A method for automatic detection and modeling of motor vehicle driver test sites
Through the drone collecting data and building a three-dimensional model, the inspection of the motor vehicle driver examination site is automatically completed, solving the problems of long and poor accuracy of manual inspection, and achieving an efficient and accurate acceptance process.
Patent Information
- Application Number
- CN202010245228.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-03-31
- Publication Date
- 2024-11-19
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The acceptance inspection and inspection of existing motor vehicle drivers mainly relies on manual measurement, resulting in large inspection workload, high time cost, poor accuracy, and may take 2 to 4 days or more.
UAVs are used to collect image data and laser point cloud data, build a three-dimensional model through an image processing cloud platform, and compare it with a pre-established acceptance standard model to automatically complete the inspection.
The automation and digitalization of examination room inspection has been realized, which significantly saves time, improves detection efficiency and accuracy, and ensures the objectivity and fairness of the inspection.
Smart Images

Figure CN111580128B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of image processing technology, and in particular to a method for automatic detection and modeling of a motor vehicle driver examination site. Background Art
[0002] With the continuous and rapid growth of the number of motor vehicles, cars have become a means of transportation for ordinary people, and the demand for taking the motor vehicle driver test to obtain a driver's license has increased year by year. The motor vehicle driver test includes the motor vehicle driver's on-site driving skills test (i.e., the subject two test). The qualified test site is the basis for the fairness and justice of the subject two test. However, the acceptance inspection of motor vehicle driver test sites in my country is mainly carried out by manual measurement and manual judgment. Manual on-site acceptance inspection has problems such as large inspection workload, high time cost, and poor accuracy. For example, it may take 2 to 4 days or even longer to inspect a single test site, and it will be more time-consuming and labor-intensive if review is required. Summary of the invention
[0003] In view of the above-mentioned problems, the present application is proposed in order to provide a method for automatic detection and modeling of motor vehicle driver examination sites that overcomes the above-mentioned problems or at least partially solves the above-mentioned problems, thereby saving the time and manpower spent on examination site detection, realizing the automation of examination site detection and acceptance, and ensuring the accuracy, objectivity and fairness of examination site acceptance.
[0004] According to one aspect of the present application, a method for automatic detection and modeling of a motor vehicle driver test site is provided, the method for automatic detection and modeling of a motor vehicle driver test site comprising:
[0005] Using a drone to collect image data and laser point cloud data of the motor vehicle driver test site to be tested, and transmitting the image data and the laser point cloud data to an image processing cloud platform; the drone is equipped with a camera and a laser radar LIDAR;
[0006] Using the image processing cloud platform to perform modeling based on the image data and the laser point cloud data, a three-dimensional model of the motor vehicle driver examination site is obtained, wherein the three-dimensional model includes quantitative feature information of the examination site;
[0007] The detection result is obtained according to the quantitative feature information of the examination room in the three-dimensional model and the corresponding quantitative feature information of the examination room in the pre-established three-dimensional model of the examination room acceptance standard.
[0008] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, the method described above is implemented.
[0009] The technical solution of the embodiment of the present application, by collecting image data and laser point cloud data of the motor vehicle driver test site to be tested, modeling is performed based on the image data and laser point cloud data, a three-dimensional model of the motor vehicle driver test site is obtained, and the test results are obtained by comparing the quantitative feature information of the test site in the three-dimensional model with the corresponding quantitative feature information of the test site in the three-dimensional model of the test site acceptance standard. On the one hand, compared with the traditional manual test site detection scheme, based on image modeling and feature comparison, the motor vehicle driver test site and its facilities are accurately detected, the test site detection is automated and digitized, and the problems of large detection workload and high time cost existing in manual on-site detection are effectively solved, and the detection efficiency is improved; on the other hand, the detection modeling is performed based on image data and laser point cloud data through the image processing platform, which ensures the accuracy, objectivity and fairness of the detection, and is convenient for management monitoring and later review. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:
[0011] Figure 1 A schematic flow chart of a method for automatic detection and modeling of a motor vehicle driver test site according to an embodiment of the present application is shown;
[0012] Figure 2 A schematic diagram showing the examination room image acquisition process according to an embodiment of the present application is shown;
[0013] Figure 3 A schematic diagram of the structure of a computer-readable storage medium according to an embodiment of the present application is shown. DETAILED DESCRIPTION
[0014] The exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.
