Method and device for detecting condensation in pipeline of EGR system, electronic equipment and storage medium

Abstract

The present disclosure discloses a pipeline condensation detection method and device, electronic equipment and storage medium of an EGR system, and relates to the field of data processing technology. At least two images to be detected in a continuous time period inside the pipeline are obtained, and the similarity between the feature data contained in the at least two images to be detected and the reference condensation feature data is calculated in sequence. When it is determined that the similarities corresponding to the at least two images to be detected are greater than or equal to a preset similarity threshold, it is determined that condensation exists inside the pipeline. Compared with the related art, the condensation situation in the pipeline is judged by identifying at least two images to be detected in a continuous time period inside the pipeline. The judgment process is more intuitive, and there is no need to conduct a comprehensive analysis based on the environmental parameters in the pipeline to obtain the final detection result, eliminating the influence of the environmental parameters on the detection result, and when based on image recognition, according to at least two images to be detected in a continuous time period, the misjudgment rate is reduced and the accuracy of the detection result is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of vehicle technology, and in particular to a pipeline condensation detection method and device, electronic equipment and storage medium of an EGR system. Background Art

[0002] Exhaust Gas Recirculation (EGR) is a process in which a portion of the exhaust gas is separated after combustion in an internal combustion engine and introduced into the intake side for re-combustion.

[0003] When the engine is running, the exhaust gas will be introduced into the intake system and eventually participate in combustion. Therefore, a large amount of water vapor in the exhaust gas will also be introduced into the intake system. The exhaust gas will then be cooled by the cooling system in the intake system. If the temperature is too low, liquid water will appear in the intake pipe due to the liquefaction of water vapor. If the liquid water enters the engine cylinder through the pipe, it will affect the normal operation of the engine. In order to prevent liquid water from entering the engine through the pipe, the humidity, temperature and other parameters in the pipe are currently mainly detected by sensors to determine whether liquid water is generated in the pipe. However, the detection method of using sensors to detect environmental parameters in the pipe to determine whether there is liquid water in the pipe has the problem of being affected by environmental factors, resulting in inaccurate detection results. Summary of the invention

[0004] The present invention provides a pipeline condensation detection method and device, electronic equipment and storage medium of an EGR system. The main purpose is to solve the problem that the detection method of using a sensor to detect environmental parameters in the pipeline to determine whether there is liquid water in the pipeline is affected by environmental factors, resulting in inaccurate detection results.

[0005] According to a first aspect of the present disclosure, a method for detecting condensation in a pipeline of an EGR system is provided, comprising:

[0006] Acquire at least two images to be detected in consecutive time periods inside the pipeline;

[0007] sequentially calculating the similarities between the feature data contained in the at least two images to be detected and the reference condensation feature data;

[0008] When it is determined that the similarities respectively corresponding to at least two images to be detected are both greater than or equal to a preset similarity threshold, it is determined that condensation exists inside the pipeline.

[0009] Optionally, the sequentially calculating the similarity between the feature data contained in the at least two images to be detected and the reference condensation feature data includes:

[0010] Calling a preset condensation detection algorithm to calculate a first similarity between first image feature data of a first image to be detected and the reference condensation feature data;

[0011] After determining that the first similarity is greater than or equal to a preset threshold, the second similarity between the second image feature data of the second image to be detected and the reference condensation feature data is calculated until all the images to be detected are calculated.

[0012] Optionally, when it is determined that the similarities respectively corresponding to at least two images to be detected are both greater than or equal to a preset similarity threshold, determining that condensation exists inside the pipeline includes:

[0013] When it is determined that the similarity corresponding to the feature data of a previous image in at least two images to be detected is less than the similarity corresponding to the feature data of a subsequent image, it is determined that condensation exists inside the pipeline.

[0014] Optionally, before sequentially calculating the similarities between the feature data contained in the at least two images to be detected and the reference condensation feature data, the method includes:

[0015] Identify the pixel position of the feature data in the image to be detected;

[0016] According to the projection relationship between the image to be detected and the internal environment of the pipeline, a target position corresponding to the pixel position in the internal environment of the pipeline is determined, and the target position is a position where condensation occurs in the pipeline.

[0017] Optionally, determining the target position corresponding to the pixel position in the internal environment of the pipeline according to the projection relationship between the image to be detected and the internal environment of the pipeline includes:

[0018] According to the intrinsic parameter matrix of the camera that collects the image to be detected, the pixel position is converted into the first target three-dimensional coordinates in the camera coordinate system;

[0019] The target position is determined according to the transformation relationship between the camera coordinate system and the environment coordinate system in the pipeline and the first target three-dimensional coordinates.

[0020] Optionally, acquiring at least two images to be detected in consecutive time periods within the pipeline includes:

[0021] Acquire a video to be detected that includes the environment inside the pipeline, wherein the video to be detected includes a timestamp corresponding to each video frame;

[0022] The video to be detected is divided into a preset number of video frames based on the timestamp, and the video frames with the preset data amount are used as the images to be detected, so as to obtain at least two images to be detected inside the pipeline in a continuous time period.

