An image recognition method and device for road detection, an electronic device, and a medium

Abstract

The application relates to the field of road detection technology, in particular to an image recognition method and device for road detection, electronic equipment and a medium, which comprises the following steps: acquiring a to-be-detected region image; performing image preprocessing on the to-be-detected region image to obtain a pretreated image; performing feature extraction on the pretreated image, and judging whether suspicious features are contained in a plurality of features; if the suspicious features exist, determining a suspicious position according to the suspicious features; and generating an adjustment instruction according to the suspicious position, wherein the adjustment instruction controls an image acquisition device to move. The application has the effect of improving the accuracy of determining an abnormal position according to a detection image in a road detection process.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of road detection technology, and particularly relates to a road detection image recognition method and device, electronic equipment and medium. BACKGROUND

[0002] With the development of economy, the use frequency of highways is more and more frequent, so the degree of damage to the road is also more and more large. In the use process of the highway, the integrity of the roadbed (slope, retaining wall) and the facilities along the line (signs, protective facilities, etc.) may directly affect the traffic safety, so it is particularly crucial to overhaul the road and to carry out preventive maintenance on the road according to the overhaul result.

[0003] In the related art, the road overhaul is usually performed by a road overhaul vehicle to collect images of a road to be detected, to perform image recognition on the collected images, and to mark positions with abnormalities in the images, so as to facilitate maintenance personnel to overhaul. However, since the position of the image collection device on the road overhaul vehicle is fixed, the clarity of the image collection on the positions with possible abnormalities is low when the road is overhauled, which may reduce the accuracy of determining the abnormalities according to the collected images. SUMMARY

[0004] In order to improve the accuracy of determining abnormal positions according to collected images, the present application provides, in particular, a road detection image recognition method and device, electronic equipment and medium.

[0005] In a first aspect, the present application provides a road detection image recognition method, which adopts the following technical solution:

[0006] A road detection image recognition method comprises the following steps:

[0007] An image of a region to be detected is acquired.

[0008] The image of the region to be detected is preprocessed to obtain a preprocessed image.

[0009] Features of the preprocessed image are extracted, and it is determined whether the features contain suspicious features.

[0010] If there are suspicious features, suspicious positions are determined according to the suspicious features.

[0011] Adjustment instructions are generated according to the suspicious positions, and the adjustment instructions control the image collection device to adjust the position in multiple degrees of freedom.

[0012] By adopting the technical scheme, the image of the to-be-detected region is acquired, image preprocessing is performed on the to-be-detected region image to obtain a preprocessed image, feature extraction is performed on the preprocessed image, it is judged whether there is a suspicious feature from the extracted multiple features, if there is a suspicious feature, the corresponding suspicious position is determined according to the suspicious feature, and finally the adjustment instruction is generated according to the suspicious position. The adjustment instruction is used to control the image acquisition device to move. Through the abnormal position in the to-be-detected region image, the image acquisition device is adjusted in multiple degrees of freedom, the clarity of the image of the suspicious position is improved, and the accuracy of determining the abnormality in the road detection process is improved.

[0013] In a possible implementation manner, the image preprocessing of the to-be-detected region image comprises:

[0014] The to-be-detected region image is subjected to image enhancement processing to obtain an enhanced image.

[0015] The enhanced image is subjected to binarization processing to obtain the preprocessed image.

[0016] By adopting the technical scheme, the to-be-detected region image is subjected to image enhancement processing to obtain an enhanced image, and the enhanced image is subjected to binarization processing to obtain a preprocessed image. The image preprocessing of the to-be-detected region image facilitates improving the efficiency and accuracy of feature extraction through the preprocessed image.

[0017] In a possible implementation manner, the determination of the suspicious position according to the suspicious feature comprises:

[0018] The suspicious feature is marked, and a marked image is generated;

[0019] The marked image is imported into a pre-established coordinate system to determine coordinate information of the mark;

[0020] The suspicious position is determined according to the coordinate information.

[0021] By adopting the technical scheme, the suspicious feature is marked to generate a marked image, the marked image is imported into a pre-established coordinate system to determine coordinate information of the mark, and finally the suspicious position is determined. The suspicious position is determined by determining the mark coordinate of the suspicious mark, and the accuracy of determining the suspicious position is improved.

[0022] In a possible implementation manner, the generation of the adjustment instruction according to the suspicious position comprises:

[0023] A moving direction and a rotation angle are determined according to the suspicious position, and the moving direction and the rotation angle constitute a direction adjustment instruction.

[0024] determine target image acquisition parameters according to the to-be-detected region image, and generate displacement adjustment instructions according to the target image acquisition parameters;

[0025] The direction adjustment instructions and the displacement adjustment instructions constitute the adjustment instructions.

[0026] By adopting the technical solution, the moving direction and the rotation angle of the image acquisition device are determined through the suspicious position first, and the direction adjustment instructions are generated, then the moving displacement of the image acquisition device is determined through the target image acquisition parameters of the to-be-detected region image, and the displacement adjustment instructions are generated, the direction adjustment instructions and the displacement adjustment instructions jointly constitute the adjustment instructions, and the adjustment of the direction and the displacement of the image acquisition device helps to improve the definition when the image acquisition device acquires images of the suspicious position.

[0027] In a possible implementation manner, a detection type is determined according to the to-be-detected region image, and the detection type includes a slope and a flat ground.

[0028] When the detection type is the slope, a slope height is predicted according to the to-be-detected region image.