[0015] At present, the acceptance inspection of motor vehicle driver examination sites in my country is mainly carried out by manual measurement. The inspection tasks include manual measurement of the graphics, markings and examination site facilities of the motor vehicle driver examination items, and manual comparison with the "Motor Vehicle Driver Examination Site and Its Facilities Setting Specification" (GA029-2017). The existing manual acceptance inspection method has problems such as large inspection workload and high time cost. The acceptance inspection of a single examination site may take 2 to 4 days or even longer, and it is more time-consuming and labor-intensive if review is required. In this regard, the embodiment of the present application proposes a method for automatic detection and modeling of motor vehicle driver examination sites, by collecting image data and laser point cloud data, constructing a three-dimensional model of the examination site according to the image and laser point cloud data, and matching the quantitative features in the three-dimensional model of the examination site with the quantitative features in the three-dimensional model of the standard examination site, and obtaining the inspection results, thereby saving the time and labor spent on the examination site inspection, making the examination site inspection acceptance digital and automated, and ensuring the precision and accuracy of the inspection, facilitating monitoring management and later review.
[0016] Figure 1 A flow chart of a method for automatic detection and modeling of a motor vehicle driver test site according to an embodiment of the present application, see Figure 1 The method of the embodiment of the present application comprises the following steps:
[0017] Step S101, using a drone to collect image data and laser point cloud data of the motor vehicle driver test site to be tested, and transmitting the image data and the laser point cloud data to an image processing cloud platform; the drone is equipped with a camera and a laser radar LIDAR. That is to say, the drone equipped with a special camera for surveying and mapping and a laser radar LIDAR (Light Detection and Ranging) is used to collect images and laser point clouds of the motor vehicle driver test site to be tested, and then store them, and then use 5G (5th-Generation) or 4G (4th-Generation) mobile communication methods to transmit the images.
[0018] Step S102, using the image processing cloud platform to model according to the image data and the laser point cloud data, to obtain a three-dimensional model of the motor vehicle driver examination site, wherein the three-dimensional model includes quantitative feature information of the examination site; the image processing cloud platform is a cloud platform that uses AI (Artificial Intelligence) algorithms, and its main function is to splice image data and laser point cloud data, and to identify targets. The three-dimensional model is, for example, a three-dimensional model of the examination site generated by the fusion of the 2D plane image and the laser point cloud data in the aforementioned steps. The quantitative feature information of the examination site refers to the features of the examination site that can be expressed in numerical values, such as the length and width of the garage of the examination site, and the width of the lanes of the examination site.
[0019] Step S103, obtaining a detection result according to the quantitative feature information of the examination room in the three-dimensional model of the stereoscopic model and the corresponding quantitative feature information of the examination room in the three-dimensional model of the pre-established examination room acceptance standard.
[0020] In this step, the test site to be tested is determined to be qualified or unqualified through feature comparison, and the test result is obtained. In practical applications, in order to improve and ensure the accuracy of the test results, the embodiment of the present application can also include implementing a part of the comparison function through the image processing platform, and then manually reviewing and confirming. The specific manual review and confirmation, for example, is performed in real time by the vehicle management office staff logging into the image processing cloud platform in the form of an account and password.
[0021] Depend on Figure 1 As shown, the method for automatic detection and modeling of the motor vehicle driver test site in the embodiment of the present application collects image data and laser point cloud data of the driver test site, models the site based on the image data and laser point cloud data, obtains a three-dimensional model of the test site, and determines whether the test site is qualified or unqualified based on the comparison results of the quantitative feature information of the test site in the three-dimensional model and the corresponding quantitative feature information of the test site in the standard three-dimensional model. Therefore, compared with the traditional manual on-site detection, the accurate detection of the motor vehicle driver test site and its facilities based on data modeling and model comparison solves the problems of large workload and high time cost of manual on-site detection, and ensures the accuracy of detection; moreover, it improves the automation, digitization level and detection efficiency of detection.