[0023] According to a second aspect of the present disclosure, a pipeline condensation detection device for an EGR system is provided, comprising:

[0024] An acquisition unit, used for acquiring at least two images to be detected in a continuous time period inside the pipeline;

[0025] A calculation unit, used for sequentially calculating the similarity between the feature data contained in the at least two images to be detected and the reference condensation feature data;

[0026] The first determining unit is configured to determine that condensation exists inside the pipeline when it is determined that the similarities respectively corresponding to at least two images to be detected are both greater than or equal to a preset similarity threshold.

[0027] Optionally, the computing unit is further configured to:

[0028] Calling a preset condensation detection algorithm to calculate a first similarity between first image feature data of a first image to be detected and the reference condensation feature data;

[0029] After determining that the first similarity is greater than or equal to a preset threshold, the second similarity between the second image feature data of the second image to be detected and the reference condensation feature data is calculated until all the images to be detected are calculated.

[0030] Optionally, the first determination unit is further configured to determine that condensation exists inside the pipeline when it is determined that a similarity corresponding to feature data of a previous image in at least two images to be detected is less than a similarity corresponding to feature data of a subsequent image.

[0031] Optionally, the device comprises:

[0032] An identification unit, configured to identify pixel positions of the feature data in the images to be detected before the calculation unit sequentially calculates similarities between the feature data contained in the at least two images to be detected and the reference condensation feature data;

[0033] The second determining unit is used to determine a target position corresponding to the pixel position in the internal environment of the pipeline according to a projection relationship between the image to be detected and the internal environment of the pipeline, wherein the target position is a position where condensation occurs in the pipeline.

[0034] Optionally, the second determining unit is further configured to:

[0035] According to the intrinsic parameter matrix of the camera that collects the image to be detected, the pixel position is converted into the first target three-dimensional coordinates in the camera coordinate system;

[0036] The target position is determined according to the transformation relationship between the camera coordinate system and the environment coordinate system in the pipeline and the first target three-dimensional coordinates.

[0037] Optionally, the acquiring unit is further used for:

[0038] Acquire a video to be detected that includes the environment inside the pipeline, wherein the video to be detected includes a timestamp corresponding to each video frame;

[0039] The video to be detected is divided into a preset number of video frames based on the timestamp, and the video frames with the preset data amount are used as the images to be detected, so as to obtain at least two images to be detected inside the pipeline in a continuous time period.

[0040] According to a third aspect of the present disclosure, a vehicle is provided, comprising the pipe condensation detection device of the EGR system according to the second aspect.

[0041] According to a fourth aspect of the present disclosure, there is provided an electronic device, including:

[0042] at least one processor; and

[0043] a memory communicatively connected to the at least one processor; wherein,

[0044] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in the first aspect.

[0045] According to a fifth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the method described in the first aspect.

[0046] According to a sixth aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0047] The pipeline condensation detection method, device, electronic device and storage medium of the EGR system provided by the present disclosure obtain at least two images to be detected in a continuous time period inside the pipeline, calculate the similarity between the feature data contained in the at least two images to be detected and the reference condensation feature data in sequence, and determine that condensation exists inside the pipeline when it is determined that the similarities corresponding to the at least two images to be detected are greater than or equal to the preset similarity threshold. Compared with the related art, the condensation situation in the pipeline is judged by identifying at least two images to be detected in a continuous time period inside the pipeline, and the judgment process is more intuitive, and there is no need to conduct a comprehensive analysis based on the environmental parameters in the pipeline to obtain the final detection result, eliminating the influence of the environmental parameters on the detection result, and when based on image recognition, according to at least two images to be detected in a continuous time period, the misjudgment rate is reduced and the accuracy of the detection result is improved.

[0048] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0050] Figure 1 A schematic flow chart of a pipeline condensation detection method for an EGR system provided in an embodiment of the present disclosure;

[0051] Figure 2 A schematic diagram of a process for training a preset condensation detection model provided in an embodiment of the present disclosure;

[0052] Figure 3 A schematic diagram of the structure of a pipeline condensation detection device for an EGR system provided in an embodiment of the present disclosure;

[0053] Figure 4 A schematic structural diagram of another pipeline condensation detection device for an EGR system provided in an embodiment of the present disclosure;

[0054] Figure 5 A schematic block diagram of an exemplary electronic device 400 provided for an embodiment of the present disclosure. DETAILED DESCRIPTION

[0055] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0056] The following describes a pipeline condensation detection method, device, electronic device, and storage medium for an EGR system according to an embodiment of the present disclosure with reference to the accompanying drawings.

[0057] Figure 1 A schematic flow chart of a method for detecting condensation in a pipeline of an EGR system provided in an embodiment of the present disclosure.