[0029] A height adjustment rate of the image acquisition device is determined based on a moving rate of the image acquisition device and the predicted slope height.

[0030] The height of the image acquisition device is adjusted based on the height adjustment rate.

[0031] By adopting the technical solution, the detection type of the to-be-detected region image is determined first, the detection type includes the slope and the flat ground, if the detection type of the to-be-detected region image is the slope, the height of the slope is predicted according to the to-be-detected image, and the height adjustment rate of the image acquisition device is determined based on the preset height and the moving rate of the image acquisition, that is, the moving rate of the road detection vehicle, and the height of the image acquisition device is adjusted based on the height adjustment rate, and the height adjustment of the image acquisition device facilitates to improve the acquisition range of the to-be-detected region image.

[0032] In a possible implementation manner, the to-be-detected region image is regionally divided according to a preset rule, and a plurality of region images are formed.

[0033] A preset standard comparison image of each region image is determined according to a preset comparison library.

[0034] Whether a to-be-detected object in the to-be-detected region image is complete is determined according to the preset standard comparison image.

[0035] If not, an incomplete position is determined, and a repair instruction is generated according to the incomplete position.

[0036] By adopting the technical scheme, the to-be-detected region image is divided into multiple region images according to a preset rule, a preset standard comparison image of each region image is determined according to a preset comparison library, the integrity of the to-be-detected object in each region image is judged according to the corresponding preset standard comparison image of each region image, when the to-be-detected object in the region image is incomplete, the incomplete position is determined and a repair instruction is generated according to the incomplete position, the to-be-detected image is divided into region images, and the incomplete position is determined according to the region image, thereby improving the accuracy of determining the incomplete position.

[0037] In a possible implementation manner, the judging whether the to-be-detected object in the to-be-detected region image is complete includes:

[0038] determining a to-be-detected object region in each region image and determining initial boundary information of the to-be-detected object region;

[0039] importing the initial boundary information and preset boundary information into a preset coordinate system, performing coincidence matching, and generating a matching value, wherein the preset boundary information is boundary information corresponding to a preset to-be-detected object region in a preset standard comparison image;

[0040] judging whether the to-be-detected object in each region image is complete according to the matching value.

[0041] By adopting the technical scheme, the to-be-detected object region in each region image is determined, the initial boundary information of the to-be-detected object region is determined, the initial boundary information is matched with the preset boundary information in the preset standard comparison image to generate a matching value, and finally whether the to-be-detected object in the region image is complete is judged according to the generated matching value, thereby improving the accuracy of judging whether the to-be-detected object is complete.

[0042] In a second aspect, the present application provides an image recognition device for road detection, which adopts the following technical scheme:

[0043] An image recognition device for road detection includes:

[0044] An acquisition module is configured to acquire a to-be-detected region image.

[0045] A preprocessing module is configured to perform image preprocessing on the to-be-detected region image to obtain a preprocessed image.

[0046] A feature extraction module is configured to perform feature extraction on the preprocessed image and judge whether suspicious features are included in multiple features.

[0047] A position determination module is configured to determine a suspicious position according to the suspicious features if the suspicious features exist.

[0048] The generating instruction module is configured to generate an adjusting instruction according to the suspicious position, and the adjusting instruction controls the image acquisition device to perform multi-degree-of-freedom displacement adjustment.

[0049] By using the above technical solution, the image of the to-be-detected region is acquired, and the to-be-detected region image is preprocessed to obtain a preprocessed image. Then, feature extraction is performed on the preprocessed image, and it is determined whether there is a suspicious feature from the extracted multiple features. If there is a suspicious feature, the corresponding suspicious position is determined according to the suspicious feature. Finally, an adjusting instruction is generated according to the suspicious position. The adjusting instruction is used to control the image acquisition device to move. Through the abnormal position in the to-be-detected region image, multi-degree-of-freedom position adjustment of the image acquisition device is realized, the clarity of the image of the suspicious position is improved, and the accuracy of determining the abnormality in the road detection process is improved.

[0050] In a third aspect, the present application provides an electronic device, which adopts the following technical solution:

[0051] An electronic device, comprising:

[0052] at least one processor;

[0053] a memory;

[0054] at least one application program, wherein the at least one application program is stored in the memory and is configured to be executed by the at least one processor, and the at least one application program is configured to execute the above-mentioned road detection image recognition method.

[0055] In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution:

[0056] A computer-readable storage medium, comprising: a computer program stored therein and capable of being loaded and executed by a processor to execute the above-mentioned road detection image recognition method.

[0057] In summary, the present application has at least one of the following beneficial technical effects:

[0058] 1. By acquiring the image of the to-be-detected region and pre-processing the to-be-detected region image to obtain a preprocessed image, feature extraction is performed on the preprocessed image, and it is determined whether there is a suspicious feature from the extracted multiple features. If there is a suspicious feature, the corresponding suspicious position is determined according to the suspicious feature. Finally, an adjusting instruction is generated according to the suspicious position. The adjusting instruction is used to control the image acquisition device to move. Through the abnormal position in the to-be-detected region image, multi-degree-of-freedom position adjustment of the image acquisition device is realized, the clarity of the image of the suspicious position is improved, and the accuracy of determining the abnormality in the road detection process is improved.