[0022] Figure 2 A schematic diagram of the examination room image acquisition process according to an embodiment of the present application is shown below. Figure 2 The method for automatic detection and modeling of a motor vehicle driver test site in an embodiment of the present application is described. Generally speaking, the method for automatic detection and modeling of a motor vehicle driver test site in an embodiment of the present application mainly includes three steps: first, data collection; second, data processing; and third, detection results. The following are described separately.
[0023] 1. Collect data.
[0024] See also Figure 2As shown on the left, the collected data here is the image data and laser point cloud data of the motor vehicle driver test site to be tested collected by a drone, which are stored synchronously on the drone, and the image data and lidar data (i.e., laser point cloud data) are synchronously transmitted back. Specifically, in the embodiment of the present application, a drone equipped with a special camera for surveying and mapping and a laser radar LIDAR (same as lidar) is used to collect images and laser point clouds of the motor vehicle driver test site to be tested. In order to improve the accuracy, the embodiment of the present application places a standard ruler in the aerial survey scene (i.e., the physical environment of the test site) as an accuracy standard to judge the accuracy of the aerial survey.
[0025] After acquiring the image data and laser point cloud data, the drone transmits the image data and the laser point cloud data to the image processing cloud platform in real time via the fifth generation mobile communication 5G or the fourth generation mobile communication 4G; or, the image data and the laser point cloud data are transmitted to the image processing cloud platform in an offline data manner. In other words, the drone is used to collect images of the examination room and laser point clouds, and image acquisition and laser point cloud data acquisition are performed on the graphics, markings, etc. of each examination item, and then the drone-collected images are transmitted back in real time and monitored on the ground in real time via the 5G / 4G network, thereby avoiding the problem of unclear and incomplete examination room image acquisition due to obstruction by objects such as trees, and improving the effectiveness and accuracy of image acquisition.
[0026] The embodiment of the present application utilizes the camera on the drone to capture images of the motor vehicle driver examination site to be tested and combines them with the 5G network to transmit images in real time, thereby solving the problem that existing handheld shooting equipment cannot restore the examination site and examination items in proportion, improving the level of automation of examination site detection, and ensuring the accuracy of detection.
[0027] In practical applications, using a drone to collect image data of the motor vehicle driver test site to be tested includes: using the first camera carried by the drone to vertically shoot the motor vehicle driver test site to be tested to obtain a bird's-eye view image, and using the second camera carried by the drone to tilt the height and slope information of the motor vehicle driver test site to be tested at a preset angle to obtain a side-view image. For example, the drone is equipped with five fixed-focus cameras for data collection, wherein one fixed-focus camera (i.e., the first camera) vertically shoots the details of the test site to obtain a bird's-eye view image, and four fixed-focus cameras (i.e., the second cameras) measure the height and slope of the test site at an angle of 35 degrees to obtain a side-view image, and stereoscopic image collection is achieved by fusing and splicing the synchronous images collected by the five cameras.
[0028] 2. Data processing.
[0029] See also Figure 2As shown in the middle, the image data collected in the previous step is modeled using an image processing cloud platform (mainly for automatic splicing and recognition) to obtain a three-dimensional model of the motor vehicle driver test site. The three-dimensional model of the test site includes: graphic marking information corresponding to the test items and test site facility information. In one embodiment, the designated test items include one or more of the following test items: reversing into the garage, side parking, ramp fixed-point parking and starting, right-angle turning, curve driving, pile test, passing a single-sided bridge, passing continuous obstacles, passing a width-limited gate, driving on an undulating road, turning around on a narrow road, simulating highway driving, and simulating continuous sharp bends in mountainous areas. The graphic marking information corresponding to each designated test item includes one or more of the following quantitative information: garage length, garage width, lane width, garage distance from the control line, and the distance between the starting line and the stop line and the outer edge of the garage; the test site facility information includes one or more of the following quantitative information: lane width, lateral clearance width, curb width, and curb clearance height.
