Method and system for the camera-based determining of the speed of a vehicle in traffic

Abstract

The method (10) for the camera-based determining of the speed of a vehicle (60) comprises the following method steps: creating a first, second, third and fourth image capture; identifying features (61) in the individual image captures; checking for consistency of the individual features (61); discarding the inconsistent features (61), determining the spatial position of the features (61) using a triangulation method; determining a speed vector for each feature (61); determining the speed for each feature (61) from the speed vector determined for the corresponding feature (61) and the difference between the first time and the second time; and determining an average speed of the vehicle (60) by averaging the determined speeds for each feature (61).

Description

[0001] Method and system for camera-based determination of the speed of a vehicle in road traffic

[0002] The present invention relates to a method for camera-based determination of the speed of a vehicle in road traffic and a corresponding system for camera-based speed determination.

[0003] Numerous methods for determining speed are known from the state of the art and contribute to monitoring road traffic and increasing road safety.

[0004] Over the past ten years, the number of accidents involving personal injury attributable to exceeding the local speed limit has decreased by approximately 20%. Nevertheless, excessive speeding is still considered the cause of one in three traffic fatalities in Germany. Monitoring vehicle speeds in road traffic therefore makes an important contribution to reducing the number of traffic accidents and fatalities by raising drivers' awareness of the need to comply with the speed limit.

[0005] To date, various technical solutions exist for measuring the speed of vehicles in road traffic. The methods known so far can be divided into active and passive methods.

[0006] Active methods typically include those methods in which the measuring device emits electromagnetic radiation in the near-optical range (THz, lasers, and lidar) or in the radio range (GHz, radar, and microwaves). The subsequent measurement of the reflection of the radiation emitted by the measuring device allows the speed of a vehicle to be determined. The Doppler effect (phase shift of a single pulse) or the change in the signal propagation time of several consecutive pulses can be used to determine the vehicle's speed. There are also laser-based methods, for example, which allow the vehicle speed to be determined by evaluating geometric relationships.

[0007] The active measurement methods discussed above can certainly deliver precise results. However, these active measurement methods have the disadvantage that they typically require high energy due to the required radiation emission. Consequently, it is often difficult to power active speed measurement devices with batteries, accumulators, and / or solar cells. Furthermore, measurement systems for active vehicle speed measurement are usually associated with high acquisition costs.

[0008] In addition to the active measurement methods mentioned above, passive measurement methods are also used for speed measurement. Unlike active measurement methods, passive methods do not emit radiation. In contrast, passive methods measure the time a vehicle takes to travel a known distance. Such methods are also known as distance-time methods. Sensors used can be light barriers, brightness sensors, pressure sensors (also known as overrun sensors), or induction sensors, which generate a signal when a vehicle passes. Some of these methods require structural measures, which significantly increase the costs involved.

[0009] A disadvantage of the described methods is that all of them require a separate camera to record the vehicle data.

[0010] Due to the above disadvantages, the object of the present invention is to provide a simple, cost-effective, and precise method for determining the speed of a vehicle in road traffic. To achieve this object, the present invention proposes a method for camera-based determination of the speed of a vehicle in road traffic, which method comprises the following method steps:

[0011] Creating a first image at a first time using a first image sensor and creating a second image at the first time using a second image sensor;

[0012] Creating a third image at a second time using the first image sensor and creating a fourth image at the second time using the second image sensor;

[0013] Recognition of features in individual images;

[0014] comparing the features in the images recorded by the first image sensor with the features contained in the images recorded by the second image sensor at the same time, whereby those features which are contained only in the images recorded by the first image sensor or in the images recorded by the second image sensor are discarded;

[0015] Comparing the corresponding features in the images recorded by the first image sensor with the features in the images recorded by the second image sensor with regard to their epipolar geometry, whereby those features for which the epipolar conditions are not met are discarded;

[0016] Comparing the features in the images captured by the first image sensor, discarding those features that are not included in both images captured by the first image sensor and those features that are not included in both images captured by the second image sensor; determining features whose position has remained unchanged from the first time to the second time and removing those features whose position has remained unchanged;

[0017] Determining the spatial position of the features using a triangulation method;

[0018] Determination of a velocity vector for each feature that was not previously rejected;

[0019] Determining the velocity for each feature from the velocity vector determined for the corresponding feature;

[0020] Determination of an average speed of the vehicle by averaging the determined speeds for each feature.

