CN113436190A - Lane line quality calculation method and device based on lane line curve coefficient and automobile - Google Patents
Lane line quality calculation method and device based on lane line curve coefficient and automobile Download PDFInfo
- Publication number
- CN113436190A CN113436190A CN202110875981.4A CN202110875981A CN113436190A CN 113436190 A CN113436190 A CN 113436190A CN 202110875981 A CN202110875981 A CN 202110875981A CN 113436190 A CN113436190 A CN 113436190A
- Authority
- CN
- China
- Prior art keywords
- lane line
- curve
- coefficient
- container
- line data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
- G06T2207/30252—Vehicle exterior; Vicinity of vehicle
- G06T2207/30256—Lane; Road marking
Abstract
The invention provides a lane line quality calculation method based on a lane line curve coefficient, which comprises the following steps: storing the lane line data fused and output by the multi-sensor data fusion device of one frame of vehicle into a container every time the lane line data is newly received; the lane line data includes: lane line curve coefficients and lane line lengths; when n +1 th frame lane line data to be stored is received, discarding the first frame lane line data currently stored in a container to enable the maximum buffer amount in the container to be n frames of lane line data; when a new frame of lane line data is stored in the container, respectively solving the standard deviation of each lane line curve coefficient; taking the standard deviation of each lane curve coefficient obtained by solving and a preset standard deviation upper limit value of each lane curve coefficient to be small, and selecting a final standard deviation of each lane curve coefficient; and calculating the lane line quality based on the final standard deviation of the curve coefficient of each lane line.
Description
Technical Field
The invention relates to the technical field of automatic driving perception fusion, in particular to a lane line quality calculation method and device based on lane line curve coefficients and an automobile.
Background
In the automatic driving perception fusion technology, lane line fusion is a necessary process for realizing transverse control. The lateral control of an autonomous vehicle generally generates a central trajectory line through a lane line, so that the vehicle travels along the trajectory to achieve a lateral control effect. When the lane line is used for transverse control, the effect of the transverse control is determined not only by the quality of a control algorithm, but also by whether the lane line at the input end is good enough to be used for the transverse control. The existing automatic driving system mainly judges whether the lane line can be used for control at a control end, and mostly depends on a single characteristic directly output by a sensor end to judge the quality of a certain aspect of the lane line, and does not have comprehensive prior knowledge to represent the quality of the lane line, so that the judging method is too single and cannot comprehensively represent the quality of the lane line.
The existing method for representing the quality of the lane line mainly has the following defects:
the single characteristic of the image end output by the sensor is relied on to represent the quality of the lane line, such as the segmentation confidence coefficient of the lane line segmented from the original image and the fitting error of the sensor when the sensor is used for lane line fitting. The single characteristic often represents the quality of a certain specific characteristic of the lane line, and the overall quality of the lane line cannot be represented;
the characteristics of different sensor outputs, which are used for representing the quality of the lane line, are inherently different due to different sensor types, and the characteristics are difficult to be unified under the same standard when the lane lines are fused.
Disclosure of Invention
The invention provides a lane line quality calculation method and device based on lane line curve coefficients and an automobile, which are used for automatic driving perception fusion and fusion of assessment of the quality of lane lines, so that the quality of the lane lines is represented in a quantitative mode, and reliable priori knowledge is provided for transverse control of an automatic driving system to judge whether the lane lines meet the control requirements.
The technical scheme of the invention is as follows:
the invention provides a lane line quality calculation method based on a lane line curve coefficient, which comprises the following steps:
storing the lane line data fused and output by the multi-sensor data fusion device of one frame of vehicle into a container every time the lane line data is newly received; the lane line data includes: coefficient of lane curve A0、A1、A2And lane line length range; wherein, when connectingWhen n +1 th frame lane line data to be stored is received, discarding the first frame lane line data currently stored in the container, so that the maximum buffer storage amount in the container is n frames of lane line data;
when a new frame of lane line data is stored in the container, the curve coefficient A of each lane line is measured0、A1、A2Respectively solving standard deviations;
the curve coefficient A of each line obtained by solving0、A1、A2Standard deviation of (a) and a preset coefficient A of each line curve0、A1、A2The upper limit value of the standard deviation is taken to be small, and the curve coefficient A of each line is selected0、A1、A2The final standard deviation of (d);
based on curve coefficient A of each lane line0、A1、A2And calculating the lane line quality.