[0030] That is to say, the motor vehicle driver test site inspection mainly tests the test items such as reversing into the garage, side parking, parking and starting at a fixed point on a slope, right-angle turning, curve driving, pole test, passing a single-sided bridge, passing continuous obstacles, passing a width-limited gate, driving on an undulating road, turning around on a narrow road, simulating driving on a highway, simulating driving on a mountain road with continuous sharp bends, etc. The test data includes the length of the garage, width of the garage, lane width, distance between the garage and the control line, and distance between the starting line and the stop line and the outer edge of the garage. In addition, the construction boundaries of the test area roads, including lane width, lateral clearance width, curb width, curb clearance height, etc.
[0031] It should be emphasized that an embodiment of applying the method of the present application to the detection of a motor vehicle driver examination site is described here. In other embodiments, more or fewer test items than the aforementioned description may be tested according to the standard specifications and requirements for the detection of a motor vehicle driver examination site. Similarly, for specific detection contents, the aforementioned description is in the form of examples, and those skilled in the art understand that the detection items and detection contents, such as the graphic marking information of the test items, should be determined according to the actual situation of the test items and are not limited to the aforementioned examples. That is, the aforementioned examples of detection contents and detection items do not represent all detection contents and all detection items.
[0032] In addition, these data are key quantitative feature data for examination room acceptance and affect the acceptance results. By collecting the quantitative feature data and reflecting them on the three-dimensional model of the examination room, subsequent data comparison is convenient.
[0033] It should be noted that the construction of the three-dimensional model of the examination room is the key to model comparison and examination room detection, so it will be emphasized here.
[0034] In one embodiment, the image processing cloud platform is used to perform modeling based on the image data and the laser point cloud data to obtain a three-dimensional model of the motor vehicle driver examination site, including: obtaining the position coordinates of the feature points on the image corresponding to the spatial points based on the three-dimensional position information of the spatial points of the motor vehicle driver examination site to be detected, as well as the camera intrinsic parameters, rotation matrix and translation vector; reconstructing a sparse point cloud based on the position coordinates of the feature points, and constructing a dense point cloud based on the sparse point cloud and reference points, wherein the reference points are determined by the laser point cloud data collected by a ground-based laser radar; constructing a point cloud mesh based on the dense point cloud, constructing a three-dimensional semantic model based on the point cloud mesh, and constructing a three-dimensional vector model based on the three-dimensional semantic model, and using the three-dimensional vector model as the three-dimensional model of the motor vehicle driver examination site.
[0035] That is, automatic detection modeling can be divided into the following three steps:
[0036] The first step is spatial point data collection . This includes two situations: image acquisition and ground-based laser radar acquisition. Image acquisition is carried out using a drone equipped with five fixed-focus cameras. Among them, one camera is directly facing the details of the examination room, and four cameras are tilted at 35 degrees to measure the height and slope of the examination room. Stereo image acquisition can be achieved by fusing and stitching the synchronized images acquired by these five cameras. Ground-based laser radar acquisition is aimed at the interference factors such as occlusion, reflection, shadow, occlusion, etc. that may be encountered in video image reconstruction, and the laser data of the ground-based laser radar LIDAR is used for fusion and completion. Accurate acquisition is achieved through the reflection information of geophysical points, and it is transmitted to the image processing platform through the mobile communication network. Note: The difference between ground-based laser radar and airborne laser radar is that the installation location is different. The airborne laser radar is installed on the drone, and the ground-based laser radar is set on the ground of the examination site to be tested. Both are used to collect laser point cloud data.
[0037] Step 2: Calculate image feature points The position coordinates of feature points on any image in the overhead or side-view image are calculated by combining the camera intrinsic parameters, camera rotation and translation data, and spatial point information of the camera on the drone through the following formulas (1) and (2):
[0038]
[0039]
[0040] Among them, formula (2) is obtained by transforming formula (1), and the matrix in formula (2) is It represents the position coordinates of the feature points on the image, K represents the camera intrinsic parameter, R represents the camera rotation matrix, and t represents the camera translation vector. Represents the position coordinates of a point in space.