[0021] The advantage of camera-based speed measurement is that little or no structural work is required and the method can be implemented cost-effectively. The use of two image sensors, whose sensor surfaces are particularly inclined relative to each other, makes it possible to determine the position of a vehicle or individual detected features in three dimensions, thus recording the vehicle's speed in three dimensions. The two image sensors can be arranged next to each other (as a left and right image sensor) or one above the other (as an upper and lower image sensor).

[0022] The two image sensors each create an image at a first point in time and a second point in time. However, the invention is not limited to image recording at two points in time. Rather, the image sensors can also be designed to create numerous image recordings, so that the information from a large number of image recordings can be used to determine the speed. In this respect, the two points in time mentioned are to be understood as a minimum within the scope of the present invention, so that the individual image sensors create at least two image recordings. The first image sensor and the second image sensor can together represent a camera system. The image sensors can be arranged in a housing. In addition, the camera system can have recording optics, each of which is arranged between an image sensor and the recording area to be monitored (in which the vehicles are to be recorded).

[0023] Specific features are detected in each image generated by the image sensors. One of numerous state-of-the-art feature recognition algorithms can be used for this purpose. For example, the SIFT (Scale Invariant Feature Transform), SURF (Speed ​​Up Robust Feature), BRIEF (Binary Robust Independent Elementary Feature), or ORB (Oriented FAST and Rotated BRIEF) methods can be used to detect features. A Harris Corner Detector can also be used for feature detection to detect prominent corner points within an image.

[0024] The features can be a prominent point within the image. For example, a corner point within an image that exhibits a particularly high contrast can be detected. Alternatively, the feature can also refer to a line or another geometric shape (e.g., a rectangle or a trapezoid). In practice, several tens, several hundred, several thousand, or even more features can be detected within an image or belong to the same vehicle.Since processing a large number of features increases the computational effort and consequently the required computing time, on the one hand, and reduces inaccuracy due to expected errors in the detection of individual features in the image recordings, the present invention performs a specific selection of the detected features before the vehicle speed is ultimately determined based on the detected features. This significantly increases the efficiency of the measurement method and reduces the method's susceptibility to errors.

[0025] During selection, the features recorded by the first image sensor are compared with the features recorded by the second image sensor. If a feature, for example a prominent point, was detected at a first time t1 in the first image recording and this feature does not appear in the second image recording (also recorded at time t1), this feature is discarded for further processing. After this process step, only those features remain that are actually contained in the images recorded by both image sensors at the same time. In this way, an initial consistency check is performed, with all inconsistent features being disregarded for the speed calculation.

[0026] A known feature in an image can, for example, be described by a vector. Ideally, two corresponding features in two images have identical values. Consequently, two features are considered identical if the feature vectors are identical. In practice, however, the comparison of two features can be achieved not only by determining the same feature based on the identity of two feature vectors, but also by first calculating a distance between two feature vectors. The identity of the two features is then assumed if the distance between the feature vectors is smaller than a predetermined threshold. The Hamming distance or the Euclidean distance, for example, can be used to calculate the distance between two feature vectors.

[0027] In addition, the features in the images captured by the first image sensor and the second image sensor at the same time are compared with each other with regard to their epipolar geometry. If the epipolar condition is met for two features, it is assumed that these points are consistent. In this case, the features are retained. If the epipolar condition is not met for two features, this is an indication that the features are inconsistent; the inconsistent features are removed and neglected in the subsequent velocity determination. In general, the epipolar condition is met for two points if the following equation (also known as the epipolar equation) is met: p FPi = 0 or

[0028] Here, pi denotes a point in the first image, p2 a point in the second image, and F the fundamental matrix. The fundamental matrix is ​​calculated from the geometric relationship between the two cameras (translation and rotation) and the intrinsic camera parameters. It describes a mapping rule between the coordinate systems of both cameras. If the features represent individual points, it can be directly determined by checking the above condition whether the two features correspond. If the features represent geometric shapes, for example, one or more points of the geometric shape of the respective image recordings can be compared with each other with regard to their epipolar geometry.For example, in the case of a line, the epipolar condition test can be applied to the two endpoints of the line to check whether the lines detected in two image acquisitions describe the same feature.