Preferably, the curve coefficient a is set for each lane line every time a new frame of lane line data is stored in the container0、A1、A2The steps of respectively solving the standard deviation comprise:
calculating the curve coefficient A of each lane line when storing a new frame of lane line data in the container in a recursion mode0、A1、A2The mean value of (a);
each lane curve coefficient A based on last calculation0、A1、A2Calculating the curve coefficient A of each lane line when a new frame of lane line data is stored in the container0、A1、A2Standard deviation of (2).
Preferably, before storing a frame of new lane line data fused and output by a multi-sensor data fusion device of the vehicle into the container, the method further comprises:
carrying out validity detection on the lane line data so as to filter the invalid lane line curve coefficient value in the lane line data;
the specific steps of detecting the validity of the lane line data comprise:
number of lane linesAccording to the curve coefficient A of each lane line0、A1、A2Are all numerical type, and the curve coefficient A of each lane line0、A1、A2The values of (a) are all not zero, and the lane length range in the lane data is greater than zero.
Preferably, the coefficient A is based on the curve of each lane line0、A1、A2The step of calculating lane line quality comprises:
based on curve coefficient A of each lane line0、A1、A2Respectively calculating the curve coefficient A of each lane line0、A1、A2Coefficient standard deviation quality of (2);
based on curve coefficient A of each lane line0、A1、A2And calculating the lane line quality.
Preferably, by the formula:
calculating the curve coefficient A of the lane line0、A1、A2Lane line quality qua l ity (A)i) Wherein a isiIs a coefficient of a curve of a lane line A0、A1、A2Coefficient of any one lane line curve in (A), stdev (A)i) Is the standard deviation of the ith lane line curve coefficient, threstd (A)i) The standard deviation is the preset standard deviation upper limit value of the ith lane line curve coefficient.
Preferably, the curve coefficient a of each lane line when a new frame of lane line data is stored in the container is calculated in a recursive manner0、A1、A2The step of averaging comprises:
by the formula:
separately counting the storage of a new frame in a containerCurve coefficient A of each lane line in lane line data0、A1、A2Mean value ofn;En-1For the mean value of the coefficients of the corresponding lane line curve calculated when storing the lane line data of the previous frame in the container, when n is equal to 1, En-1Is zero; x is the number ofnStoring the corresponding lane line coefficient value in the new frame of lane line data in the container; n is the corresponding bit number when the received new frame of lane line data is stored in the container;
each lane curve coefficient A based on last calculation0、A1、A2Calculating the curve coefficient A of each lane line when a new frame of lane line data is stored in the container0、A1、A2The step of standard deviation of (a) comprises:
firstly, according to the formula:
respectively calculating the variance of curve coefficients a0, a1 and a2 of each lane line when a new frame of lane line data is stored in the container; fn-1F is a variance of a corresponding lane line curve coefficient calculated when storing the lane line data of the previous frame in the container when n is equal to 1n-1Is zero; x is the number ofnStoring the corresponding lane line coefficient value in the new frame of lane line data in the container; n is the corresponding bit number when the received new frame of lane line data is stored in the container; en-1For the mean value of the coefficients of the corresponding lane line curve calculated when storing the lane line data of the previous frame in the container, when n is equal to 1, En-1Is zero;
and then through the formula:
respectively calculating curve coefficient A of each lane line when storing a new frame of lane line data in a container0、A1、A2Standard deviation σ of (a).
Preferably, before the step of storing the newly received frame of lane line data in the container, the method further includes:
judging whether the receiving time interval between the currently received new frame of lane line data and the received previous frame of lane line data is checked by a preset threshold t imethresho l d or not;
if the number of the frames of lane line data exceeds the number of the frames of lane line data, discarding the currently received new frame of lane line data and all the currently stored frame of lane line data in the container, and storing the newly received lane line data into the container every time the lane line data output by the fusion of the multi-sensor data fusion device of one frame of vehicle is newly received;
otherwise, the step of storing the newly received frame of lane line data into the container is executed.