[0041] Step 3: Complete 3D modeling and reconstruction. The image feature points calculated in the second step are reconstructed through sparse point cloud, dense point cloud, point cloud network modeling, 3D semantic modeling, and 3D vector modeling, and finally 3D reconstruction is achieved. It should be emphasized that in view of the interference factors such as occlusion, reflection, shadow, occlusion, etc. in the actual examination environment measurement, the embodiment of the present application introduces the laser point cloud data collected by the ground laser radar LIDAR as a reference point in the process of dense point cloud reconstruction, thereby greatly improving the accuracy of the overall model reconstruction. In practical applications, laser point cloud data collected by airborne laser radar can also be used according to different projects and geographical conditions.
[0042] After constructing the three-dimensional model of the examination room, the image processing platform of the embodiment of the present application performs target identification and handles problems such as image distortion and target occlusion to improve the precision and accuracy of the automatic detection and modeling method of the embodiment of the present application.
[0043] In one embodiment, the image processing cloud platform is used to perform modeling based on the image data and the laser point cloud data to obtain a three-dimensional model of the motor vehicle driver examination site, including: the image processing cloud platform identifies the target test items in the three-dimensional model of the motor vehicle driver examination site based on the signs set around the target test items, the signs including color signs and shape signs; or, the image processing cloud platform identifies right-angle turn test items and curve driving test items in the three-dimensional model of the motor vehicle driver examination site based on a curve curvature threshold; or, the image processing cloud platform identifies undulating roads and ramps in the three-dimensional model of the motor vehicle driver examination site based on a slope threshold.
[0044] Specifically, when performing target identification, the embodiments of the present application mainly adopt the following methods: first, setting marks around the test items, such as adding color marks and shape marks around test items that are difficult to distinguish, so that in the captured image, the target identification of the test items can be performed through these marks. The second is the recognition of curvature features, pre-setting the curvature curvature thresholds for right-angle turns and curved driving in the test items, and identifying whether it is a right-angle turn or curved driving based on the thresholds. The third is slope feature recognition. Pre-set the slope threshold to identify and archive undulating road surfaces, ramp angles, etc.
[0045] In order to reduce image distortion: the camera of the embodiment of the present application adopts a fixed-focus mapping lens, and sets the lens distortion rate parameter value in advance to ensure accurate restoration of the image at different shooting heights and angles.
[0046] In one embodiment, the automatic detection modeling of the motor vehicle driver examination site also includes: if it is determined based on the image data collected by the drone that the test item is partially blocked, adjusting the shooting height and angle of the drone; if it is determined based on the image data collected by the drone that the test item is completely blocked, collecting laser point cloud data through the ground laser radar of the motor vehicle driver examination site to be inspected, so that the image processing cloud platform splices the laser point cloud data with the image data collected by the drone.
[0047] That is to say, in order to deal with the technical problem of test items being blocked, when the test items are blocked by objects of a certain height and a small area, the embodiment of the present application solves the problem by controlling the drone to adjust the shooting height and angle; when the test items are completely blocked, the laser point cloud data is collected by the ground laser radar and then spliced with the image data collected by the drone. The position information of the ground laser radar is obtained through GPS (Global Positioning System). Note: Before actual use, the ground laser radar and the drone camera can be calibrated to complete the unification of the coordinate system.
[0048] In order to reduce errors, the method of an embodiment of the present application, after the image processing cloud platform performs modeling based on the image data and the laser point cloud data to obtain a three-dimensional model of the motor vehicle driver examination site, also includes: detecting the reconstruction data error values of all benchmark poles in the three-dimensional model of the motor vehicle driver examination site; taking the average of the reconstruction data error values to obtain a comprehensive error value of the image data; when the comprehensive error value is not greater than a preset comprehensive error threshold, retaining the three-dimensional model of the motor vehicle driver examination site; the number of the benchmark poles is determined based on the area and terrain of the motor vehicle driver examination site, and the benchmark poles are randomly placed within the range of the motor vehicle driver examination site.
[0049] For example, according to the area and terrain of the examination room, more than three benchmark poles are randomly placed in the examination room, and the reconstructed data error values of all benchmark poles in the three-dimensional modeling are detected. The comprehensive error value of the image can be obtained by taking the average. When the comprehensive error value is not greater than the preset comprehensive error threshold, the three-dimensional model of the motor vehicle driver examination site is retained, thereby improving the accuracy of image acquisition.