[0029] In practice, the fulfillment of the epipolar condition can be defined such that the product of p2(transposed), the fundamental matrix F and pi is not exactly equal to 0, but is smaller than a predetermined limit. Furthermore, the features determined in two consecutive images captured by an image sensor are checked for consistency. In this process, those features that are not contained in both images captured by the first image sensor or by the second image sensor are discarded. As already explained above, an image sensor can be designed to generate more than two images. In this case, it can be provided that those features that are not contained in all images generated by an image sensor are discarded.

[0030] To further increase the efficiency of the method according to the invention, the present invention also determines the features whose positions have remained unchanged over time, discarding these "constant" features. This further reduces the total number of features used for speed determination to those features that actually move between the first and second points in time and therefore contain information about the vehicle's speed. This advantageously prevents unnecessary consideration of features that can be attributed, for example, to a building or a road marking.

[0031] After discarding those features that are considered inconsistent, the spatial position of the features is determined using a triangulation method. Triangulation methods are well known in the art and allow the position of an object to be determined using trigonometric relationships.

[0032] Once the position of a feature at the first and second time points has been determined, a velocity vector is determined for each feature. The velocity for each of the (consistent) features can then be determined from the time difference between two consecutive images and the displacement vector (also known as the motion vector) for each (consistent) feature. The time difference between two image recordings can generally be assumed to be known. This can be derived, for example, from the image recording frequency of the image sensors used. If the image sensors are used with an image recording frequency of 20 fps (frames per second), it can be deduced that the time difference between two consecutive image recordings is 50 ms.If two images from a video sequence are used to determine the speed, which are ten individual images apart, it can be concluded that the time difference between the images used is 500 ms. In this respect, a speed vector can be determined as the quotient of a displacement vector, which describes the movement of a feature between two images and is defined by the start position and the end position of the feature, and the time difference between the two images. According to the present invention, the speed vector can therefore be determined for the respective feature by determining the quotient of a displacement vector, which describes the movement of a feature between two images and is defined by the start position and the end position of the respective feature, and the time difference between two images.The velocity can be determined for one feature at a time by determining the magnitude of the velocity vector for the corresponding feature.

[0033] An average speed of the vehicle can then be determined by calculating an average value from the individual speeds determined for each feature. For example, an arithmetic mean can be formed from the individual speed values ​​or the median value can be determined from the individual speed values. In this case, only the features or the speed values ​​for those features that were previously considered consistent and therefore not rejected are taken into account. Even if the present invention has been described above in connection with a specific sequence, it will be obvious to a person skilled in the art that the order of the method steps is not mandatory. For example, the order of the individual consistency checks relating to the individual features can be swapped without deviating from the present invention.

[0034] According to one embodiment of the invention, the method may additionally comprise the following method step:

[0035] Checking the consistency of the velocity vectors using a RANSAC procedure, discarding those velocity vectors that were evaluated as inconsistent.

[0036] RANSAC (Random Sample Consensus) methods are generally known from the state of the art. The use of a RANSAC method allows the detection of outliers in a measurement data set. Taking into account the mean of the recorded measurement data and a specified tolerance range, those measured values ​​that lie outside the specified tolerance range are identified.

[0037] As an alternative to the RANSAC method, a special variant of the DBScan method can also be used according to the invention. In this case, the determined 3D points and velocity vectors are simultaneously "clustered" in a 6-dimensional space, i.e., consistent groups are formed with respect to a 6-dimensional metric. Each consistent group corresponds to a vehicle moving within the visible field of view.

[0038] A consistent group that is visible across multiple consecutive images, but at least across two consecutive images, is referred to as a "cluster." The velocity measurement value for a cluster is calculated as soon as its geometric center of gravity falls below a defined distance from the measuring device. The velocity values ​​of all points belonging to all consistent groups of the cluster are taken into account.