The invention provides a lane line quality calculating device based on a lane line curve coefficient, which comprises:
the storage module is used for storing the lane line data fused and output by the multi-sensor data fusion device of one frame of vehicle into a container every time the lane line data is newly received; the lane line data includes: lane curve coefficients a0, a1, a2, and lane length range; when n +1 th frame lane line data to be stored is received, discarding the first frame lane line data currently stored in a container, so that the maximum buffer amount in the container is n frames of lane line data;
a standard deviation calculation module for calculating curve coefficient A of each lane line when each new frame of lane line data is stored in the container0、A1、A2Respectively solving standard deviations;
a standard deviation limiting module for solving the obtained curve coefficient A of each line0、A1、A2Standard deviation of (a) and a preset coefficient A of each line curve0、A1、A2The upper limit value of the standard deviation is taken to be small, and the curve coefficient A of each line is selected0、A1、A2The final standard deviation of (d);
a lane line quality calculation module for calculating a lane line curve coefficient A based on each lane line0、A1、A2And calculating the lane line quality.
The invention provides an automobile which comprises the lane line quality calculating device based on the lane line curve coefficient.
The invention has the beneficial effects that:
the method for measuring the fluctuation degree of the lane line by using the coefficient standard deviation of the lane line curve is simple, efficient and convenient to calculate;
although the single characteristic representation lane line of the image end output by the sensor has different side points, the single characteristic representation lane line can reflect the fluctuation condition of the output lane line, so the invention can directly calculate the lane line quality from the fluctuation degree of the lane line, can effectively avoid the one-sidedness of the original single characteristic and can integrally measure the lane line quality;
the single characteristics of different sensors for representing the quality of the lane lines can be different, but the lane lines output by different sensors are all represented in a curve form, so that the method does not consider the processes of curve extraction, segmentation and fitting, starts from the curve, and well solves the problem that the characteristics of different sensors for representing the quality of the lane lines are difficult to unify to the same standard through fusion.
Drawings
FIG. 1 is a diagram of the steps of the method of the present invention.
Detailed Description
In the existing automatic driving technology, when lane lines are fused, due to the inherent difference of different sensors, a fusion end cannot effectively combine single characteristics of different sensors for representing the quality of the lane lines to obtain a good fusion characteristic. The transverse control is highly dependent on the quality of the lane line, but the control end often judges whether the quality of the lane line is good enough and can be used for transverse control through single characteristics output by the sensor. The single characteristics such as the lane line segmentation confidence, the fitting error and the like cannot comprehensively and comprehensively reflect the lane line quality, so the effect of judging the lane line quality by the automatic driving lateral control is also influenced. According to the defects of the lane line quality representation calculation in the lane line fusion, the invention provides a lane line quality calculation method by using a lane line curve coefficient. And calculating the standard deviation of the curve coefficients of the lane lines, and weighting and combining the standard deviations of the curve coefficients of the lane lines to obtain final measurement to represent the quality of the lane lines.
In general, in an automatic driving system, the most intuitive expression of the lane line is the fluctuation degree of the curve. The curve is determined by the coefficient, and the standard deviation of the curve coefficient of the lane line can reflect the fluctuation condition of the curve coefficient of the lane line, so that the fluctuation degree of the curve can be reflected by the standard deviation of the curve coefficient of the lane line, and the quality of the curve of the lane line can be measured.
The invention adopts the technical scheme that the lane line quality calculation method based on the curve coefficient is adopted. The method comprises the steps of determining the size of a standard deviation calculation window, judging the effectiveness of the lane line, resetting and judging the standard deviation calculation window, calculating the coefficient standard deviation, normalizing the coefficient standard deviation, calculating the quality of each coefficient standard deviation, and generating the quality of the lane line based on the coefficient standard deviation.
At present, a cubic polynomial curve is mainly adopted for lane line fitting in the automatic driving perception fusion, and the curve is as follows:
y=A0+A1*x+A2*x 2+A3*x 3
wherein A isiAnd i is 0,1,2,3, which represents the coefficient of the curve. A. the3The change of the curvature of the characteristic curve is represented, under the actual road condition, A3Cannot be directly reflected in the overall fluctuation of the curve, i.e. A3The fluctuations of the curve cannot be clearly distinguished. Therefore, the invention only adopts A when calculating the lane line quality0、A1And A2Three coefficients.
