, the present application provides a computer readable storage medium storing a computer program, which, when executed by a processor, implements the traffic image segmentation control method provided by the present application.

[0034] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0035] The first traffic image is obtained through the on-board camera of the target vehicle, the first traffic image is denoised through a wavelet filtering algorithm to obtain a second traffic image, ORB feature recognition is performed according to the second traffic image to obtain an ORB feature recognition result, the ORB feature recognition result is clustered to obtain an ORB feature clustering result, the second traffic image is isolated from the background according to the ORB feature clustering result to obtain a third traffic image, and the third traffic image has an ORB feature identifier, and the third traffic image is segmented and controlled according to the traffic image segmentation channel based on the ORB feature identifier to obtain a target traffic image segmentation result. The technical effect of improving the accuracy of traffic image segmentation is achieved, and reliable and effective reference data for the driving safety of the vehicle is provided.

[0036] The above description is only a summary of the technical solutions of the present application. In order to more clearly understand the technical means of the present application, the following detailed description can be implemented according to the content of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described. BRIEF DESCRIPTION OF DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings of the embodiments of the present application. Obviously, the drawings described in the following description only relate to some embodiments of the present application, but not limit the present application.

[0038] Figure 1 The flowchart of the traffic image segmentation control method of the present application is shown in the figure;

[0039] Figure 2 The flowchart of obtaining the ORB feature recognition result in the traffic image segmentation control method of the present application is shown in the figure;

[0040] Figure 3 The structure diagram of the traffic image segmentation control system of the present application is shown in the figure;

[0041] Figure 4 The structure diagram of the exemplary electronic device of the present application is shown in the figure.

[0042] Explanation of reference signs: processor 31, memory 32, input device 33, output device 34. DETAILED DESCRIPTION

[0043] The application provides a traffic image segmentation control method and system. The technical problem of low traffic image segmentation accuracy in the prior art is solved. The technical effect of improving the traffic image segmentation accuracy and providing reliable and effective reference data for vehicle driving safety is achieved.

[0044] Embodiment one

[0045] Please refer to the accompanying Figure 1 The application provides a traffic image segmentation control method, which is applied to a traffic image segmentation control system and specifically includes the following steps:

[0046] Obtaining a target vehicle;

[0047] Interacting with the vehicle-mounted camera of the target vehicle to obtain a first traffic image;

[0048] Connecting the vehicle-mounted camera of the target vehicle, collecting real-time images of the target vehicle according to the vehicle-mounted camera, and obtaining a first traffic image. The target vehicle can be any vehicle. The first traffic image includes real-time image information of various participating objects such as other vehicles, pedestrians, and road surfaces corresponding to the target vehicle.

[0049] According to the wavelet filtering algorithm, the first traffic image is subjected to noise reduction processing to obtain a second traffic image;

[0050] According to the wavelet filtering algorithm, the first traffic image is subjected to noise reduction processing to obtain a second traffic image;

[0051] The image signal of the first traffic image is subjected to wavelet decomposition to obtain image signal wavelet coefficients;

[0052] According to the image signal wavelet coefficients, threshold quantization is performed to obtain wavelet selection thresholds;

[0053] Based on the wavelet selection thresholds, the image signal wavelet coefficients are intercepted, noise signals less than / equal to the wavelet selection thresholds are set to zero, and effective signal information greater than the wavelet selection thresholds is obtained;

[0054] The effective signal information is subjected to filtering reconstruction to obtain the second traffic image.