[0039] According to the invention, the consistent groups can be combined into a cluster by determining the geometric center of gravity for each consistent group and checking whether this center of gravity has moved by the distance corresponding to the velocity vector of the respective consistent group within the corresponding time difference between the frames.

[0040] According to one embodiment of the method according to the invention, a time signature (often referred to as a time stamp, digital time stamp, or in English, a time stamp) is generated during the creation of each image recording, which is attached to or assigned to the image recording. The additional time signature allows the exact time at which a recording was actually taken to be precisely defined. In this way, the accuracy of the method according to the invention can be increased.

[0041] According to an embodiment of the method according to the invention, it can also be provided that the method comprises the following method step:

[0042] Checking the consistency of the direction of the individual speed vectors, whereby the angle between the individual speed vectors and a predetermined reference line is determined and an average angle is determined from the determined angle values. Those speed vectors are regarded as inconsistent for which the difference between the determined angle and the average angle is greater than a predetermined limit, and those speed vectors that are regarded as inconsistent are rejected. The reference line can, in particular, run parallel to the monitored roadway. In general, it can be expected that the determined characteristics of the vehicle indicate a direction of movement that is essentially parallel to the roadway.If the direction of movement deviates significantly from the reference line (and consequently the determined angle value exceeds a specified limit), this can be interpreted as an indicator that a determined velocity vector is inconsistent. Using this approach, individual velocity vectors that are considered inconsistent can be discarded. By subsequently limiting the calculation to the consistent velocity vectors, the accuracy of the method can be further increased. At the same time, the error susceptibility of the measurement method is reduced.

[0043] Furthermore, according to the present invention, it can be provided that the method according to the invention comprises the following method step:

[0044] Checking the consistency of the magnitude of the individual velocity vectors using a RANSAC method, discarding those velocity vectors that are considered inconsistent.

[0045] While the consistency of the direction of the velocity vectors was checked in the previously described method step, the consistency of the magnitude is checked in this method step. Outliers determined by the RANSAC method used are ignored in the subsequent calculation of the vehicle speed. A limit value can be defined, with those measured values ​​that exceed the limit value being considered outliers. By checking the consistency of the vector magnitudes and discarding inconsistent values, the efficiency of the present measurement method can be further increased while simultaneously reducing the susceptibility to errors. According to a preferred embodiment of the invention, the method according to the invention can comprise the following method step:

[0046] Comparison of the mutually corresponding features in the image recordings taken by the first image sensor at the first time and at the second time and comparison of the mutually corresponding features in the image recordings taken by the second image sensor at the first time and at the second time with regard to their epipolar geometry, wherein those features for which the epipolar condition is not fulfilled are discarded.

[0047] This allows the number of features used for the final speed determination to be further reduced, with only consistent features being considered for the speed determination. This further increases the efficiency of the method according to the invention and simultaneously improves the precision of the method, since significantly fewer measured values ​​need to be used for the final speed determination and inconsistent measured values ​​are disregarded in the speed determination.

[0048] Furthermore, to achieve the object described above, a system for determining the speed of a vehicle in road traffic is proposed, the system comprising: a first image sensor for creating a first image at a first time and a third image at a second time; a second image sensor for creating a second image at the first time and a fourth image at the second time; and an evaluation unit, the external unit being configured to recognize features in the individual image recordings; to compare the features in the image recordings recorded by the first image sensor with the features contained in the image recordings recorded by the second image sensor, and to reject those features that are only contained in the image recordings recorded by the first image sensor or in the image recordings recorded by the second image sensor;to compare the corresponding features in the images recorded by the first image sensor with the features in the images recorded by the second image sensor with regard to their epipolar geometry and to reject those features for which the epipolar condition is not met; to compare the features in the images recorded by the first image sensor with each other and to reject those features that are not contained in both images recorded by the first image sensor and those features that are not contained in both images recorded by the second image sensor; to determine the features whose position has remained unchanged from the first time to the second time and to remove those features whose position has remained unchanged; to determine the spatial position of the individual features using a triangulation method;to determine a velocity vector for each feature; the velocity for each feature from the velocity vector determined for the corresponding feature; and to determine an average velocity of the vehicle by averaging the determined velocities for each feature.