The method comprises the following detailed steps:
1. determination of standard deviation calculation window size
Before calculating the standard deviation of the curve coefficient, a first-in first-out container with a certain size is constructed, and the container is used for storing the coefficients of a certain number of historical frame lane lines. The method uses double-ended queues as storage containers. Under the actual automatic driving environment, the size of a better calculation window is greatly influenced by the speed of the vehicle. When the vehicle speed is faster, the distance traveled by the vehicle is large in a smaller window frame time, and when the vehicle speed is slower, more history frames are needed for the same distance. The larger the calculation window, the longer the distance the host vehicle travels within the window time. Although the change of the coefficient of the lane line is more gradual in general conditions, when the window is larger, the number of experienced historical frames is more, and the coefficient of the earlier historical frame in the coefficient window can not effectively reflect the coefficient condition of the current frame. That is, the current frame lane line coefficient has a larger change compared with the earlier historical frame lane line coefficient, and the standard deviation of the calculated coefficient is larger, so that the fluctuation condition of the current lane line cannot be effectively reflected. In summary of actual driving conditions, the present invention sets the size of the standard deviation calculation window to 10, i.e., calculates the standard deviation of the coefficients of 10 frames of history each time, for calculating lane line quality.
2. Lane line validity judgment
Before storing the lane line data into the double-ended queue created in step 1, validity judgment needs to be performed on the lane line first, and invalid coefficient values are filtered out to ensure validity of subsequent calculation. The method mainly judges the validity of the lane line according to the lane line coefficient and the length of the lane line, and comprises the following steps:
1) whether or not the lane line curve coefficient is non-0
condition01=(A0≠0)∨(A1≠0)∨(A2≠0)∨(A3≠0)
2) Whether the curve coefficients of the lane lines are all numerical values
condition02=(A0≠NAN)∧(A1≠NAN)∧(A2≠NAN)∧(A3≠NAN)
3) Whether the length of the lane line is greater than 0
condition03=Range>0
If conditions 01, 02, and 03 are true, the current lane line is valid and its coefficients are stored in the standard deviation calculation window, otherwise the lane line is invalid and its coefficients are not stored in the calculation window.
3. Standard deviation calculation window reset judgment
And judging whether the interval time delta t between the current frame lane line data and the previous frame lane line data exceeds a preset threshold value TimeThreshold or not, if so, indicating that the interval time between the previous frame data and the next frame data is too long, and if not, indicating that the historical data in the standard deviation calculation window is unavailable, and emptying the historical data to reset the standard deviation calculation window.
4. Calculation of standard deviation of curve coefficient of lane line
The standard deviation of the curve coefficient of the lane line is calculated in a recursion mode, and the calculation steps are as follows:
1) calculating the mean value of each curve coefficient in the window by recursion
2) Updating the mean value of the last calculation as the current mean value
3) Standard deviation of curve coefficient in recursion calculation window
4) Updating the last calculated standard deviation to the current standard deviation
Wherein E isn-1、EnRespectively the mean of the last calculation and the mean of the current calculation, Fn-1、FnRespectively, the variance calculated last time and the variance calculated currently, wherein sigma is the current standard deviation, and n is the number of corresponding coefficients in the current calculation window. In the invention, n is gradually increased from 1 to 10 at most according to calculation.
5. Normalization of coefficient standard deviation
In order to calculate a bounded quality value, the standard deviation of the individual curve coefficients calculated in step 4 is limited to a corresponding range, namely:
after normalization
StdevA0∈[0,ThresStdA0]
StdevA1∈[0,ThresStdA1]
StdevA2∈[0,ThresStdA2]
Wherein, ThreshstdA 0, ThreshstdA 1 and ThreshstdA 2 are lane line coefficients A0、A1And A2Upper limit of standard deviation of (a). The upper definite limits of the three standard deviations are respectively 0.1, 0.005 and 0.0001.
After the step is executed, the curve coefficient A of each lane line is obtained0、A1、A2The final standard deviation of (d).
6. Quality calculation of standard deviation of each coefficient
And respectively calculating the quality of each coefficient standard deviation based on the coefficient standard deviations calculated in the step 5.
7. Lane line quality generation based on coefficient standard deviation
And (4) calculating the average value by using the standard deviation quality calculated based on each coefficient obtained in the step (6) to obtain the final lane line quality.