[0059] 2. By judging the detection type of the image of the to-be-detected area, the detection type is a slope and a flat ground, if the detection type of the image of the to-be-detected area is a slope, the height of the slope is predicted according to the to-be-detected image, and the height adjustment rate of the image acquisition device is determined based on a preset height and a moving rate of the image acquisition device, that is, a moving rate of the road detection vehicle, and the height of the image acquisition device is adjusted based on the height adjustment rate, so that the collection range of the image of the to-be-detected area is facilitated to be improved by adjusting the height of the image acquisition device. BRIEF DESCRIPTION OF DRAWINGS

[0060] Figure 1 is a flowchart of a road detection image recognition method in an embodiment of the present application;

[0061] Figure 2 is a structural schematic diagram of a road detection image recognition device in an embodiment of the present application;

[0062] Figure 3 is a structural schematic diagram of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0063] The following will be described in detail with reference to the accompanying drawings. Figures 1-3 The present application will be further described in detail.

[0064] Those skilled in the art can make modifications to the present embodiments without creative contribution after reading the present specification, but as long as the modifications are within the scope of the claims of the present application, they are protected by the patent law.

[0065] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme of the embodiments of the present application will be described clearly and completely below with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0066] At present, the domestic road condition detection vehicle basically adopts fixed camera, fixed focal length and fixed image acquisition mode, which has been unable to meet the on-site detection needs of highways, especially mountainous road sections, high fill road sections and disaster-bearing bodies. It cannot effectively cover all ranges of the detection object, and because the road condition detection vehicle cannot stop when collecting images of the to-be-detected area, it cannot clearly locate some positions suspected to have abnormalities, so the accuracy may be low when judging whether there are abnormalities in the image based on the collected image.

[0067] In order to improve the accuracy of determining the abnormal position according to the image, the embodiment of the application obtains the image of the to-be-detected region, and obtains a preprocessed image by preprocessing the image of the to-be-detected region, then extracts features from the preprocessed image, judges whether there is a suspicious feature from the extracted features, if there is a suspicious feature, determines the corresponding suspicious position according to the suspicious feature, and finally generates an adjustment instruction according to the suspicious position, the adjustment instruction is used to control the image acquisition device to move, through the abnormal position in the image of the to-be-detected region, the image acquisition device is adjusted in multiple degrees of freedom, the clarity of the image of the suspicious position is improved, and the accuracy of determining the abnormal position in the road detection process is improved.

[0068] Specifically, the embodiment of the application provides an image recognition method for road detection, which is executed by an electronic device. The electronic device can be a server or a terminal device. The server can be a physical server, a server cluster composed of multiple physical servers, or a distributed system. It can also be a cloud server providing cloud computing services. The terminal device can be a smartphone, a tablet computer, a notebook computer, a desktop computer, etc., but is not limited thereto. The terminal device and the server can be directly or indirectly connected through wired or wireless communication, and the embodiment of the application does not limit this.

[0069] Reference Figure 1 , Figure 1 is a flowchart of an image recognition method for road detection in the embodiment of the application. The method comprises steps S110, S120, S130, S140, and S150.

[0070] Step S110: Obtain an image of a to-be-detected region.

[0071] Specifically, the to-be-detected region can be a mountainous road section, a high fill road section, and a general road section. Any road section that needs to be evaluated for road technical conditions can be used. The image is obtained by a road condition detection vehicle with an image acquisition device. The road condition detection vehicle is provided with multiple image acquisition devices, which facilitate image acquisition of both sides and the front of the to-be-detected region.

[0072] The image acquisition device is a high-resolution CCD mainstream matrix camera and a high-quality lens, wherein the CCD mainstream matrix camera is a large-scale integrated circuit process made semiconductor photoelectric element, has a powerful self-scanning function, good image definition, can capture images, supports multiple merged pixel modes, innovative readout technology can sufficiently reduce noise, achieve a higher sensitivity and conversion effect, and the CCD mainstream matrix camera has the advantages of small size, high reliability, high sensitivity, strong light resistance, shock resistance, magnetic field resistance, small distortion, long service life, clear image, simple operation and the like, and is widely used in the image detection field.

[0073] When the image of the to-be-detected region is acquired, the video stream can be frame-extracted according to a preset frame extraction frequency, or the image of the to-be-detected region can be acquired according to a preset frequency. The specific image acquisition mode is not limited in the embodiments of the present application.

[0074] Step S120: performing image preprocessing on the image of the to-be-detected region to obtain a preprocessed image.

[0075] Specifically, the preprocessing of the image of the to-be-detected region can improve the clarity of the image, and the processing of the image helps to improve the accuracy of determining the anomaly in the image.

[0076] Common image preprocessing methods for the image of the to-be-detected region can include grayscale value processing, binaryzation processing, dilation processing, erosion processing and cutting processing. When cutting the to-be-processed image, edge detection needs to be performed on the preprocessed image, the preprocessed image is cut through the edge of the preprocessed image, some non-key region and surrounding environment interference image region are removed, and a region of interest image is obtained.

[0077] Step S130: performing feature extraction on the preprocessed image, and determining whether the multiple features contain a suspicious feature.

[0078] Specifically, when the preprocessed image is subjected to feature extraction, all features in the preprocessed image can be extracted through a convolutional neural network. All features that can exist in the image of the to-be-detected region include road signs and signboards, and the suspicious features are road signs and signboards that can have anomalies, i.e., incomplete road signs and damaged signboards. When determining whether the multiple features contain a suspicious feature, a suspicious feature recognition model can be used for judgment. The multiple features extracted are introduced into the trained suspicious feature recognition model, and the suspicious feature contained in the multiple features can be recognized. The training process of the suspicious feature recognition model is to train the feature recognition model by using a large number of samples with artificial labels. The artificial label is the anomaly type of the corresponding feature, including incompleteness and damage.