[0049] The first image sensor and the second image sensor can be designed, for example, as CCD sensors or CMOS sensors. The sensors can be inclined relative to one another and arranged in a common sensor housing. In addition, recording optics can be provided, which are arranged between the respective image sensor and the object to be recorded. The evaluation unit can be designed, for example, as an integrated computing unit with a CPU or a GPU, an ASIC, a microcontroller, or an FPGA. In addition, the system according to the invention can have a storage unit for storing the recorded images. Alternatively, it can be provided that an external cloud storage is used on which the recorded images are stored. In this case, the invention measuring system can have a communication unit for exchanging information with the external cloud storage.

[0050] In the system according to the invention, the evaluation unit can additionally be configured to check the consistency of the velocity vectors using a RANSAC method, whereby those velocity vectors that are considered inconsistent are discarded.

[0051] The system according to the invention can also comprise a timer which is configured to generate a time signature at the time of an image recording, which is attached to or assigned to the respective image recording.

[0052] Furthermore, the system according to the invention can be provided for the evaluation unit to be configured to check the consistency of the direction of the individual velocity vectors, wherein the angle between the individual velocity vectors and a predetermined reference line is determined in each case and an angle mean is determined from the determined angle values, and those velocity vectors are regarded as inconsistent for which the difference between the determined angle value and the angle mean is greater than a predetermined limit value, and wherein those velocity vectors that are regarded as inconsistent are discarded. Furthermore, it can be provided for the evaluation unit to be configured to check the consistency of the magnitude of the individual velocity vectors using a RANSAC method, wherein those velocity vectors that are regarded as inconsistent are discarded.

[0053] Furthermore, the evaluation unit can additionally be configured to compare the mutually corresponding features in the image recordings taken by the first image sensor at the first time and at the second time with regard to their epipolar geometry and to compare the mutually corresponding features in the image recordings taken by the second image sensor at the first time and at the second time with regard to their epipolar geometry, wherein those features for which the epipolar condition is not fulfilled are discarded.

[0054] The present invention is explained in more detail below with reference to the figures. These show:

[0055] Fig. 1 shows an embodiment of the method according to the invention,

[0056] Fig. 2 shows an embodiment of the system according to the invention,

[0057] Fig. 3 a schematic representation of speed measurement in road traffic,

[0058] Fig. 4 shows a first image recording and a second image recording which were generated at a first time, and

[0059] Fig. 5 shows a third and a fourth image recording that were generated at a second point in time. Fig. 1 schematically shows an embodiment of the method 10 according to the invention. In the embodiment shown, the method 10 according to the invention comprises a total of eleven method steps 11-21. In the first method step 11, a first image recording and a second image recording are created using a first image sensor and a second image sensor at a first point in time. In the second method step 12, a third image recording and a fourth image recording are created at a second point in time. Subsequently, in the third method step 13, features are recognized in the individual image recordings. For this purpose, one of the methods for feature recognition described above (for example a Harris Corner Detector) can be used.However, the present invention is not limited to the use of a single feature recognition method. In the fourth method step 14, the features detected in the images recorded by the first image sensor are compared with the features detected in the images recorded by the second image sensor. In particular, the features in the first image can be compared with the features in the second image, and the features in the third image can be compared with the features in the fourth image. In other words, in particular, the features in the images recorded by the first image sensor and the second image sensor at the same time can be compared. In this case, those features that are only contained in the image recorded by a single image sensor are discarded for later speed determination.These features are regarded as inconsistent features, so that discarding these features contributes to increasing the precision in calculating the vehicle speed. In the fifth method step 15, the corresponding features (i.e. those features that were detected by both the first image sensor and the second image sensor) are compared with regard to their epipolar geometry, with those features for which the epipolar condition is not met being discarded. These features are also regarded as inconsistent, so that discarding these features also contributes to increasing the precision in calculating the vehicle speed. In the sixth method step 16, the features that were each recorded by an image sensor at different times are compared with each other. All features that are only contained in one image recording are regarded as inconsistent and discarded.In the seventh method step 17, those features are determined whose position remained unchanged at the first and second time points. These features, which are also considered constant features, refer to objects that are stationary in the recorded image sequence (for example, they could be the features of a building), making these features obsolete for the speed calculation. The constant features are therefore also discarded and are not taken into account in the subsequent speed calculation. By discarding the inconsistent features, the accuracy of the calculation and, at the same time, the efficiency of the method according to the invention are increased. In the eighth method step 18, the position of the features is determined using a triangulation method. Subsequently, in the ninth method step 19, a speed sector is determined for each detected (and not discarded) feature.In the tenth method step 20, the speed for each of the detected features (if not previously rejected) is determined from the speed vector determined for the corresponding feature. Finally, in the eleventh method step 21, the average speed of the vehicle is determined by communicating the previously determined speeds for each relevant (i.e., consistent) feature.