[0050] 3. Test results
[0051] In the embodiment of the present application, the test result is obtained according to the quantitative feature information of the examination room in the three-dimensional model and the corresponding quantitative feature information of the examination room in the pre-established three-dimensional model of the examination room acceptance standard. Specifically: the graphic marking information corresponding to each designated test item is matched with the graphic marking information corresponding to the corresponding test item in the three-dimensional model of the examination room acceptance standard. When the matching result is within the allowable error range, the test result of the qualified examination room is obtained; when the matching result exceeds the allowable error range, the test result of the unqualified examination room is obtained. In this way, the technical problem that the collected image cannot be automatically compared with the "Motor Vehicle Driver Examination Site and Its Facility Setting Specifications" is solved, and the beneficial effect of automatically detecting whether the test items and examination room facilities meet the requirements of the "Motor Vehicle Driver Examination Site and Its Facility Setting Specifications" is achieved. In addition, through the fusion of image data and laser point cloud data, the influence of distortion and occlusion is reduced, and the detection accuracy, objectivity and fairness are guaranteed.
[0052] It should be noted that due to the influence of factors such as image acquisition, image transmission and matching processing, there may be errors between the detection results of the embodiment of the present application and the actual examination room. Therefore, the embodiment of the present application sets the allowable error range in advance based on empirical values. If the graphic markings of the examination room to be tested are matched with the image markings of the standard examination room, and the difference between the two is within the error range, then the two are considered to be matched or the markings of the examination room are qualified.
[0053] See also Figure 2 As shown on the right, after the model is built, the embodiment of the present application can view and plot the ground stitching results locally. That is, after the partial comparison function is realized through the image processing platform, it is manually reviewed and confirmed. For example, the vehicle management office staff logs in to the image processing cloud platform with an account and password to view the comparison results, and conducts review and confirmation.
[0054] Furthermore, the method of the embodiment of the present application also includes: when the test result indicates that the test room is unqualified, marking the test room facilities and / or test item graphic markings that do not meet the requirements and generating a test report. For example, when the measurement error is within 5%, it is considered to be qualified. If it exceeds the error range, the embodiment of the present application can automatically identify the white line and automatically measure and plot the relevant data.
[0055] That is to say, if it is determined through comparison that the examination room facilities do not meet the requirements, then the examination room facilities will be marked; if it is determined through comparison that the test item graphic markings do not meet the requirements, then the test item graphic markings will be marked; if it is determined through comparison that both the examination room facilities and the test item graphic markings do not meet the requirements, then both the examination room facilities and the test item graphic markings will be marked, so that the unqualified information is clear at a glance, which is convenient for subsequent targeted rectification.
[0056] The test site facilities here include two-way two-lane roads, observation points, energy-dissipating objects or facilities set up within the lateral clearance range of the road, etc. The test item graphics and markings include test item graphics and test item markings. The test item graphics and markings vary according to the test items. For example, the graphics and markings of the reverse parking test item are obviously different from those of the curve driving test item.
[0057] Thus, the embodiment of the present application provides a method for automatic detection and modeling of a motor vehicle driver examination site. By comparing the system program with the examination site acceptance specifications, it finds out the signs, markings, and examination site facilities that do not meet the specifications and makes corresponding annotations (for example, highlighting them in yellow), thereby saving the time for troubleshooting unqualified examination site information and improving the accuracy of detection.
[0058] In summary, the method for automatic detection and modeling of a motor vehicle driver test site in the embodiment of the present application collects images and laser point cloud data of the test site to be detected, models the test site based on the images and laser point cloud data to obtain a three-dimensional model of the test site, and compares the quantitative feature information of the test site with the corresponding quantitative feature information of the test site in the three-dimensional model of the test site acceptance standard to obtain the detection result. Thus, on the one hand, it solves the problems of large workload and high time cost of manual on-site detection, and ensures the accuracy of detection; on the other hand, it improves the automation, digitization level and efficiency of test site detection, and facilitates management and monitoring.