[0079] Step S140: If there is a suspicious feature, a suspicious position is determined according to the suspicious feature.

[0080] Step S150: An adjustment instruction is generated according to the suspicious position, and the adjustment instruction controls the image acquisition device to adjust the position in multiple degrees of freedom.

[0081] Specifically, the suspicious position is the position of the suspicious feature in the image of the to-be-detected region, and the suspicious position is determined by the position of the suspicious feature in the image of the to-be-detected region. Since the device used for image acquisition of the to-be-detected region is a road detection vehicle provided with an image acquisition device, and since multiple image acquisition devices are provided on the road detection vehicle, the position of the suspicious feature is determined to facilitate subsequent adjustment of the angle of the image acquisition device through the position of the suspicious feature. The adjustment instruction is generated by an electronic device, and the generated adjustment instruction is used to control the image acquisition device to adjust.

[0082] In the embodiment of the application, the image of the to-be-detected region is acquired, and the image of the to-be-detected region is preprocessed to obtain a preprocessed image. Then, feature extraction is performed on the preprocessed image, and it is determined whether there is a suspicious feature from the extracted multiple features. If there is a suspicious feature, the corresponding suspicious position is determined according to the suspicious feature. Finally, an adjustment instruction is generated according to the suspicious position, and the adjustment instruction is used to control the image acquisition device to move. Through the abnormal position in the image of the to-be-detected region, the image acquisition device is adjusted in multiple degrees of freedom, the clarity of the image of the suspicious position is improved, and the accuracy of determining abnormalities in the road detection process is improved.

[0083] Further, the image of the to-be-detected region is preprocessed to obtain a preprocessed image in step S120, which can specifically include steps S1201 (not shown in the figure), S1202 (not shown in the figure), wherein:

[0084] Step S1201: The image of the to-be-detected region is enhanced to obtain an enhanced image.

[0085] Specifically, for the to-be-detected region image acquired in a high-noise environment, the image is sometimes too dark or too bright. The enhancement processing of the to-be-detected region image can make the effective region and the environment region in the to-be-detected region image have a large difference in gray value, thereby helping to separate the effective region in the to-be-detected region from the background and eliminate more interference.

[0086] Step S1202: The enhanced image is binarized to obtain a preprocessed image.

[0087] Specifically, the enhanced image is binarized according to a preset threshold, and the image after binarization processing presents an obvious black and white effect.

[0088] In the embodiment of the present application, the image enhancement processing is performed on the image of the to-be-detected region to obtain an enhanced image, and then the enhanced image is binarized to obtain a preprocessed image, so that the image preprocessing is performed on the image of the to-be-detected region, thereby facilitating the efficiency and accuracy of feature extraction through the preprocessed image.

[0089] Further, the step S140 of determining the suspicious position according to the suspicious feature can specifically include a step S1401 (not shown in the drawings), a step S1402 (not shown in the drawings), and a step S1403 (not shown in the drawings), wherein:

[0090] The step S1401 is to mark the suspicious feature and generate a marked image.

[0091] Specifically, the marking form can be marking by color or marking by text, and the specific marking form is not limited in the embodiment of the present application, as long as the marking of the suspicious feature can be realized. The marked image contains the suspicious feature and the marking corresponding to the suspicious feature.

[0092] The step S1402 is to import the marked image into a pre-established coordinate system to determine the coordinate information of the marking.

[0093] The step S1403 is to determine the suspicious position according to the coordinate information.

[0094] Specifically, the coordinate information of the marking includes the horizontal coordinate of the marking and the vertical coordinate of the marking, and the position of the marking is determined through the coordinate information, that is, the suspicious position is determined. The suspicious position can or can not exist abnormality, and the accuracy of determining the abnormality is improved by determining the suspicious position.

[0095] In the embodiment of the present application, the suspicious feature is marked to generate a marked image, and then the coordinate information of the marking is determined by importing the marked image into a pre-established coordinate system, and finally the suspicious position is determined. The coordinate information of the marking of the suspicious marking is determined to determine the suspicious position, thereby improving the accuracy of determining the suspicious position.

[0096] Further, the step S150 of generating the adjustment instruction according to the suspicious position can specifically include a step S1501 (not shown in the drawings), a step S1502 (not shown in the drawings), and a step S1503 (not shown in the drawings), wherein:

[0097] The step S1501 is to determine the moving direction and the rotating angle according to the suspicious position, and the moving direction and the moving angle constitute the direction adjustment instruction.

[0098] Specifically, for example, the suspicious position is in the lower right corner region of the to-be-detected region image, the moving direction is right turn, and the turning angle can be determined according to the coordinate information of the suspicious position. The direction adjustment instruction is generated by the electronic device and is used to control the image acquisition device to turn. The image acquisition device that turns can be the initial acquisition camera or an acquisition camera other than the initial acquisition camera.

[0099] The road detection vehicle is provided with a plurality of image acquisition devices, different image acquisition devices are movable, and each has a corresponding detection region. When the acquisition device needs to turn, in order to reduce the influence on subsequent image acquisition of the next road section, if the angle of the acquisition device that needs to turn exceeds a preset threshold, a standby image acquisition device is started to implement image acquisition of the suspicious position.

[0100] Step S1502: determining a target image acquisition parameter according to the to-be-detected region image, and generating a displacement adjustment instruction according to the target image acquisition parameter.

[0101] Step S1503: the direction adjustment instruction and the displacement adjustment instruction constitute an adjustment instruction.