[0060] Even if the method steps described above have been described in a predetermined order for illustrative reasons, it will be apparent to those skilled in the art that adherence to the order is not binding and that, for example, the order of the individual steps for discarding inconsistent features can be changed without deviating from the principle of the present invention. Fig. 2 shows an exemplary embodiment of the system 50 according to the invention. The system 50 has a first image sensor 52 and a second image sensor 54. The image sensors 52, 54 can in particular be designed as CCD image sensors or as CMOS image sensors. The image sensors 52, 54 can each be designed to create an image sequence from at least two image recordings.The image sensors can preferably be configured to record a video sequence that can be used to calculate individual vehicle speeds. Furthermore, the system 50 has an evaluation unit 56, which is configured to perform the method steps described in connection with Fig. 1.

[0061] Fig. 3 schematically illustrates the principle of speed measurement using the system 50 according to the invention. As shown in this figure, the system 50 according to the invention can, for example, be mounted on a roadside. The system 50 is oriented such that it can monitor a detection zone 58. All vehicles 60 that pass through the detection zone 58 can be monitored by the system 50, wherein, in particular, the speed of the vehicle 60 can be determined using the method described above. If the system 50 according to the invention has detected the speed of the vehicle 60, it can compare the determined speed with the maximum speed permitted in the respective section of road. If the maximum speed permitted is exceeded by the vehicle 60, at least one image of the vehicle 60 is saved and forwarded to the responsible authority.For example, a single image of the vehicle 60 can be forwarded to the responsible authority to identify the vehicle owner, together with the determined speed. Alternatively, the recorded image sequences can be saved and forwarded so that the accuracy of the speed calculation can be verified at a later time, should the vehicle owner question the speeding. Fig. 4 shows a first image 62 and a second image 64, each generated at a first time t1. The first image 62 was created by a first image sensor, while the second image 64 was created by a second image sensor. The two image sensors record the corresponding images from different perspectives. A vehicle 60 can be seen in the first image 62.A feature 61 was detected within the image 62. This feature 61 is a corner point of the vehicle license plate. One of the methods described above can be used for feature detection. For example, a Harris Corner Detector can be used for this. The feature 61 was also detected in the second image 64. The feature 61 can, for example, be described by a feature vector. By comparing the feature vectors of the features detected in the first image 62 and the second image 64, it can be checked whether the same feature was detected in the two image 62, 64. If this is the case, the two features 61 are retained. If the features 61 are only contained in one of the image 62, 64, the features are classified as inconsistent and consequently discarded.

[0062] Using trigonometric relationships, the position of feature 61 can be determined at the first time point. Even though Fig. 4 shows the detection of a single feature, in practice a large number of features are detected in one image recording, with a consistency check then being performed for the individual features and all features deemed inconsistent being discarded for the subsequent speed calculation. For example, several hundred or several thousand characteristic points can be detected in one image recording or on a vehicle, with a large proportion of these points or features being discarded for the subsequent calculation in order to obtain a selection of features deemed relevant for the speed calculation. Fig. 5 shows a third image recording 66 and a fourth image recording 68, which were recorded at a second time point t2.The third image 66 was taken by the first image sensor, while the fourth image 68 was taken by the second image sensor. The images 66, 68 depict the same vehicle 60 that was previously visible in Fig. 4. As shown in Fig. 5, the vehicle 60 has covered a certain distance compared to the first time t1. Consequently, the position of the feature 61 has changed from Fig. 4 to Fig. 5. This allows a speed vector for the feature 61 to be determined. A speed for the feature 61 can then be determined from the speed sector and the time difference between the images shown in Figs. 4 and 5.The speed determination described above can be performed for each of the features evaluated as consistent, whereby the average speed of the vehicle 60 can then be calculated by averaging the speeds previously determined for each individual feature. In this case, the speed vectors considered to be outliers can be neglected in the speed determination.