[0055] Preferably, the application adopts a wavelet filtering algorithm to perform noise reduction processing on the first traffic image. The wavelet filtering algorithm is a noise reduction method based on wavelet transform, which can decompose an image into wavelet coefficients at different scales and directions, and perform threshold processing on these coefficients to eliminate noise. Using wavelet decomposition, the image signal is decomposed into each scale, the wavelet coefficients belonging to noise at each scale are removed, the wavelet coefficients belonging to signal are retained and enhanced, and finally the processed wavelet coefficients are reconstructed to obtain a denoised image using inverse wavelet transform. After wavelet transform of the image signal of the first traffic image, the image signal wavelet coefficients generated by the signal contain important information of the signal. After wavelet decomposition of the image signal, the wavelet coefficients of the noise are smaller, and the wavelet coefficients of the noise are smaller than the wavelet coefficients of the signal. By selecting a suitable threshold, the wavelet coefficients greater than the threshold are considered to be generated by the signal and should be retained, and the wavelet coefficients less than / equal to the threshold are considered to be generated by noise and are set to zero, thereby achieving the purpose of denoising and obtaining a second traffic image.

[0056] Based on the second traffic image, ORB feature recognition is performed to obtain an ORB feature recognition result.

[0057] As shown in the accompanying drawings, Figure 2 Based on the second traffic image, ORB feature recognition is performed to obtain an ORB feature recognition result, including:

[0058] According to a feature detection algorithm, key point recognition is performed on the second traffic image to obtain an image key point recognition result.

[0059] Based on a vector creation algorithm, feature vector construction is performed on the image key point recognition result to obtain a feature vector construction result.

[0060] Based on the feature vector construction result, the ORB feature recognition result is generated.

[0061] According to a feature detection algorithm, key point recognition is performed on the second traffic image to obtain an image key point recognition result, including:

[0062] Based on the second traffic image, pixel recognition is performed to obtain an image pixel recognition result.

[0063] The image pixel recognition result is analyzed for neighborhood pixel change degree to obtain a neighborhood pixel change degree analysis result.

[0064] A preset neighborhood pixel change degree is obtained.

[0065] It is determined whether the neighborhood pixel change degree analysis result is greater than the preset neighborhood pixel change degree.

[0066] If the neighborhood pixel change degree analysis result is greater than the preset neighborhood pixel change degree, the image key point recognition result is generated.

[0067] An image pixel recognition result is obtained by performing pixel recognition on the second traffic image. The image pixel recognition result includes a pixel value corresponding to each pixel point in the second traffic image. Then, neighborhood pixel change degree analysis is performed on the image pixel recognition result to obtain a neighborhood pixel change degree analysis result. The neighborhood pixel change degree analysis result includes a plurality of neighborhood pixel change degrees. The neighborhood pixel change degree is data information for representing the degree of change in pixel values between adjacent pixel points in the image pixel recognition result. The higher the degree of change in pixel values between adjacent pixel points, the greater the corresponding neighborhood pixel change degree.

[0068] Further, it is judged whether each neighborhood pixel change degree in the neighborhood pixel change degree analysis result is greater than the preset neighborhood pixel change degree. If the neighborhood pixel change degree is greater than the preset neighborhood pixel change degree, the adjacent pixel points corresponding to the neighborhood pixel change degree are added to the image key point recognition result. The image key point recognition result includes a plurality of adjacent pixel points corresponding to a plurality of neighborhood pixel change degrees greater than the preset neighborhood pixel change degree.

[0069] Further, according to a vector creation algorithm, a plurality of adjacent pixel points in the image key point recognition result are subjected to vector construction to obtain a feature vector construction result, and the feature vector construction result is added to the ORB feature recognition result. The vector creation algorithm is a vector construction method in the prior art. The feature vector construction result includes a vector corresponding to each adjacent pixel point. The ORB feature recognition result includes a vector corresponding to each adjacent pixel point and a pixel value corresponding to the adjacent pixel point.

[0070] Based on the ORB feature recognition result, clustering is performed to obtain an ORB feature clustering result;

[0071] The ORB feature recognition result is clustered, that is, the same vectors in the ORB feature recognition result are classified into a class to obtain an ORB feature clustering result. The ORB feature clustering result includes a plurality of ORB feature results. Each ORB feature result includes a plurality of same vectors in the ORB feature recognition result. Each same vector has a corresponding pixel value.