[0058] like Figure 1 As shown, the method comprises the following steps:

[0059] Step 101: Acquire at least two images to be detected in a continuous time period inside a pipeline.

[0060] As a refinement of the above-mentioned step 101, in response to a condensation detection instruction, which is a start instruction for performing condensation detection in the pipeline, at least two images to be detected that include a continuous time period of the environment in the pipeline are obtained, and the images to be detected are obtained by photographing the environment in the pipeline, and the method of obtaining the images to be detected includes but is not limited to a panoramic camera, that is, the type of camera that collects the images to be detected is not limited in this disclosure.

[0061] The purpose of acquiring at least two images to be detected within a continuous time period is to prevent the discrimination error of condensation caused by the recognition result of one image to be detected. As a feasible method of an embodiment of the present application, the continuous time period may include but is not limited to one second or two seconds. Taking the continuous time period of one second as an example, at least two images to be detected in one second are acquired, such as 2 images to be detected, 3 images to be detected, or 5 images to be detected. It should be noted that the more images to be detected, the more accurate the detection result, but the corresponding data overhead is also relatively increased. Therefore, the specific continuous time period and the specific values ​​of the images to be detected can be flexibly adjusted according to actual needs.

[0062] Step 102: sequentially calculating the similarities between the feature data contained in the at least two images to be detected and the reference condensation feature data.

[0063] As a refinement of the above step 102, the judgment of whether there is condensation in the pipeline is realized by performing image recognition on at least two images to be detected, and the image recognition of the images to be detected is mainly to determine the similarity between the feature data included in the images to be detected and the reference condensation feature data. It should be understood that the feature data is the data obtained by feature extraction of the images to be detected, and the reference condensation feature data is the data already existing in the database corresponding to the preset condensation detection algorithm. In this embodiment, the reference condensation feature data is the data corresponding to the existence of condensation. The preset condensation detection algorithm is a deep learning algorithm. The feature data and the reference condensation feature data are calculated based on the preset condensation algorithm to obtain the similarity between the feature data and the reference condensation feature data. In short, it is to judge whether the pipeline environment included in the image to be detected tends to have condensation.

[0064] Step 103 : When it is determined that the similarities corresponding to at least two images to be detected are both greater than or equal to a preset similarity threshold, it is determined that condensation exists inside the pipeline.

[0065] As a refinement of the above-mentioned step 103, in order to determine whether condensation exists in the pipeline, when it is determined that the similarity between the feature data of at least two images to be detected and the reference condensation feature data is greater than or equal to a preset similarity threshold, it is determined that condensation exists in the pipeline. The preset similarity threshold is a pre-set value. This embodiment does not limit the specific value of the preset similarity threshold. It can be seen from the above description that the determination of the presence of condensation in the pipeline is achieved by judging whether the feature data is the same as the reference condensation feature data. The reference condensation feature data is data corresponding to the presence of condensation in the pipeline.

[0066] The pipeline condensation detection method of the EGR system provided by the present disclosure obtains at least two images to be detected in a continuous time period inside the pipeline, calculates the similarity between the feature data contained in the at least two images to be detected and the reference condensation feature data in sequence, and determines that condensation exists inside the pipeline when it is determined that the similarities corresponding to the at least two images to be detected are greater than or equal to a preset similarity threshold. Compared with the related art, the condensation situation in the pipeline is judged by identifying at least two images to be detected in a continuous time period inside the pipeline, and the judgment process is more intuitive, and there is no need to conduct a comprehensive analysis based on the environmental parameters in the pipeline to obtain the final detection result, eliminating the influence of the environmental parameters on the detection result, and when based on image recognition, according to at least two images to be detected in a continuous time period, the false judgment rate is reduced and the accuracy of the detection result is improved.

[0067] As a refinement of the embodiment of the present disclosure, when executing step 102 to sequentially calculate the similarity between the feature data contained in the at least two images to be detected and the reference condensation feature data, the following implementation method can be adopted but is not limited to, for example: calling a preset condensation detection algorithm to calculate the first similarity between the first image feature data of the first image to be detected and the reference condensation feature data, and after determining that the first similarity is greater than or equal to a preset threshold, calculating the second similarity between the second image feature data of the second image to be detected and the reference condensation feature data, until all the images to be detected are calculated.

[0068] In practical applications, the preset condensation detection algorithm is applied to a preset condensation detection model, the image to be detected is input into the preset condensation detection model, and the similarity between the feature data and the reference condensation feature data is detected based on the preset condensation detection model. The reference condensation feature data is contained in a database of the preset condensation detection model, and the similarity between the feature data contained in the first image to be detected in at least two images to be detected and the reference condensation feature data is obtained. After determining that the first similarity is greater than or equal to a preset threshold, it indicates that condensation may exist in the continuous time period. In order to improve the accuracy of condensation recognition, the second similarity between the second image feature data of the second image to be detected in at least two images to be detected and the reference condensation feature data is calculated until all images to be detected are calculated.