The method of the invention has the following effects:
the method for measuring the fluctuation degree of the lane line by using the coefficient standard deviation of the lane line curve is simple, efficient and convenient to calculate;
although the single characteristic representation lane line of the image end output by the sensor has different side points, the single characteristic representation lane line can reflect the fluctuation condition of the output lane line, so the invention can directly calculate the lane line quality from the fluctuation degree of the lane line, can effectively avoid the one-sidedness of the original single characteristic and can integrally measure the lane line quality;
the single characteristics of different sensors for representing the quality of the lane lines can be different, but the lane lines output by different sensors are all represented in a curve form, so that the method does not consider the processes of curve extraction, segmentation and fitting, starts from the curve, and well solves the problem that the characteristics of different sensors for representing the quality of the lane lines are difficult to unify to the same standard through fusion.
The invention provides a lane line quality calculating device based on a lane line curve coefficient, which comprises:
the storage module is used for storing the lane line data fused and output by the multi-sensor data fusion device of one frame of vehicle into a container every time the lane line data is newly received; the lane line data includes: lane curve coefficients a0, a1, a2, and lane length range; when n +1 th frame lane line data to be stored is received, discarding the first frame lane line data currently stored in a container, so that the maximum buffer amount in the container is n frames of lane line data;
a standard deviation calculation module for calculating curve coefficient A of each lane line when each new frame of lane line data is stored in the container0、A1、A2Respectively solving standard deviations;
a standard deviation limiting module for solving the obtained curve coefficient A of each line0、A1、A2Standard deviation of (a) and a preset coefficient A of each line curve0、A1、A2The upper limit value of the standard deviation is taken to be small, and the curve coefficient A of each line is selected0、A1、A2The final standard deviation of (d);
a lane line quality calculation module for calculating a lane line curve coefficient A based on each lane line0、A1、A2And calculating the lane line quality.
The invention provides an automobile which comprises the lane line quality calculating device based on the lane line curve coefficient.
Claims (9)
1. A lane line quality calculation method based on lane line curve coefficients is characterized by comprising the following steps:
storing the lane line data fused and output by the multi-sensor data fusion device of one frame of vehicle into a container every time the lane line data is newly received; the lane line data includes: coefficient of lane curve A0、A1、A2And lane line length range; when n +1 th frame lane line data to be stored is received, discarding the first frame lane line data currently stored in a container, so that the maximum buffer amount in the container is n frames of lane line data;
when a new frame of lane line data is stored in the container, the curve coefficient A of each lane line is measured0、A1、A2Respectively solving standard deviations;
solving the curve coefficient A of each lane line0、A1、A2Standard deviation of (a) and a preset curve coefficient A of each lane line0、A1、A2The upper limit value of the standard deviation is taken to be small, and the curve coefficient A of each line is selected0、A1、A2The final standard deviation of (d);
based on curve coefficient A of each lane line0、A1、A2And calculating the lane line quality.
2. The method of claim 1, wherein the curve coefficient A is determined for each lane line every time a new frame of lane line data is stored in the container0、A1、A2The steps of respectively solving the standard deviation comprise:
calculating the curve coefficient A of each lane line when storing a new frame of lane line data in the container in a recursion mode0、A1、A2The mean value of (a);
each lane curve coefficient A based on last calculation0、A1、A2Calculating the curve coefficient A of each lane line when a new frame of lane line data is stored in the container0、A1、A2Standard deviation of (2).
3. The method according to claim 1, wherein before storing a frame of new lane line data fused and output by a multi-sensor data fusion device of a vehicle into a container, the method further comprises:
carrying out validity detection on the lane line data so as to filter the invalid lane line curve coefficient value in the lane line data;
the specific steps of detecting the validity of the lane line data comprise:
each lane line curve coefficient a in the lane line data0、A1、A2Are all numerical type, and the curve coefficient A of each lane line0、A1、A2The values of (a) are all not zero, and the lane length range in the lane data is greater than zero.
4. The method of claim 1, wherein a coefficient A is based on each lane line curve0、A1、A2The step of calculating lane line quality comprises:
based on curve coefficient A of each lane line0、A1、A2Respectively calculating the curve coefficient A of each lane line0、A1、A2Coefficient standard deviation quality of (2);
based on curve coefficient A of each lane line0、A1、A2And calculating the lane line quality.
5. The method of claim 4, characterized by the formula:
calculating the curve coefficient A of the lane line0、A1、A2Lane line quality (A)i) Wherein a isiIs a coefficient of a curve of a lane line A0、A1、A2Coefficient of any one lane line curve in (A), stdev (A)i) Is the standard deviation of the ith lane line curve coefficient, threstd (A)i) The standard deviation is the preset standard deviation upper limit value of the ith lane line curve coefficient.