[0072] Based on the ORB feature clustering result, the second traffic image is subjected to background isolation to obtain a third traffic image, and the third traffic image has an ORB feature identifier;

[0073] The method further includes:

[0074] Based on the ORB feature clustering result, a plurality of center ORB features are obtained;

[0075] Based on the plurality of center ORB features, curve fitting is performed to obtain a center ORB feature curve;

[0076] Based on the center ORB feature curve, the background center ORB feature is positioned based on the preset ORB feature;

[0077] Based on the background center ORB feature, the background ORB feature clustering result is obtained according to the ORB feature clustering result;

[0078] Based on the background ORB feature clustering result, the second traffic image is background isolated to obtain the third traffic image.

[0079] A plurality of vectors corresponding to the multi-class ORB feature result are set as a plurality of center ORB features, and curve construction is performed on the plurality of center ORB features to obtain a center ORB feature curve corresponding to the plurality of center ORB features. Then, based on the preset ORB feature, a preset ORB feature curve is constructed. The preset ORB feature includes a background vector and a background pixel value determined by the one traffic image segmentation control system in advance. The center ORB feature curve and the preset ORB feature curve are compared by coincidence, and the plurality of center ORB features corresponding to the preset ORB feature curve and the preset ORB feature curve in the center ORB feature curve are set as a plurality of background center ORB features. Then, the ORB feature clustering result is data extracted according to the plurality of background center ORB features to obtain a background ORB feature clustering result, and the pixel points corresponding to the background ORB feature clustering result in the second traffic image are deleted to obtain a third traffic image. And, the third traffic image has an ORB feature identifier. The ORB feature identifier is a vector corresponding to adjacent pixel points in the third traffic image, and a pixel value corresponding to the adjacent pixel points.

[0080] A traffic image segmentation channel is built;

[0081] The traffic image segmentation channel is built, including:

[0082] A traffic image segmentation record set is obtained;

[0083] Based on a preset division operator, the traffic image segmentation record set is divided to obtain a training data sequence, a test data sequence and a verification data sequence;

[0084] Based on a full connection neural network, the training data sequence, the test data sequence and the verification data sequence are trained, tested and verified to generate the traffic image segmentation channel.

[0085] Based on the ORB feature identification, the third traffic image is executed for segmentation control according to the traffic image segmentation channel, and a target traffic image segmentation result is obtained.

[0086] Based on big data, a traffic image segmentation record set is collected. The traffic image segmentation record set includes a plurality of historical traffic image segmentation records. Each historical traffic image segmentation record includes a historical traffic image, a historical ORB feature identification, and a historical segmented traffic image. Then, a preset division operator includes a data division ratio determined by the one traffic image segmentation control system. For example, the preset division operator is 6:2:2. Then, 60% of the data information in the traffic image segmentation record set is divided into a training data sequence. 20% of the data information in the traffic image segmentation record set is divided into a test data sequence. 20% of the data information in the traffic image segmentation record set is divided into a verification data sequence. Based on the fully connected neural network, the training data sequence is continuously self-trained and learned to a convergence state, that is, the traffic image segmentation channel is obtained. The test data sequence and the verification data sequence are input into the traffic image segmentation channel, and the test data sequence and the verification data sequence are tested and verified to improve the accuracy of the traffic image segmentation channel. Then, the ORB feature identification and the third traffic image are input into the traffic image segmentation channel to obtain the target traffic image segmentation result. The traffic image segmentation accuracy is improved.

[0087] In summary, the traffic image segmentation control method provided by the application has the following technical effects:

[0088] Through the vehicle-mounted camera of the target vehicle, a first traffic image is obtained. Through a wavelet filtering algorithm, the first traffic image is denoised to obtain a second traffic image. According to the second traffic image, ORB feature recognition is performed to obtain an ORB feature recognition result, and the ORB feature recognition result is clustered to obtain an ORB feature clustering result. According to the ORB feature clustering result, the second traffic image is isolated from the background to obtain a third traffic image, and the third traffic image has an ORB feature identification. Based on the ORB feature identification, the third traffic image is executed for segmentation control according to the traffic image segmentation channel, and a target traffic image segmentation result is obtained. The traffic image segmentation accuracy is improved, and reliable and effective reference data for vehicle driving safety is provided.