[0102] Specifically, the target image acquisition parameter can include the focal length, exposure mode, gain, chroma, and definition of the target image. The to-be-detected region image is judged to be clear or not through the target image acquisition parameter. Different exposure modes have different effects on image quality. If the generated image is overexposed, the image brightness is high, and then many details can be lost. On the contrary, if the generated image is underexposed, the image brightness is low, and the details of the image can also be lost. The definition is the definition of the suspicious position when the suspicious position in the to-be-detected region image is determined. The Brenner gradient function, Tenengrad gradient function, variance function, and energy gradient function can all be used to determine the target definition of the image. When the Tenengrad gradient function is used to determine the target definition of the image, the Sobel operator is used to extract the horizontal and vertical gradient values of the suspicious region, and then the square sum is calculated to determine the target definition.

[0103] The height adjustment instruction is used to control the image acquisition device to change in height, so as to improve the target definition of the to-be-detected region image by increasing or decreasing the height of the image acquisition device.

[0104] In the embodiments of the present application, the moving direction and the rotation angle of the image acquisition device are first determined through the suspicious position, and a direction adjustment instruction is generated. Then, the moving displacement of the image acquisition device is determined through the target image acquisition parameter of the image of the to-be-detected area, and a displacement adjustment instruction is generated. The direction adjustment instruction and the displacement adjustment instruction jointly constitute the adjustment instruction. By adjusting the direction and displacement of the image acquisition device, the clarity of the image acquisition device when acquiring images of the suspicious position is improved.

[0105] Further, in order to expand the image acquisition range, the embodiments of the present application further include steps S1 (not shown in the figure), S2 (not shown in the figure), S3 (not shown in the figure), and S4 (not shown in the figure), wherein:

[0106] Step S1: determining the detection type according to the image of the to-be-detected area.

[0107] The detection type includes a slope and a flat ground.

[0108] Specifically, the slope includes an upper slope and a lower slope. The detection type can be determined by judging whether the preset reference object exists in the image of the to-be-detected area. The preset reference object can be a railing or a bridge pier. The specific reference object is not limited in the embodiments of the present application, as long as the detection type of the image of the to-be-detected area can be determined according to the preset reference object.

[0109] Step S2: when the detection type is a slope, predicting the slope height according to the image of the to-be-detected area.

[0110] Specifically, the slope height can be predicted by determining the size of the reference object in the image of the to-be-detected area. By predicting the slope height, the height limit value of lifting or lowering can be determined. When the road detection vehicle is on an uphill, the height of the image acquisition device needs to be lifted. When the road detection vehicle is on a downhill, the height of the image acquisition device needs to be lowered. If the height of the image acquisition device is not adjusted, the coverage range of the image of the to-be-detected area may be reduced during uphill or downhill.

[0111] Step S3: determining the height adjustment rate of the image acquisition device based on the moving rate of the image acquisition device and the predicted slope height.

[0112] Step S4: adjusting the height of the image acquisition device based on the height adjustment rate.

[0113] Specifically, the moving speed of the image acquisition device, i.e., the moving speed of the road detection vehicle, can be calculated according to the positions and times of the same object appearing in the plurality of images to be detected, or the actual moving speed of the road detection vehicle can be directly obtained. The specific obtaining method is not limited in the embodiments of the present application, as long as the moving speed of the road detection vehicle can be determined.

[0114] The height adjustment speed is the height adjustment speed of raising or lowering the position of the image acquisition device to the height limit.

[0115] In the embodiments of the present application, the detection type of the image to be detected is first determined, the detection type is a slope and a flat ground, if the detection type of the image to be detected is a slope, the height of the slope is predicted according to the image to be detected, and the height adjustment speed of the image acquisition device is determined based on the preset height and the moving speed of the image acquisition, i.e., the moving speed of the road detection vehicle, and the height of the image acquisition device is adjusted based on the height adjustment speed, so as to improve the collection range of the image to be detected by adjusting the height of the image acquisition device.

[0116] Further, it further includes steps Sa (not shown in the figure), step Sb (not shown in the figure), step Sc (not shown in the figure), and step Sd (not shown in the figure), wherein:

[0117] Step Sa: The image to be detected is divided into a plurality of region images according to a preset rule.

[0118] Specifically, the region to be detected can be a flat ground or a slope. The preset rule is to divide the image to be detected according to the image proportion of the image to be detected, which can be modified according to requirements, and is not limited in the embodiments of the present application, as long as the image to be detected can be divided. By dividing the image to be detected, the accuracy of comparison can be improved.

[0119] Step Sb: Determine the preset standard comparison image of each region image according to the preset comparison library.

[0120] Specifically, the preset comparison library stores a plurality of region images, and the preset comparison library stores a plurality of preset standard comparison images. The preset standard comparison image is a historical region image formed by dividing a historical region to be detected according to a preset rule in a historical road detection process. The preset standard comparison image of each region image can be determined from a plurality of historical region images by using a target recognition algorithm. The content of the preset comparison library can be added or deleted according to requirements.

[0121] Step Sc: Determine whether the object to be detected in the image to be detected is complete according to the preset standard comparison image.

[0122] Specifically, the preset standard contrast image contains a preset to-be-detected object similar to the region image, the to-be-detected region is matched with the preset to-be-detected region, and then it is judged whether the to-be-detected object in the to-be-detected region image is complete. There are many ways to match the to-be-detected region with the preset to-be-detected region, for example, the boundary information corresponding to the to-be-detected region and the preset to-be-detected region can be determined respectively, and the respective boundary information is matched; or the region image and the preset standard image are superimposed, and the superimposed region area is matched.