[0069] Applying the preset condensation detection algorithm to the preset condensation detection model means constructing the preset condensation detection model based on the preset condensation detection algorithm, and the preset condensation detection model calculates the similarity between the feature data and the reference condensation feature data based on the preset condensation detection algorithm. The specific process is as follows: inputting the image to be detected into the preset condensation detection model. Since the preset condensation detection model contains the preset condensation detection algorithm, the preset condensation detection model can detect the similarity between the feature data and the reference condensation feature data.

[0070] In some embodiments, the preset condensation detection algorithm may be, but not limited to, computer vision algorithms such as SIFT, SURF, and ORB, and the preset condensation detection model may be, but not limited to, deep learning and training models such as the VGG series. Exemplarily, the preset condensation detection model is an AlexNet model.

[0071] As an extension of the disclosed embodiment, this embodiment provides a method for extracting the feature data in the image to be detected, including: preprocessing the image to be detected to improve the accuracy of recognition. Preprocessing includes graying, filtering, binarization, morphological processing, etc. Among them, graying converts the image to be detected from RGB color space to grayscale color space, which is convenient for subsequent processing; filtering can remove image noise and improve image quality; binarization can convert the image into a black and white image, which is convenient for contour detection and feature extraction; morphological processing can perform operations such as corrosion and expansion on the binarized image to further extract features. The edges of objects in the image to be detected can be extracted through contour detection. Some classic algorithms such as Canny edge detection algorithm, Sobel algorithm, Laplacian algorithm, etc. can be used. For each object in the image to be detected, some feature data needs to be extracted to distinguish different objects. Common features include shape, texture, color, etc. Among them, shape features can be extracted by calculating the object's contour, perimeter, area and other indicators; texture features can be extracted by calculating the object's grayscale co-occurrence matrix, wavelet transform, etc.; color features can be extracted by calculating the object's RGB or HSV value. For the feature data that needs to be extracted for image recognition of EGR mixed gas condensation, there may be some redundant or less useful features. Feature selection and dimensionality reduction techniques, such as principal component analysis, linear discriminant analysis, t-SNE, etc., can be used to reduce the feature dimension or select the most representative feature data.

[0072] As a refinement of the above embodiment, the method can also be adopted but not limited to the following manner: when it is determined that the similarities corresponding to at least two images to be detected are both greater than or equal to a preset similarity threshold, when it is determined that condensation exists inside the pipe, when it is determined that the similarity corresponding to the previous image feature data (such as the first image feature data) in at least two images to be detected is less than the similarity corresponding to the next image feature data (such as the second image feature data), it is determined that condensation exists inside the pipe. When the similarity corresponding to the previous image feature data is less than the similarity corresponding to the next image feature data, it can be determined that condensation exists, so as to prevent misjudgment of condensation recognition results, and whether condensation exists can be determined by judging the two images to be detected, which further saves computing resources.

[0073] In order to determine whether there is condensation in the pipeline, after calculating the similarity between the feature data and the reference condensation feature data based on the preset condensation detection algorithm, if it is determined that the similarities corresponding to at least two images to be detected are both less than the preset similarity threshold, it is determined that there is no condensation in the pipeline.

[0074] As a refinement of the above embodiment, after successively calculating the similarity between the feature data contained in the at least two images to be detected and the reference condensation feature data, the method can also be implemented in the following manner, but not limited to, for example: identifying the pixel position of the feature data in the image to be detected; determining the target position corresponding to the pixel position in the environment in the pipeline according to the projection relationship between the image to be detected and the environment in the pipeline, the target position being the position of condensation in the pipeline. This method can not only accurately identify condensation, but also identify the specific position of condensation.

[0075] It should be noted that the feature data is a part of the picture in the image to be detected, and the image to be detected is composed of pixel points. The position of the pixel point corresponding to the feature data on the image to be detected is the similarity position, so as to identify and determine the pixel position of the feature data in the image to be detected; in the above description, it is known that the image to be detected is obtained by photographing the environment inside the pipeline, that is, the environment inside the pipeline included in the image to be detected has a projection relationship with the actual environment inside the pipeline. Based on the mapping relationship, the target position corresponding to the pixel position in the environment inside the pipeline can be determined. The feature data is condensation, and the target position is the position of condensation in the pipeline.

[0076] As a refinement of the above embodiment, when determining the target position corresponding to the pixel position in the environment inside the pipeline according to the projection relationship between the image to be detected and the environment inside the pipeline, the following implementation method can be adopted but is not limited to, for example: according to the intrinsic parameter matrix of the camera that acquires the image to be detected, the pixel position is converted into the first target three-dimensional coordinates in the camera coordinate system; the target position is determined according to the transformation relationship between the camera coordinate system and the coordinate system of the environment inside the pipeline and the first target three-dimensional coordinates.