[0089] Embodiment two

[0090] Based on the traffic image segmentation control method in the foregoing embodiments, the same inventive concept is provided. The application also provides a traffic image segmentation control system. Please refer to the accompanying drawings Figure 3 , the system includes:

[0091] A vehicle obtaining module is configured to obtain a target vehicle.

[0092] a camera interaction module, configured to interact with a vehicle-mounted camera of the target vehicle to obtain a first traffic image;

[0093] a noise reduction processing module, configured to perform noise reduction processing on the first traffic image according to a wavelet filtering algorithm to obtain a second traffic image;

[0094] an ORB feature recognition module, configured to perform ORB feature recognition based on the second traffic image to obtain an ORB feature recognition result;

[0095] an ORB feature clustering module, configured to perform clustering based on the ORB feature recognition result to obtain an ORB feature clustering result;

[0096] an image background isolation module, configured to perform background isolation on the second traffic image based on the ORB feature clustering result to obtain a third traffic image, and the third traffic image has an ORB feature identifier;

[0097] a building module, configured to build a traffic image segmentation channel;

[0098] a segmentation control module, configured to perform segmentation control on the third traffic image based on the ORB feature identifier and according to the traffic image segmentation channel to obtain a target traffic image segmentation result.

[0099] Further, the system further comprises:

[0100] a wavelet decomposition module, configured to perform wavelet decomposition on an image signal of the first traffic image to obtain image signal wavelet coefficients;

[0101] a threshold quantization module, configured to perform threshold quantization according to the image signal wavelet coefficients to obtain a wavelet selection threshold;

[0102] an effective signal determination module, configured to intercept the image signal wavelet coefficients based on the wavelet selection threshold, set noise signals less than / equal to the wavelet selection threshold to zero, and obtain effective signal information greater than the wavelet selection threshold;

[0103] a signal filtering and reconstruction module, configured to perform filtering and reconstruction on the effective signal information to obtain the second traffic image.

[0104] Further, the system further comprises:

[0105] The key point recognition module is configured to recognize key points in the second traffic image according to a feature detection algorithm, and obtain an image key point recognition result.

[0106] The vector construction module is configured to construct a feature vector based on the image key point recognition result according to a vector creation algorithm, and obtain a feature vector construction result.

[0107] The ORB feature recognition result generation module is configured to generate the ORB feature recognition result based on the feature vector construction result.

[0108] Further, the system further comprises:

[0109] The image pixel recognition result obtaining module is configured to recognize pixels in the second traffic image, and obtain an image pixel recognition result.

[0110] The neighborhood pixel change degree analysis module is configured to analyze neighborhood pixel change degrees of the image pixel recognition result, and obtain a neighborhood pixel change degree analysis result.

[0111] The preset neighborhood pixel change degree obtaining module is configured to obtain a preset neighborhood pixel change degree.

[0112] The change degree judgment module is configured to judge whether the neighborhood pixel change degree analysis result is greater than the preset neighborhood pixel change degree.

[0113] The first execution module is configured to generate the image key point recognition result if the neighborhood pixel change degree analysis result is greater than the preset neighborhood pixel change degree.

[0114] Further, the system further comprises:

[0115] The center ORB feature obtaining module is configured to obtain a plurality of center ORB features based on the ORB feature clustering result.

[0116] The curve fitting module is configured to perform curve fitting based on the plurality of center ORB features, and obtain a center ORB feature curve.

[0117] The feature curve positioning module is configured to position the center ORB feature curve based on a preset ORB feature, and obtain a background center ORB feature.

[0118] The second execution module is configured to obtain a background ORB feature clustering result based on the background center ORB feature and according to the ORB feature clustering result.

[0119] The third traffic image obtaining module is configured to isolate the background of the second traffic image based on the background ORB feature clustering result, and obtain the third traffic image.