[0123] Step Sd: if not complete, determining an incomplete position, and generating a repair instruction according to the incomplete position.

[0124] Specifically, the incomplete position is a position in the region image where the to-be-detected region does not match the preset to-be-detected region in the preset standard contrast image, that is, a position in the region image that needs to be repaired. The repair instruction is used to remind the relevant staff to repair the incomplete position in time, thereby reducing the probability of traffic accidents caused by not repairing in time.

[0125] In the embodiment of the application, the to-be-detected region image is divided into a plurality of region images according to a preset rule, the preset standard contrast image of each region image is determined according to a preset comparison library, the completeness of the to-be-detected object in each region image is judged according to the preset standard contrast image corresponding to each region image, when the to-be-detected object in the region image is not complete, the incomplete position is determined and the repair instruction is generated according to the incomplete position, the to-be-detected image is divided into region images, and the accuracy of determining the incomplete position is improved by determining the incomplete position according to the region image.

[0126] Further, the step Sc of judging whether the to-be-detected object in the to-be-detected region image is complete can specifically include steps Sc1 (not shown in the drawing), Sc2 (not shown in the drawing), and Sc3 (not shown in the drawing), wherein:

[0127] Step Sc1: determining a to-be-detected region in each region image, and determining initial boundary information of the to-be-detected region.

[0128] Specifically, the initial boundary information is a plurality of boundary point coordinates of the to-be-detected region, and the region connected by the plurality of boundary point coordinates is the to-be-detected region. The boundary point information can also be determined by an OpenCV (Open Source Computer Vision Library) function. In the embodiment of the application, the way of determining the boundary information in the image is not limited as long as the boundary information can be determined.

[0129] Step Sc2: Import the initial boundary information and the preset boundary information into the preset coordinate system, perform coincidence matching, and generate matching values.

[0130] Among them, the preset boundary information is the boundary information corresponding to the preset detection object region in the preset standard comparison image.

[0131] Step Sc3: Determine whether the object to be detected in each region image is complete based on the matching value.

[0132] Specifically, the objects to be detected can be road signs, roadbeds, and roadside traffic facilities. When the matching value is equal to or higher than the preset standard matching value, it is determined that the object to be detected in the area image is complete; when the matching value is lower than the preset standard matching value, it is determined that the object to be detected in the area image is incomplete. The preset standard matching value can be modified according to needs and is not specifically limited in this embodiment.

[0133] In this embodiment, the detection area in each region image is determined, and the initial boundary information of the detection area is determined. The initial boundary information is matched with the preset boundary information in the preset standard comparison image to generate a matching value. Finally, the accuracy of determining whether the detection object in the region image is complete is improved based on the generated matching value.

[0134] The above embodiments describe an image recognition method for road detection from the perspective of method flow. The following embodiments describe an image recognition device for road detection from the perspective of virtual module or virtual unit. For details, please refer to the following embodiments.

[0135] This application provides an image recognition device for road detection, such as... Figure 2 As shown, the device may specifically include an acquisition module 210, a preprocessing module 220, a feature extraction module 230, and a generation instruction module 240, wherein:

[0136] The acquisition module 210 is used to acquire the image of the region to be detected;

[0137] The preprocessing module 220 is used to perform image preprocessing on the image of the region to be detected to obtain a preprocessed image;

[0138] The feature extraction module 230 is used to extract features from the preprocessed image and determine whether there are suspicious features among multiple features;

[0139] Location determination module 240 is used to determine a suspicious location based on suspicious features if suspicious features are found.

[0140] The instruction generation module 250 is used to generate adjustment instructions based on the suspected location, and the adjustment instructions control the image acquisition device to perform multi-degree-of-freedom displacement adjustment.

[0141] In a possible implementation, the preprocessing module 220 includes:

[0142] a first preprocessing unit, configured to perform image enhancement processing on the to-be-detected region image to obtain an enhanced image;

[0143] a second preprocessing unit, configured to perform binarization processing on the enhanced image to obtain a preprocessed image.

[0144] In a possible implementation, the position determination module 240 includes:

[0145] a marking unit, configured to mark the suspicious feature and generate a marked image;

[0146] a coordinate determination unit, configured to import the marked image into a pre-established coordinate system and determine coordinate information of the marked image;

[0147] a suspicious position determination unit, configured to determine a suspicious position according to the coordinate information.

[0148] In a possible implementation, the instruction generation module 250 includes:

[0149] a direction adjustment instruction determination unit, configured to determine a moving direction and a rotation angle according to the suspicious position, and combine the moving direction and the moving angle to form a direction adjustment instruction;

[0150] a displacement adjustment instruction determination unit, configured to determine a target image acquisition parameter according to the to-be-detected region image, and generate a displacement adjustment instruction according to the target image acquisition parameter;

[0151] an instruction combination unit, configured to combine the direction adjustment instruction and the displacement adjustment instruction to form an adjustment instruction.

[0152] In a possible implementation, the method further includes:

[0153] a detection type determination module, configured to determine a detection type according to the to-be-detected region image, the detection type including a slope and a flat ground;

[0154] a height prediction module, configured to predict a slope height according to the to-be-detected region image when the detection type is the slope;

[0155] a height adjustment rate determination module, configured to determine a height adjustment rate of the image acquisition device based on a moving rate of the image acquisition device and the predicted slope height;

[0156] an adjustment module, configured to adjust the height of the image acquisition device based on the height adjustment rate.