[0059] It should be noted that:
[0060] The algorithm and display provided herein are not inherently related to any particular computer, virtual device or other equipment. Various general purpose devices can also be used together with the teachings based on this. According to the above description, it is obvious to construct the structure required for this type of device. In addition, the application is not directed to any specific programming language either. It should be understood that various programming languages can be utilized to realize the content of the application described herein, and the description of the specific language above is for the purpose of disclosing the best mode of the application.
[0061] In the description provided herein, a large number of specific details are described. However, it is understood that the embodiments of the present application can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.
[0062] Similarly, it should be understood that in order to streamline the present application and help understand one or more of the various inventive aspects, in the above description of the exemplary embodiments of the present application, the various features of the present application are sometimes grouped together into a single embodiment, figure, or description thereof. However, the disclosed method should not be interpreted as reflecting the following intention: the claimed application requires more features than the features clearly stated in each claim. More specifically, as reflected in the claims below, the inventive aspects are less than all the features of the single embodiment disclosed above. Therefore, the claims following the specific embodiment are hereby expressly incorporated into the specific embodiment, wherein each claim itself serves as a separate embodiment of the present application.
[0063] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition they may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.
[0064] In addition, those skilled in the art will appreciate that, although some embodiments described herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present application and form different embodiments. For example, in the claims below, any one of the claimed embodiments may be used in any combination.
[0065] The various component embodiments of the present application can be implemented in hardware, or implemented in software modules running on one or more processors, or implemented in a combination thereof. It should be understood by those skilled in the art that a microprocessor or a digital signal processor (Digital Signal Processor, DSP) can be used in practice to implement some or all functions of some or all components according to the embodiments of the present application. The application can also be implemented as a device or apparatus program (for example, a computer program and a computer program product) for executing a part or all of the methods described herein. Such a program implementing the present application can be stored on a computer-readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.
[0066] Figure 3 A schematic diagram of the structure of a computer-readable storage medium according to an embodiment of the present application is shown. The computer-readable storage medium 300 stores a computer-readable program code for executing the method according to the present application, which can be read by a processor of an electronic device. When the computer-readable program code is executed by the electronic device, the electronic device executes each step of the method described above. Specifically, the computer-readable program code stored in the computer-readable storage medium can execute the method shown in the above embodiment. The computer-readable program code can be compressed in an appropriate form.
[0067] It should be noted that the above embodiments illustrate the present application rather than limit the present application, and that those skilled in the art may design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets should not be constructed as a limitation to the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of multiple such elements. The present application may be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In a unit claim that lists several devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc. does not indicate any order. These words may be interpreted as names.
[0068] The above is only a specific embodiment of the present invention. Under the above teachings of the present invention, those skilled in the art can make other improvements or modifications based on the above embodiments. Those skilled in the art should understand that the above specific description is only to better explain the purpose of the present invention, and the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims
1. A method for automatic detection and modeling of a motor vehicle driver test site, characterized in that: The method for automatic detection and modeling of a motor vehicle driver test site includes: Using a drone to collect image data and laser point cloud data of the motor vehicle driver test site to be tested, and transmitting the image data and the laser point cloud data to an image processing cloud platform; the drone is equipped with a camera and a laser radar LIDAR; Using the image processing cloud platform to perform modeling based on the image data and the laser point cloud data, a three-dimensional model of the motor vehicle driver examination site is obtained, wherein the three-dimensional model includes quantitative feature information of the examination site; Obtaining a test result based on the test room quantitative feature information in the three-dimensional model and the corresponding test room quantitative feature information in the pre-established test room acceptance standard three-dimensional model; The test room quantitative feature information includes graphic marking information corresponding to each designated test item; According to the quantitative feature information of the examination room in the three-dimensional model and the corresponding quantitative feature information of the examination room in the pre-established three-dimensional model of the examination room acceptance standard, the detection results include: Match the graphic marking information corresponding to each designated test item with the graphic marking information