[0120] Further, the system further comprises:

[0121] The segmentation record set obtaining module is configured to obtain a traffic image segmentation record set.

[0122] The record set division module is configured to divide the traffic image segmentation record set based on a preset division operator, and obtain a training data sequence, a test data sequence and a verification data sequence.

[0123] The third execution module is configured to train, test and verify based on a full connection neural network according to the training data sequence, the test data sequence and the verification data sequence, and generate the traffic image segmentation channel.

[0124] The traffic image segmentation control system provided by the embodiment of the present application can execute the traffic image segmentation control method provided by any embodiment of the present application, and has the corresponding function modules and beneficial effects of the execution method.

[0125] Each module included is only divided according to the function logic, but is not limited to the above division, as long as the corresponding function can be realized; in addition, the specific name of each function module is only for easy mutual differentiation, and is not used to limit the protection scope of the present application.

[0126] Embodiment three

[0127] Figure 4 The structural schematic diagram of the electronic device provided by the embodiment three of the present application shows the block diagram of the exemplary electronic device suitable for realizing the embodiment of the present application. Figure 4 The displayed electronic device is only an example, and should not bring any limitation to the function and use range of the embodiment of the present application. Figure 4 As shown, the electronic device comprises a processor 31, a memory 32, an input device 33 and an output device 34; the number of processors 31 in the electronic device can be one or more, Figure 4 For example, the processor 31 in the electronic device, the memory 32, the input device 33 and the output device 34 can be connected through a bus or other means, Figure 4 For example, the connection through the bus.

[0128] The memory 32 is a computer readable storage medium, which can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the traffic image segmentation control method in the embodiment of the application. The processor 31 executes the software programs, instructions and modules stored in the memory 32, thereby performing various function applications and data processing of the computer device, i.e. implementing the traffic image segmentation control method.

[0129] The application provides a traffic image segmentation control method, wherein the method is applied to a traffic image segmentation control system, and the method comprises the following steps: obtaining a first traffic image through a vehicle-mounted camera of a target vehicle; performing noise reduction processing on the first traffic image through a wavelet filtering algorithm to obtain a second traffic image; performing ORB feature recognition according to the second traffic image to obtain an ORB feature recognition result, and performing clustering on the ORB feature recognition result to obtain an ORB feature clustering result; performing background isolation on the second traffic image according to the ORB feature clustering result to obtain a third traffic image, and the third traffic image has an ORB feature identifier; performing segmentation control on the third traffic image according to a traffic image segmentation channel based on the ORB feature identifier to obtain a target traffic image segmentation result. The technical problem of low traffic image segmentation accuracy in the prior art is solved. The technical effect of improving the traffic image segmentation accuracy and providing reliable and effective reference data for improving the driving safety of vehicles is achieved.

[0130] Although the application has been described in detail through the above embodiments, the application is not limited to the above embodiments only, and can include more other equivalent embodiments without departing from the concept of the application, and the scope of the application is determined by the appended claims.

Claims

1. A traffic image segmentation control method characterized by comprising: The method comprises: obtaining a target vehicle; interacting with a vehicle-mounted camera of the target vehicle to obtain a first traffic image; performing noise reduction processing on the first traffic image according to a wavelet filtering algorithm to obtain a second traffic image; performing ORB feature recognition based on the second traffic image to obtain an ORB feature recognition result; performing clustering based on the ORB feature recognition result to obtain an ORB feature clustering result; performing background isolation on the second traffic image based on the ORB feature clustering result to obtain a third traffic image, and the third traffic image has an ORB feature identifier; building a traffic image segmentation channel; the building of the traffic image segmentation channel comprises: obtaining a traffic image segmentation record set; dividing the traffic image segmentation record set based on a preset division operator to obtain a training data sequence, a test data sequence, and a verification data sequence; training, testing, and verifying based on a fully connected neural network according to the training data sequence, the test data sequence, and the verification data sequence to generate the traffic image segmentation channel; performing segmentation control on the third traffic image based on the ORB feature identifier and the traffic image segmentation channel to obtain a target traffic image segmentation result.