[0157] In a possible implementation, the method further includes:

[0158] a region division module, configured to divide the to-be-detected region image according to a preset rule to form a plurality of region images;

[0159] a standard contrast image determination module, configured to determine a preset standard contrast image of each region image according to a preset contrast library;

[0160] a judgment module, configured to judge whether the to-be-detected object in the to-be-detected region image is complete according to the preset standard contrast image;

[0161] a generation instruction module, configured to, if the to-be-detected object is not complete, determine an incomplete position and generate a repair instruction according to the incomplete position.

[0162] In a possible implementation manner, the judgment module comprises:

[0163] a determination initial boundary unit, configured to determine a to-be-detected object region in each region image and determine initial boundary information of the to-be-detected object region;

[0164] a matching unit, configured to import the initial boundary information and preset boundary information into a preset coordinate system, perform coincidence matching, and generate a matching value, wherein the preset boundary information is boundary information corresponding to a preset to-be-detected object region in the preset standard contrast image;

[0165] a completeness judgment unit, configured to judge whether the to-be-detected object in each region image is complete according to the matching value.

[0166] An electronic device is provided in the embodiments of the present application, as shown in Figure 3 as shown in Figure 3 The electronic device 300 shown in the embodiment comprises a processor 301 and a memory 303. The processor 301 and the memory 303 are connected, for example, through a bus 302. Optionally, the electronic device 300 can further comprise a transceiver 304. It should be noted that the transceiver 304 is not limited to one in actual application, and the structure of the electronic device 300 does not constitute a limitation on the embodiments of the present application.

[0167] The processor 301 can be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic device, transistor logic device, hardware component, or any combination thereof. It can implement or execute various exemplary logical blocks, modules and circuits described in connection with the disclosure. The processor 301 can also be a combination of computing functions, such as one or more microprocessor combinations, combinations of DSP and microprocessor, etc.

[0168] The bus 302 can include a path for transmitting information between the above-mentioned components. The bus 302 can be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus 302 can be divided into an address bus, a data bus, a control bus, etc. For convenience of representation, Figure 3 Only one thick line is used in the middle, but it does not mean that there is only one bus or one type of bus.

[0169] The memory 303 can be a ROM (Read Only Memory) or other type of static storage device that can store static information and instructions, a RAM (Random Access Memory) or other type of dynamic storage device that can store information and instructions, an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, an optical disk storage (including a compact disk, a laser disk, an optical disk, a digital versatile disk, a Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and capable of being accessed by a computer, but not limited to.

[0170] The memory 303 is configured to store application program codes for implementing the solutions of the present application, and the processor 301 is configured to execute the application program codes stored in the memory 303.

[0171] The electronic device includes, but is not limited to, a mobile terminal such as a mobile phone, a notebook computer, a digital broadcast receiver, a PDA (Personal Digital Assistant), a PAD (Tablet Personal Computer), a PMP (Portable Multimedia Player), a car terminal (for example, a car navigation terminal), and the like, and a stationary terminal such as a digital TV, a desktop computer, and the like. It can also be a server or the like. Figure 3 The electronic device shown is merely an example and should not impose any limitation on the functions and use range of the embodiments of the present disclosure.

[0172] The computer readable storage medium of the embodiments of the present application stores a computer program, and when the computer program runs on a computer, the computer can execute the corresponding content in the foregoing method embodiments. Compared with the related art, in the embodiments of the present application, an image of a to-be-detected region is acquired, and an image pre-processing is performed on the to-be-detected region image to obtain a pre-processed image, then a feature extraction is performed on the pre-processed image, and it is judged from the extracted multiple features whether there is a suspicious feature, if there is a suspicious feature, a corresponding suspicious position is determined according to the suspicious feature, and finally an adjustment instruction is generated according to the suspicious position, the adjustment instruction is used to control the image acquisition device to move, through the abnormal position in the to-be-detected region image, the image acquisition device is adjusted in multiple degrees of freedom, the clarity of the image of the suspicious position is improved, and the accuracy of determining the abnormality in the road detection process is improved.

[0173] It should be understood that, although each step in the flowchart of the accompanying drawings is shown in sequence according to the direction of the arrow, these steps are not necessarily executed in sequence according to the direction of the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and they can be executed in other sequences. Moreover, at least part of the steps in the flowchart of the accompanying drawings can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or sub-steps or stages of other steps.

[0174] The above only describes some embodiments of the present application. It should be noted that, for those skilled in the art, without departing from the principles of the present application, some improvements and refinements can be made, which should also be considered as the protection scope of the present application.