[0077] In an embodiment of the present application, the camera that collects the at least two images to be detected is a camera that photographs the environment inside the pipeline, the intrinsic parameter matrix is ​​a camera parameter corresponding to the camera, and the intrinsic parameter matrix is ​​obtained by calibrating the camera. In order to determine the target position according to the pixel position, the pixel position is converted into a first target three-dimensional coordinate in a camera coordinate system according to the intrinsic parameter matrix, and the first target three-dimensional coordinate is transformed into a three-dimensional coordinate corresponding to the target position in the environment coordinate system inside the pipeline according to a transformation relationship between the camera coordinate system and the coordinate system of the environment inside the pipeline, thereby determining the target position.

[0078] In an embodiment of the present disclosure, a training method for the preset condensation detection model is provided. Figure 2A schematic diagram of a process for training a preset condensation detection model provided in an embodiment of the present disclosure, such as Figure 2 As shown, the steps include:

[0079] Step 201 : acquiring training images, wherein the training images include a first training image in which condensation occurs in the pipeline environment, and a second training image in which condensation does not occur in the pipeline environment.

[0080] In order to more clearly illustrate the steps involved in the above-mentioned embodiment, this embodiment provides an exemplary description here, for example: based on a notebook and a camera, the camera is installed in the air intake pipe to record video images in real time. During the recording process, in order to make the preset anti-pinch model for subsequent training more robust, it is necessary to collect image information at multiple time points. When the collected image information is video data, it is necessary to convert the video data into images based on video processing software, and the converted images are the training images to be obtained, that is, the first training image containing condensation in the environment inside the pipeline, and the second training image containing no condensation in the environment inside the pipeline. The method of converting video data into images includes: through Convertto JPG or ffmpeg software, the video stream frame is converted into 24 images per second, so as to realize the conversion of the video into a training image, and the video image is obtained from the video stream frame.

[0081] Step 202: Perform data enhancement processing on the first training image and the second training image to obtain an enhanced first training image and an enhanced second training image.

[0082] As a refinement of the above step 202, the enhancement processing of the first training image and the second training image includes: performing image flipping and rotation, image cropping and scaling, image shifting and edge filling on the first training image and the second training image to obtain the enhanced first training image and the second training image. The training images cover more scenes, and the data volume in each scene is also guaranteed to be sufficient. This embodiment does not limit the enhancement method of the first training image and the second training image.

[0083] In some embodiments, the enhancement processing of the first training image and the second training image further includes denoising, graying, binarization and other operations on the first training image and the second training image to facilitate subsequent feature extraction and analysis.

[0084] Step 203: input the enhanced first training image and the enhanced second training image into the preset condensation detection model, and train the preset condensation detection model according to preset training conditions.

[0085] As a refinement of the above-mentioned step 203, the training images need to be divided into training sample sets before executing step 203. In the present embodiment, the division is performed in a ratio of 8:2 between the training set and the test set, that is, the sum of the number of the first training images and the second training images is divided into training image samples with different functions in a ratio of 8:2. The present disclosure does not limit the division ratio of the training sample set.

[0086] The training of the preset condensation detection model according to the preset training conditions includes: designing model training parameters: 50 epochs, 256 as a batch size, the learning rate is set to 0.01, and the number of layers and neurons of the neural network obey the AlexNet algorithm and the training model. After the 50 epoch training is completed, a preset condensation detection model will be generated. The above exemplary description is only exemplary, and this embodiment does not limit the setting of specific model training parameters.

[0087] In some embodiments, after generating a preset condensation detection model, the preset condensation detection model is evaluated, and based on the evaluation results, the model is optimized, such as adjusting hyperparameters, increasing data sets, adjusting model structures, etc., to further improve the performance of the model.

[0088] An epoch means: all training image samples are sent to the neural network model, completing a forward calculation and back propagation process. When performing an epoch training, the training image samples used need to be input into the neural network model in batches. The 256 is a batchsize, which means that each batch contains 256 training image samples. The learning rate determines whether the objective function can converge to the local minimum and when it converges to the minimum. The lower the learning rate, the slower the change speed of the loss function. Using a low learning rate can ensure that the preset condensation detection model will not miss any local minimum.

[0089] In some embodiments, the preset condensation detection model adopts the Alexnet model, and the Alexnet model is adjusted, specifically including: adopting a deeper network structure, using 5 layers of convolution and 3 layers of fully connected layers, a total of 8 layers, thereby increasing the nonlinear expression ability and learning ability of the model; introducing the Dropout layer to randomly inactivate the fully connected layer to reduce the risk of overfitting of the model; using the ReLU activation function instead of the traditional Sigmoid activation function to make the model training faster and reduce the gradient disappearance problem; using data enhancement technology to crop, flip, scale and other operations on the input image, thereby increasing the number of samples in the data set and increasing the generalization ability of the model; using two GPUs for parallel computing to accelerate the training and reasoning process of the model.