2. The method of claim 1, wherein, performing noise reduction processing on the first traffic image according to a wavelet filtering algorithm to obtain a second traffic image comprises: performing wavelet decomposition on an image signal of the first traffic image to obtain image signal wavelet coefficients; performing threshold quantization according to the image signal wavelet coefficients to obtain a wavelet selection threshold; performing truncation on the image signal wavelet coefficients based on the wavelet selection threshold, setting noise signals less than or equal to the wavelet selection threshold to zero, and obtaining effective signal information greater than the wavelet selection threshold; performing filtering reconstruction on the effective signal information to obtain the second traffic image.

3. The method of claim 1, wherein, performing ORB feature recognition based on the second traffic image to obtain an ORB feature recognition result comprises: performing key point recognition on the second traffic image according to a feature detection algorithm to obtain an image key point recognition result; performing feature vector construction on the image key point recognition result based on a vector creation algorithm to obtain a feature vector construction result; generating the ORB feature recognition result based on the feature vector construction result.

4. The method of claim 3, wherein, performing key point recognition on the second traffic image according to a feature detection algorithm to obtain an image key point recognition result comprises: performing pixel recognition on the second traffic image to obtain an image pixel recognition result; performing neighborhood pixel change degree analysis on the image pixel recognition result to obtain a neighborhood pixel change degree analysis result; obtaining a preset neighborhood pixel change degree; determining whether the neighborhood pixel change degree analysis result is greater than the preset neighborhood pixel change degree; if the neighborhood pixel change degree analysis result is greater than the preset neighborhood pixel change degree, generating the image key point recognition result.

5. The method of claim 1, wherein, performing background isolation on the second traffic image based on the ORB feature clustering result to obtain a third traffic image comprises: obtaining a plurality of center ORB features based on the ORB feature clustering result; Performing curve fitting based on the plurality of center ORB features, to obtain a center ORB feature curve; Positioning the center ORB feature curve based on the preset ORB feature, to obtain a background center ORB feature; Based on the background center ORB feature, obtaining a background ORB feature clustering result according to the ORB feature clustering result; Based on the background ORB feature clustering result, performing background isolation on the second traffic image, to obtain the third traffic image.

6. A traffic image segmentation control system characterized by comprising: The system is used to execute the method of any one of claims 1 to 5, and the system comprises: A vehicle obtaining module is used to obtain a target vehicle; A camera interaction module is used to interact with the vehicle-mounted camera of the target vehicle, to obtain a first traffic image; A noise reduction processing module is used to perform noise reduction processing on the first traffic image according to a wavelet filtering algorithm, to obtain a second traffic image; An ORB feature recognition module is used to perform ORB feature recognition based on the second traffic image, to obtain an ORB feature recognition result; An ORB feature clustering module is used to perform clustering based on the ORB feature recognition result, to obtain an ORB feature clustering result; An image background isolation module is used to perform background isolation on the second traffic image based on the ORB feature clustering result, to obtain a third traffic image, and the third traffic image has an ORB feature identifier; A building module is used to build a traffic image segmentation channel, specifically comprising: Obtaining a traffic image segmentation record set; Dividing the traffic image segmentation record set based on a preset division operator, to obtain a training data sequence, a test data sequence, and a verification data sequence; Based on a fully connected neural network, training, testing, and verifying according to the training data sequence, the test data sequence, and the verification data sequence, to generate the traffic image segmentation channel; 7. An electronic device, comprising: A segmentation control module is used to perform segmentation control on the third traffic image based on the ORB feature identifier and the traffic image segmentation channel, to obtain a target traffic image segmentation result. The electronic device comprises: A memory is used to store executable instructions; 8. A computer-readable storage medium having stored thereon a computer program, characterized in that, A processor is used to execute the executable instructions stored in the memory, to implement the traffic image segmentation control method of any one of claims 1 to 5. The program is executed by the processor to implement the traffic image segmentation control method of any one of claims 1 to 5.

Images