[0063] According to a further aspect of the present invention, a method for camera-based determination of the speed of a vehicle in road traffic is also proposed, wherein the method comprises the following method steps:

[0064] Creating a first image at a first time using a first image sensor and creating a second image at the first time using a second image sensor;

[0065] Creating a third image at a second time using the first image sensor and creating a fourth image at the second time using the second image sensor;

[0066] Recognition of features in individual images;

[0067] comparing the features in the images recorded by the first image sensor with the features contained in the images recorded by the second image sensor at the same time, whereby those features which are contained only in the images recorded by the first image sensor or in the images recorded by the second image sensor are discarded;

[0068] Comparing the corresponding features in the images recorded by the first image sensor with the features in the images recorded by the second image sensor with regard to their epipolar geometry, whereby those features for which the epipolar conditions are not met are discarded;

[0069] comparing the features in the images captured by the first image sensor, discarding those features that are not included in both images captured by the first image sensor and those features that are not included in both images captured by the second image sensor;

[0070] Determining features whose position has remained unchanged from the first time point to the second time point and removing those features whose position has remained unchanged;

[0071] Determining the spatial position of the features using a triangulation method;

[0072] Determination of a velocity vector for each feature that was not previously rejected;

[0073] Determining the velocity for each feature from the velocity vector determined for the corresponding feature and the difference between the first time point and the second time point;

[0074] Determination of an average vehicle speed by averaging the determined speeds for each characteristic. LIST OF REFERENCE SYMBOLS

[0075] Process first process step second process step third process step fourth process step fifth process step sixth process step seventh process step eighth process step ninth process step tenth process step eleventh process step

[0076] System first image sensor second image sensor

[0077] Evaluation unit

[0078] Detection range

[0079] vehicle

[0080] Feature first image capture second image capture third image capture fourth image capture