corresponding to the corresponding test item in the three-dimensional model of the test site acceptance standard. When the matching result is within the allowable error range, the test result of the examination room passing is obtained; When the matching result exceeds the allowable error range, the test result of the examination room being unqualified is obtained; The examination room quantitative characteristic information also includes examination room facility information; The designated examination items include one or more of the following examination items: Reversing into a garage, parallel parking, parking and starting at a fixed point on a ramp, right-angle turns, curve driving, pole tests, passing a single-sided bridge, passing continuous obstacles, passing a width-limited gate, driving on undulating roads, U-turns on narrow roads, simulated highway driving, and simulated driving on mountain roads with continuous sharp bends; The graphic marking information corresponding to each designated test item includes one or more of the following quantitative information: garage length, garage width, lane width, distance between the garage and the control line, and the distance between the starting line and the stop line and the outer edge of the garage; The test site facility information includes one or more of the following quantitative information: lane width, lateral clearance width, curb width, curb clearance height; Using the image processing cloud platform to perform modeling based on the image data and the laser point cloud data to obtain a three-dimensional model of the motor vehicle driver test site includes: The position coordinates of the feature points on the image corresponding to the spatial points are calculated based on the three-dimensional position information of the spatial points of the motor vehicle driver test site to be tested, as well as the camera internal parameters, the rotation matrix and the translation vector; Reconstructing a sparse point cloud according to the position coordinates of the feature points, and constructing a dense point cloud according to the sparse point cloud and reference points, wherein the reference points are determined by laser point cloud data collected by a ground laser radar; Constructing a point cloud mesh according to the dense point cloud, constructing a three-dimensional semantic model according to the point cloud mesh, constructing a three-dimensional vector model according to the three-dimensional semantic model, and using the three-dimensional vector model as the three-dimensional model of the motor vehicle driver test site; Using the image processing cloud platform to perform modeling based on the image data and the laser point cloud data to obtain a three-dimensional model of the motor vehicle driver test site includes: The image processing cloud platform identifies the target test item in the three-dimensional model of the motor vehicle driver test site according to the identification set around the target test item, wherein the identification includes a color identification and a shape identification; Alternatively, the image processing cloud platform identifies the right-angle turn test item and the curve driving test item in the three-dimensional model of the motor vehicle driver test site according to the bend curvature threshold; Alternatively, the image processing cloud platform identifies the undulating road surface and ramps in the three-dimensional model of the motor vehicle driver test site according to the slope threshold; After the image processing cloud platform is used to perform modeling based on the image data and the laser point cloud data to obtain a three-dimensional model of the motor vehicle driver test site, the method further includes: Detecting reconstruction data error values of all benchmark poles in the three-dimensional model of the motor vehicle driver test site; averaging the reconstructed data error values to obtain a comprehensive error value of the image data; When the comprehensive error value is not greater than a preset comprehensive error threshold, retaining the three-dimensional model of the motor vehicle driver test site; The number of the benchmark poles is determined according to the area and topography of the motor vehicle driver examination site, and the benchmark poles are randomly placed within the motor vehicle driver examination site.
2. The method according to claim 1, characterized in that The method further includes: When the test results indicate that the examination room is unqualified, the examination room facilities and / or examination item graphic markings that do not meet the requirements shall be marked and a test report shall be generated.
3. The method according to claim 1, characterized in that The transmitting the image data and the laser point cloud data to the image processing cloud platform comprises: The image data and the laser point cloud data are transmitted to the image processing cloud platform in real time via the fifth generation mobile communication 5G or the fourth generation mobile communication 4G; or, the image data and the laser point cloud data are transmitted to the image processing cloud platform in an offline data manner; The image data collected by drones at the motor vehicle driver test site to be tested includes: The first camera carried on the drone is used to vertically shoot an overhead shot of the motor vehicle driver examination site to be tested to obtain an overhead shot image, and the second camera carried on the drone is used to tilt at a preset angle to shoot sideways the height and slope information of the motor vehicle driver examination site to be tested to obtain a side shot image.
4. The method according to any one of claims 1 to 3, characterized in that The method further comprises: If it is determined based on the image data collected by the drone that the test item is partially blocked, adjusting the shooting height and angle of the drone; If it is determined based on the image data collected by the drone that the test item is completely blocked, the laser point cloud data is collected by the ground laser radar at the motor vehicle driver test site to be tested, so that the image processing cloud platform splices the laser point cloud data with the image data collected by the drone.
5. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, the method according to any one of claims 1 to 4 is implemented.