[0090] Step 204, after the preset condensation detection model is trained, the test image is input into the trained preset condensation detection model, and a detection result of the trained preset condensation detection model for the test image is obtained.

[0091] As a refinement of the above step 204, the purpose of inputting the test image into the preset condensation detection model is to test the effect of the preset condensation detection model in identifying the presence of condensation in the pipeline, that is, to test the accuracy of the recognition result.

[0092] Step 205, confirming that the training of the preset condensation detection model is completed.

[0093] If the test result meets the preset accuracy condition, step 205 is executed.

[0094] The preset accuracy conditions include: the accuracy rate, recall rate, and F-value indicators of the preset condensation detection model in identifying the presence of condensation conditions in the pipeline. The trained preset condensation detection model must ultimately meet the preset parameters of the indicators. When the preset condensation detection model meets the preset parameters, it is determined that the training of the preset condensation detection model is completed.

[0095] In some embodiments, after determining that the preset condensation detection model has been trained, the actual application of the preset condensation detection model includes: inputting the image to be detected into the trained preset condensation detection model, and the preset condensation detection model will output the probability that each pixel in the image to be detected belongs to liquid water. By setting a probability threshold and marking pixels with a probability higher than the threshold as liquid water, the location of the liquid water can be obtained, and whether there is condensation in the pipeline can be determined. In actual applications, it is also necessary to pay attention to factors such as the quality of the collected image data to be detected and changes in the environment, and adjust and update the model in real time.

[0096] In some embodiments, the preset condensation detection model can use manually annotated images, or collect real data through sensors and other devices, annotate the existence of condensation, and perform training and verification. The trained preset condensation detection model classifies the feature data to determine whether condensation exists in the pipeline.

[0097] Step 206: retrain the preset condensation detection model according to the preset training conditions.

[0098] If the test result does not meet the preset accuracy condition, step 206 is executed.

[0099] Contrary to the judgment condition of the above step 205, if the preset condensation detection model does not meet the setting of the indicator, the preset condensation detection model is retrained according to the preset training condition.

[0100] As a refinement of the above embodiment, when executing step 101 to obtain at least two images to be detected in a continuous time period inside the pipeline, the following implementation method can be adopted but is not limited to, for example: obtaining a video to be detected containing the environment inside the pipeline, wherein the video to be detected includes timestamps corresponding to each video frame; based on the timestamps, the video to be detected is divided into a preset number of video frames, and the video frames with the preset data amount are used as the images to be detected.

[0101] The content described in the above embodiment is explained in the refinement of step 201, wherein the video to be detected including the environment in the pipeline is obtained, wherein the video to be detected includes the timestamp corresponding to each video frame, that is, the video data is collected based on the camera and then converted into image data based on the preset video processing software. In this embodiment, the video stream frame of the video data is converted into 24 images per second, thereby obtaining a training image sample. It should be understood that this embodiment does not limit the number of images converted from the video stream frame.

[0102] In some embodiments, the method disclosed in the present invention for detecting whether there is condensation in the pipeline based on a preset condensation detection model can also be used in combination with any of the following methods, including: directly shutting down EGR in a low temperature area to prevent liquid water in the intake pipeline; using a temperature sensor to detect the temperature and pressure at the inlet and outlet of the EGR valve, after the air filter, before the intercooler, and after the intercooler; using a humidity sensor to detect the humidity in the mixed gas in the EGR pipeline; using a gas analyzer to analyze and detect various components in the EGR mixed gas to obtain the component distribution and concentration of the EGR mixed gas; by fusing and processing the detection data of multiple sensors, comprehensive information about the EGR mixed gas can be obtained, including temperature, pressure, flow, humidity, and component distribution.

[0103] In summary, this embodiment can achieve the following effects:

[0104] 1. Based on image recognition, at least two images to be detected in continuous time periods are used to reduce the misjudgment rate and improve the accuracy of the detection results.

[0105] 2. Use the preset condensation detection model to identify whether there is condensation in the intake duct. Since the preset condensation detection model is independent of other vehicle components, the signal is not coupled with the signal of other components, it has strong independence and is not prone to errors, thereby improving the accuracy of the detection results.

[0106] 3. During the EGR opening process, whether there is condensation in the intake duct can be identified in advance based on the preset condensation detection model, providing the possibility for accurate control of EGR.

[0107] Corresponding to the above-mentioned EGR system pipeline condensation detection method, the present invention also proposes a EGR system pipeline condensation detection device. Since the device embodiment of the present invention corresponds to the above-mentioned method embodiment, the details not disclosed in the device embodiment can be referred to the above-mentioned method embodiment, and will not be repeated in the present invention.