6. The method of claim 2, wherein the lane line curve coefficients a for each frame of lane line data stored in the container are calculated recursively0、A1、A2The step of averaging comprises:
by the formula:
respectively calculating curve coefficient A of each lane line when storing a new frame of lane line data in a container0、A1、A2Mean value ofn;En-1For the mean value of the coefficients of the corresponding lane line curve calculated when storing the lane line data of the previous frame in the container, when n is equal to 1, En-1Is zero; x is the number ofnStoring the corresponding lane line coefficient value in the new frame of lane line data in the container; n is the corresponding bit number when the received new frame of lane line data is stored in the container;
each lane curve coefficient A based on last calculation0、A1、A2Calculating the curve coefficient A of each lane line when a new frame of lane line data is stored in the container0、A1、A2The step of standard deviation of (a) comprises:
firstly, according to the formula:
respectively calculating curve coefficient A of each lane line when storing a new frame of lane line data in a container0、A1、A2The variance of (a); fn-1F is a variance of a corresponding lane line curve coefficient calculated when storing the lane line data of the previous frame in the container when n is equal to 1n-1Is zero; x is the number ofnStoring the corresponding lane line coefficient value in the new frame of lane line data in the container; n is the corresponding bit number when the received new frame of lane line data is stored in the container;En-1For the mean value of the coefficients of the corresponding lane line curve calculated when storing the lane line data of the previous frame in the container, when n is equal to 1, En-1Is zero;
and then through the formula:
respectively calculating curve coefficient A of each lane line when storing a new frame of lane line data in a container0、A1、A2Standard deviation σ of (a).
7. The method of claim 1, wherein prior to the step of storing the newly received frame of lane line data in a container, the method further comprises:
judging whether the receiving time interval between the currently received new frame of lane line data and the received previous frame of lane line data exceeds a preset threshold value timethreshold or not;
if the number of the frames of lane line data exceeds the number of the frames of lane line data, discarding the currently received new frame of lane line data and all the currently stored frame of lane line data in the container, and storing the newly received lane line data into the container every time the lane line data output by the fusion of the multi-sensor data fusion device of one frame of vehicle is newly received;
otherwise, the step of storing the newly received frame of lane line data into the container is executed.
8. A lane line quality calculation apparatus based on a lane line curve coefficient, comprising:
the storage module is used for storing the lane line data fused and output by the multi-sensor data fusion device of one frame of vehicle into a container every time the lane line data is newly received; the lane line data includes: coefficient of lane curve A0、A1、A2And lane line length range; when the n +1 th frame lane line data to be stored is received, the first frame lane line data currently stored in the container is processedDiscarding, wherein the maximum buffer storage amount in the container is n frames of lane line data;
a standard deviation calculation module for calculating curve coefficient A of each lane line when each new frame of lane line data is stored in the container0、A1、A2Respectively solving standard deviations;
a standard deviation limiting module for solving the obtained curve coefficient A of each line0、A1、A2Standard deviation of (a) and a preset coefficient A of each line curve0、A1、A2The upper limit value of the standard deviation is taken to be small, and the curve coefficient A of each line is selected0、A1、A2The final standard deviation of (d);
a lane line quality calculation module for calculating a lane line curve coefficient A based on each lane line0、A1、A2And calculating the lane line quality.