Claims

CLAIMS Method (10) for camera-based determination of the speed of a vehicle (60) in road traffic, wherein the method (10) comprises the following procedural steps: Creating (11) a first image capture (62) at a first time using a first image sensor (52) and creating a second image capture (64) at the first time using a second image sensor (54); Creating (12) a third image capture (66) at a second time using the first image sensor (52) and creating a fourth image capture (68) at the second time using the second image sensor (54); Recognition (13) of features (61) in the individual image recordings; comparison (14) of the features (61) in the image recordings (62, 66) taken by the first image sensor (52) with the features (61) contained in the image recordings (64, 68) taken by the second image sensor (54) at the same time, discarding those features (61) that are contained only in the image recordings (62, 66) taken by the first image sensor (52) or in the image recordings (64, 68) taken by the second image sensor (54); Comparison (15) of the corresponding features (61) in the images (62, 66) taken by the first image sensor (52) with the features (61) in the images (64, 68) taken by the second image sensor (54) with respect to their epipolar geometry, discarding those features (61) for which the epipolar condition is not met; Comparison (16) of the features (61) in the image recordings (62, 66) taken by the first image sensor (52), wherein those features (61) that are not included in both image recordings (64, 68) taken by the first image sensor (52) and those features (61) that are not included in both images (64, 68) taken by the second image sensor (54) are discarded; Determination (17) of features (61) whose position has remained unchanged from the first time to the second time and removal of those features (61) whose position has remained unchanged; Determination (18) of the spatial position of the features (61) using a triangulation procedure; Determination (19) of a velocity vector for each feature (61); Determination (20) of the velocity for each feature (61) from the velocity vector determined for the corresponding feature (61); and Determination (21) of an average speed of the vehicle (60) by averaging the determined speeds for each feature (61). Method (10) according to claim 1, further characterized by the following method step: Verification of the consistency of the velocity vectors using a RANSAC method, wherein those velocity vectors that were evaluated as inconsistent are discarded. Method (10) according to claim 1 or 2, characterized in that a time signature is generated during the creation of each image acquisition (62, 64, 66, 68), which is appended to or assigned to the image acquisition (62, 64, 66, 68). Method (10) according to any one of claims 1 to 3, further characterized by the following method step: Verification of the consistency of the direction of the individual velocity vectors, wherein the angle between each individual velocity vector and a predetermined reference line is determined, an angle mean is calculated from the determined angle values, and those velocity vectors are considered inconsistent for which the difference between the determined angle and the angle mean is greater than a predetermined limit, and wherein those velocity vectors considered inconsistent are discarded. Method (10) according to any one of claims 1 to 4, further characterized by the following method step: Verification of the consistency of the magnitudes of the individual velocity vectors using a RANSAC method, wherein those velocity vectors deemed inconsistent are discarded. Method (10) according to any one of claims 1 to 5, further characterized by the following method step: Comparison of the corresponding features (61) in the images (62, 66) taken by the first image sensor (52) at the first and second time points, and comparison of the corresponding features (61) in the images (64, 68) taken by the second image sensor (54) at the first and second time points with respect to their epipolar geometry, discarding those features (61) for which the epipolar condition is not met. System (50) for determining the speed of a vehicle (60) in road traffic, comprising a first image sensor (52) for producing a first image capture (62) at a first time and a third image capture (66) at a second time; a second image sensor (54) for producing a second image capture (64) at the first time and a fourth image capture (68) at the second time; and an evaluation unit (56), wherein the external unit (56) is configured to to identify features (61) in the individual image recordings (62, 64, 66, 68); to compare the features (61) in the image recordings (62, 66) taken by the first image sensor (52) with the features (61) contained in the image recordings (64, 68) taken by the second image sensor (54) and to discard those features (61) that are contained only in the image recordings (62, 66) taken by the first image sensor (52) or in the image recordings (64, 68) taken by the second image sensor (54); to compare the corresponding features (61) in the images (62, 66) taken by the first image sensor (52) with the features (61) in the images (64, 68) taken by the second image sensor (54) with respect to their epipolar geometry and to discard those features (61) for which the epipolar condition is not met;to compare the features (61) in the images (62, 66) taken by the first image sensor (52) and to discard those features (61) that are not included in both images (62, 66) taken by the first image sensor (52) and those features (61) that are not included in both images (64, 68) taken by the second image sensor (54); to determine the features (61) whose position has remained unchanged from the first time to the second time; to remove those features (61) whose position has remained unchanged; to determine the spatial position of each feature (61) using a triangulation method; to determine a velocity vector for each feature (61); to determine the velocity for each feature (61) from the velocity vector determined for the corresponding feature (61); and to determine an average speed of the vehicle (60) by averaging the determined velocities for each feature (61). System (50) according to claim 7, characterized in that the evaluation unit (56) is additionally configured to check the consistency of the velocity vectors using a RANSAC method, discarding those velocity vectors that were evaluated as inconsistent.System (50) according to claim 7 or 8, characterized by a timer configured to generate a time signature at the time of an image acquisition (62, 64, 66, 68), which is appended to or assigned to the respective image acquisition (62, 64, 66, 68). System (50) according to any one of claims 7 to 9, characterized in that the evaluation unit (56) is configured to check the consistency of the direction of the individual velocity vectors, wherein the angle between each individual velocity vector and a predetermined reference line is determined, and an angular mean value is calculated from the determined angular values, and those velocity vectors are considered inconsistent where the difference between the determined angular value and the angular mean value is greater than a predetermined limit, and wherein those velocity vectors considered inconsistent are discarded. System (50) according to any one of claims 7 to 10, characterized in that the evaluation unit (56) is further configured to check the consistency of the magnitude of the individual velocity vectors using a RANSAC method, discarding those velocity vectors that are considered inconsistent. System (50) according to any one of claims 7 to 11, characterized in that the evaluation unit (56) is additionally configured to compare the corresponding features (61) in the image recordings (62, 66) taken by the first image sensor (52) at the first and second time points and to compare the corresponding features (61) in the image recordings (64, 68) taken by the second image sensor at the first and second time points with respect to their epipolar geometry, discarding those features (61) for which the epipolar condition is not met.

Images