Claims

1. An image recognition method for road detection, characterized by, The method comprises the following steps: acquiring an image of a region to be detected; performing image preprocessing on the image of the region to be detected to obtain a preprocessed image; extracting features from the preprocessed image and determining whether the features include suspicious features, wherein the suspicious features are incomplete road signs and damaged signs; if the suspicious features exist, determining a suspicious position according to the suspicious features, wherein the suspicious position is a position of the suspicious features in the image of the region to be detected; generating an adjustment instruction according to the suspicious position, wherein the adjustment instruction controls a multi-degree-of-freedom position adjustment of an image acquisition device; wherein the generation of the adjustment instruction according to the suspicious position comprises: determining a moving direction and a rotating angle according to the suspicious position, wherein the moving direction and the rotating angle constitute a direction adjustment instruction; determining a target image acquisition parameter according to the image of the region to be detected, and generating a displacement adjustment instruction according to the target image acquisition parameter; the direction adjustment instruction and the displacement adjustment instruction constitute the adjustment instruction; wherein the method further comprises: determining a detection type according to the image of the region to be detected, wherein the detection type includes a slope and a flat ground; when the detection type is a slope, predicting a slope height according to the image of the region to be detected, determining a height limit value of lifting or lowering according to the predicted slope height; determining a height adjustment rate of the image acquisition device based on a moving rate of the image acquisition device and the predicted slope height; adjusting the height of the image acquisition device based on the height adjustment rate, wherein the height adjustment rate is an adjustment rate when the position of the image acquisition device is lifted or lowered to the corresponding height limit value; wherein the method further comprises: dividing the image of the region to be detected into a plurality of region images according to a preset rule, determining a preset standard comparison image for each region image according to a preset comparison library, determining whether a detected object in the image of the region to be detected is complete according to the preset standard comparison image, determining an incomplete position if the detected object is incomplete, and generating a repair instruction according to the incomplete position, wherein the incomplete position is a position where a detected object region in a region image does not match a preset detected object region in a preset standard comparison image, and the repair instruction is used to remind relevant staff to repair the incomplete position in time; wherein the determination of whether the detected object in the image of the region to be detected is complete comprises: determining a detected object region in each region image and determining initial boundary information of the detected object region; importing the initial boundary information and preset boundary information into a preset coordinate system, performing coincidence matching, and generating a matching value, wherein the preset boundary information is boundary information corresponding to a preset detected object region in a preset standard comparison image; determining whether the detected object in each region image is complete according to the matching value.

2. The image recognition method for road detection according to claim 1, wherein The image preprocessing on the image of the region to be detected to obtain a preprocessed image comprises: performing image enhancement processing on the image of the region to be detected to obtain an enhanced image; performing binaryzation processing on the enhanced image to obtain the preprocessed image.

3. The image recognition method for road detection according to claim 1, characterized by, The determination of the suspicious position according to the suspicious features comprises: labeling the suspicious features and generating a labeled image; Import the mark image into a pre-established coordinate system to determine coordinate information of the mark; Determine a suspicious position according to the coordinate information.

4. An image recognition device for road detection, characterized by comprising: The method comprises the steps of: An acquisition module is configured to acquire a region image to be detected; A preprocessing module is configured to perform image preprocessing on the region image to be detected to obtain a preprocessed image; A feature extraction module is configured to perform feature extraction on the preprocessed image and determine whether suspicious features exist in the features, wherein the suspicious features are incomplete road signs and damaged signboards; A position determination module is configured to determine a suspicious position according to the suspicious features if the suspicious features exist, wherein the suspicious position is a position of the suspicious features in the region image to be detected; A generation instruction module is configured to generate an adjustment instruction according to the suspicious position, wherein the adjustment instruction controls a multi-degree-of-freedom displacement adjustment of an image acquisition device; The generation instruction module comprises: A direction adjustment instruction unit is configured to determine a moving direction and a rotating angle according to the suspicious position, wherein the moving direction and the rotating angle constitute a direction adjustment instruction; A displacement adjustment instruction unit is configured to determine a target image acquisition parameter according to the region image to be detected and generate a displacement adjustment instruction according to the target image acquisition parameter; An instruction combination unit is configured to combine the direction adjustment instruction and the displacement adjustment instruction to form the adjustment instruction; The device further comprises: A detection type determination module is configured to determine a detection type according to the region image to be detected, wherein the detection type comprises a slope and a flat ground; A height prediction module is configured to predict a slope height according to the region image to be detected when the detection type is a slope, and determine a height limit value of lifting or lowering according to the predicted slope height; A height adjustment rate determination module is configured to determine a height adjustment rate of the image acquisition device based on a moving rate of the image acquisition device and the predicted slope height; An adjustment module is configured to adjust a height of the image acquisition device based on the height adjustment rate, wherein the height adjustment rate is an adjustment rate when the position of the image acquisition device is lifted or lowered to a corresponding height limit value; A region division module is configured to divide the region image to be detected into a plurality of region images according to a preset rule; A standard comparison image determination module is configured to determine a preset standard comparison image of each region image according to a preset comparison library; A judgment module is configured to determine whether a to-be-detected object in the region image to be detected is complete according to the preset standard comparison image; A generation instruction module is configured to determine an incomplete position and generate a repair instruction according to the incomplete position if the to-be-detected object is incomplete, wherein the incomplete position is a position where a to-be-detected object region in the region image does not match a preset to-be-detected object region in the preset standard comparison image, and the repair instruction is used to remind relevant staff to repair the incomplete position in time; The judging module is configured to determine a to-be-detected object region in each region image and determine initial boundary information of the to-be-detected object region, import the initial boundary information and preset boundary information into a preset coordinate system, perform coincidence matching, and generate a matching value, wherein the preset boundary information is boundary information corresponding to a preset to-be-detected object region in a preset standard comparison image, and determine whether the to-be-detected object in each region image is complete according to the matching value.

5. An electronic device, comprising: The electronic device includes: at least one processor; a memory; at least one application program, wherein the at least one application program is stored in the memory and is configured to be executed by the at least one processor, and the at least one application program is configured to execute the image recognition method for road detection according to any one of claims 1-3.

6. A computer-readable storage medium, characterized in that, including: a computer program stored in the memory and capable of being loaded and executed by the processor to perform the image recognition method for road detection according to any one of claims 1-3.

Images