[0108] Figure 3 A schematic diagram of a pipeline condensation detection device for an EGR system provided in an embodiment of the present disclosure is shown in FIG. Figure 3 As shown, including:

[0109] An acquisition unit 31 is used to acquire at least two images to be detected in a continuous time period inside the pipeline;

[0110] A calculation unit 32, used for sequentially calculating the similarity between the feature data contained in the at least two images to be detected and the reference condensation feature data;

[0111] The first determination unit 33 is configured to determine that condensation exists inside the pipeline when it is determined that the similarities respectively corresponding to at least two images to be detected are both greater than or equal to a preset similarity threshold.

[0112] Furthermore, in a possible implementation of this embodiment, as Figure 4 As shown, the calculation unit 32 is also used for:

[0113] Calling a preset condensation detection algorithm to calculate a first similarity between first image feature data of a first image to be detected and the reference condensation feature data;

[0114] After determining that the first similarity is greater than or equal to a preset threshold, the second similarity between the second image feature data of the second image to be detected and the reference condensation feature data is calculated until all the images to be detected are calculated.

[0115] Furthermore, in a possible implementation of this embodiment, as Figure 4 As shown, the first determination unit 33 is further used to determine that condensation exists inside the pipeline when it is determined that the similarity corresponding to the feature data of the previous image in at least two images to be detected is less than the similarity corresponding to the feature data of the next image.

[0116] Furthermore, in a possible implementation of this embodiment, as Figure 4 As shown, the device comprises:

[0117] An identification unit 34, configured to identify pixel positions of the feature data in the image to be detected before the calculation unit sequentially calculates the similarities between the feature data contained in the at least two images to be detected and the reference condensation feature data;

[0118] The second determination unit 35 is used to determine the target position corresponding to the pixel position in the internal environment of the pipeline according to the projection relationship between the image to be detected and the internal environment of the pipeline, and the target position is the position of condensation in the pipeline.

[0119] Furthermore, in a possible implementation of this embodiment, as Figure 4 As shown, the second determining unit 35 is further used for:

[0120] According to the intrinsic parameter matrix of the camera that collects the image to be detected, the pixel position is converted into the first target three-dimensional coordinates in the camera coordinate system;

[0121] The target position is determined according to the transformation relationship between the camera coordinate system and the environment coordinate system in the pipeline and the first target three-dimensional coordinates.

[0122] Furthermore, in a possible implementation of this embodiment, as Figure 4 As shown, the acquisition unit 31 is also used for:

[0123] Acquire a video to be detected that includes the environment inside the pipeline, wherein the video to be detected includes a timestamp corresponding to each video frame;

[0124] The video to be detected is divided into a preset number of video frames based on the timestamp, and the video frames with the preset data amount are used as the images to be detected, so as to obtain at least two images to be detected inside the pipeline in a continuous time period.

[0125] The pipeline condensation detection device of the EGR system provided by the present disclosure obtains at least two images to be detected in a continuous time period inside the pipeline, calculates the similarity between the feature data contained in the at least two images to be detected and the reference condensation feature data in sequence, and determines that condensation exists inside the pipeline when it is determined that the similarities corresponding to the at least two images to be detected are greater than or equal to a preset similarity threshold. Compared with the related art, the condensation situation in the pipeline is judged by identifying at least two images to be detected in a continuous time period inside the pipeline, and the judgment process is more intuitive, and there is no need to conduct a comprehensive analysis based on the environmental parameters in the pipeline to obtain the final detection result, eliminating the influence of the environmental parameters on the detection result, and when based on image recognition, according to at least two images to be detected in a continuous time period, the misjudgment rate is reduced and the accuracy of the detection result is improved.

[0126] It should be noted that the above explanation of the method embodiment is also applicable to the device of this embodiment, and the principle is the same, which is no longer limited in this embodiment.

[0127] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0128] Figure 5 A schematic block diagram of an example electronic device 400 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0129] like Figure 5 As shown, the device 400 includes a computing unit 401, which can perform various appropriate actions and processes according to a computer program stored in a ROM (Read-Only Memory) 402 or a computer program loaded from a storage unit 408 to a RAM (Random Access Memory) 403. In the RAM 403, various programs and data required for the operation of the device 400 can also be stored. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An I / O (Input / Output) interface 405 is also connected to the bus 404.

[0130] A number of components in the device 400 are connected to the I / O interface 405, including: an input unit 406, such as a keyboard, a mouse, etc.; an output unit 407, such as various types of displays, speakers, etc.; a storage unit 408, such as a disk, an optical disk, etc.; and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the device 400 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0131] The computing unit 401 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a CPU (Central Processing Unit), a GPU (Graphic Processing Units), various dedicated AI (Artificial Intelligence) computing chips, various computing units running machine learning model algorithms, a DSP (Digital Signal Processor), and any appropriate processor, controller, microcontroller, etc. The computing unit 401 performs the various methods and processes described above, such as a pipeline condensation detection method for an EGR system. For example, in some embodiments, the pipeline condensation detection method for an EGR system may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 408. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded into the RAM 403 and executed by the computing unit 401, one or more steps of the method described above may be performed. Alternatively, in other embodiments, the computing unit 401 may be configured to execute the aforementioned pipeline condensation detection method of the EGR system in any other appropriate manner (for example, by means of firmware).

[0132] Various embodiments of the systems and techniques described above herein may be implemented in digital electronic circuit systems, integrated circuit systems, FPGAs (Field Programmable Gate Arrays), ASICs (Application-Specific Integrated Circuits), ASSPs (Application Specific Standard Products), SOCs (System On Chips), CPLDs (Complex Programmable Logic Devices), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs that may be executed and / or interpreted on a programmable system including at least one programmable processor that may be a dedicated or general-purpose programmable processor that may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0133] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0134] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a RAM, a ROM, an EPROM (Electrically Programmable Read-Only-Memory) or a flash memory, an optical fiber, a CD-ROM (Compact Dis sc Read-Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0135] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (Cathode-Ray Tube) or LCD (Liquid Crystal Display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the 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 input, voice input, or tactile input).

[0136] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: LAN (Local Area Network), WAN (Wide Area Network), the Internet, and blockchain networks.

[0137] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may 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 to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services ("Virtual Private Server", or "VPS" for short). The server may also be a server of a distributed system, or a server combined with a blockchain.

[0138] It should be noted that artificial intelligence is a discipline that studies how computers can simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, planning, etc.), and includes both hardware-level and software-level technologies. Artificial intelligence hardware technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, and big data processing; artificial intelligence software technologies mainly include computer vision technology, speech recognition technology, natural language processing technology, as well as machine learning / deep learning, big data processing technology, knowledge graph technology, and other major directions.

[0139] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0140] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for detecting condensation in a pipeline of an EGR system, characterized in that: include: Acquire at least two images to be detected in consecutive time periods inside the pipeline; sequentially calculating the similarities between the feature data contained in the at least two images to be detected and the reference condensation feature data; When it is determined that the similarities respectively corresponding to at least two images to be detected are both greater than or equal to a preset similarity threshold, it is determined that condensation exists inside the pipeline.

2. The method according to claim 1, characterized in that The sequentially calculating the similarity between the feature data contained in the at least two images to be detected and the reference condensation feature data comprises: Calling a preset condensation detection algorithm to calculate a first similarity between first image feature data of a first image to be detected and the reference condensation feature data; After determining that the first similarity is greater than or equal to a preset threshold, the second similarity between the second image feature data of the second image to be detected and the reference condensation feature data is calculated until all the images to be detected are calculated.

3. The method according to claim 2, characterized in that When it is determined that the similarities respectively corresponding to at least two images to be detected are both greater than or equal to a preset similarity threshold, determining that condensation exists inside the pipeline includes: When it is determined that the similarity corresponding to the feature data of a previous image in at least two images to be detected is less than the similarity corresponding to the feature data of a subsequent image, it is determined that condensation exists inside the pipeline.

4. The method according to claim 3, characterized in that Before sequentially calculating the similarities between the feature data contained in the at least two images to be detected and the reference condensation feature data, the method includes: Identify the pixel position of the feature data in the image to be detected; According to the projection relationship between the image to be detected and the internal environment of the pipeline, a target position corresponding to the pixel position in the internal environment of the pipeline is determined, and the target position is a position where condensation occurs in the pipeline.

5. The method according to claim 4, characterized in that Determining the target position corresponding to the pixel position in the internal environment of the pipeline according to the projection relationship between the image to be detected and the internal environment of the pipeline includes: According to the intrinsic parameter matrix of the camera that collects the image to be detected, the pixel position is converted into the first target three-dimensional coordinates in the camera coordinate system; The target position is determined according to the transformation relationship between the camera coordinate system and the environment coordinate system in the pipeline and the first target three-dimensional coordinates.

6. The method according to any one of claims 1 to 5, characterized in that The acquiring of at least two images to be detected in a continuous time period inside the pipeline comprises: Acquire a video to be detected that includes the environment inside the pipeline, wherein the video to be detected includes a timestamp corresponding to each video frame; The video to be detected is divided into a preset number of video frames based on the timestamp, and the video frames with the preset data amount are used as the images to be detected, so as to obtain at least two images to be detected inside the pipeline in a continuous time period.

7. A pipeline condensation detection device for an EGR system, characterized in that: include: An acquisition unit, used for acquiring at least two images to be detected in a continuous time period inside the pipeline; A calculation unit, used for sequentially calculating the similarity between the feature data contained in the at least two images to be detected and the reference condensation feature data; The first determining unit is configured to determine that condensation exists inside the pipeline when it is determined that the similarities respectively corresponding to at least two images to be detected are both greater than or equal to a preset similarity threshold.

8. A vehicle, characterized in that: The vehicle includes the pipe condensation detection device of the EGR system according to claim 7.

9. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 6.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-6.

Images