9. An automobile characterized by comprising the lane line quality calculation apparatus based on the lane line curve coefficient of claim 8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110875981.4A CN113436190B (en) | 2021-07-30 | 2021-07-30 | Lane line quality calculation method and device based on lane line curve coefficient and automobile |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110875981.4A CN113436190B (en) | 2021-07-30 | 2021-07-30 | Lane line quality calculation method and device based on lane line curve coefficient and automobile |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113436190A true CN113436190A (en) | 2021-09-24 |
CN113436190B CN113436190B (en) | 2023-03-14 |
Family
ID=77762576
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110875981.4A Active CN113436190B (en) | 2021-07-30 | 2021-07-30 | Lane line quality calculation method and device based on lane line curve coefficient and automobile |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113436190B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113886634A (en) * | 2021-09-30 | 2022-01-04 | 重庆长安汽车股份有限公司 | Lane line offline data visualization method and device |
CN115049997A (en) * | 2022-06-07 | 2022-09-13 | 北京百度网讯科技有限公司 | Method and device for generating edge lane line, electronic device and storage medium |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150227800A1 (en) * | 2014-02-07 | 2015-08-13 | Toyota Jidosha Kabushiki Kaisha | Marking line detection system and marking line detection method |
CN107590438A (en) * | 2017-08-16 | 2018-01-16 | 中国地质大学(武汉) | A kind of intelligent auxiliary driving method and system |
CN109409202A (en) * | 2018-09-06 | 2019-03-01 | 惠州市德赛西威汽车电子股份有限公司 | Robustness method for detecting lane lines based on dynamic area-of-interest |
CN109840463A (en) * | 2017-11-27 | 2019-06-04 | 北京图森未来科技有限公司 | A kind of Lane detection method and apparatus |
US20190251372A1 (en) * | 2018-02-13 | 2019-08-15 | Kpit Technologies Ltd | System and method for lane detection |
CN110160540A (en) * | 2019-06-12 | 2019-08-23 | 禾多科技(北京)有限公司 | Lane line data fusion method based on high-precision map |
CN111487971A (en) * | 2020-04-23 | 2020-08-04 | 重庆长安汽车股份有限公司 | Automatic driving transverse control method and system for vehicle |
CN111516673A (en) * | 2020-04-30 | 2020-08-11 | 重庆长安汽车股份有限公司 | Lane line fusion system and method based on intelligent camera and high-precision map positioning |
CN112818804A (en) * | 2021-01-26 | 2021-05-18 | 重庆长安汽车股份有限公司 | Parallel processing method and system for target level lane line, vehicle and storage medium |
CN113128307A (en) * | 2020-01-10 | 2021-07-16 | 阿里巴巴集团控股有限公司 | Lane line detection processing method, and related device and system |
-
2021
- 2021-07-30 CN CN202110875981.4A patent/CN113436190B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150227800A1 (en) * | 2014-02-07 | 2015-08-13 | Toyota Jidosha Kabushiki Kaisha | Marking line detection system and marking line detection method |
CN107590438A (en) * | 2017-08-16 | 2018-01-16 | 中国地质大学(武汉) | A kind of intelligent auxiliary driving method and system |
CN109840463A (en) * | 2017-11-27 | 2019-06-04 | 北京图森未来科技有限公司 | A kind of Lane detection method and apparatus |
US20190251372A1 (en) * | 2018-02-13 | 2019-08-15 | Kpit Technologies Ltd | System and method for lane detection |
CN109409202A (en) * | 2018-09-06 | 2019-03-01 | 惠州市德赛西威汽车电子股份有限公司 | Robustness method for detecting lane lines based on dynamic area-of-interest |
CN110160540A (en) * | 2019-06-12 | 2019-08-23 | 禾多科技(北京)有限公司 | Lane line data fusion method based on high-precision map |
CN113128307A (en) * | 2020-01-10 | 2021-07-16 | 阿里巴巴集团控股有限公司 | Lane line detection processing method, and related device and system |
CN111487971A (en) * | 2020-04-23 | 2020-08-04 | 重庆长安汽车股份有限公司 | Automatic driving transverse control method and system for vehicle |
CN111516673A (en) * | 2020-04-30 | 2020-08-11 | 重庆长安汽车股份有限公司 | Lane line fusion system and method based on intelligent camera and high-precision map positioning |
CN112818804A (en) * | 2021-01-26 | 2021-05-18 | 重庆长安汽车股份有限公司 | Parallel processing method and system for target level lane line, vehicle and storage medium |
Non-Patent Citations (3)
Title |
---|
RUMAISA RAMADHANI 等: "Line Detection Using Arranging Coordinate Point Method", 《2019 6TH INTERNATIONAL CONFERENCE ON ELECTRICAL ENGINEERING, COMPUTER SCIENCE AND INFORMATICS (EECSI)》 * |
吴俊丽: "基于单目视觉的无人驾驶汽车轨迹跟踪控制系统研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
王珍: "智能驾驶中的车道线检测算法研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113886634A (en) * | 2021-09-30 | 2022-01-04 | 重庆长安汽车股份有限公司 | Lane line offline data visualization method and device |
CN113886634B (en) * | 2021-09-30 | 2024-04-12 | 重庆长安汽车股份有限公司 | Lane line offline data visualization method and device |
CN115049997A (en) * | 2022-06-07 | 2022-09-13 | 北京百度网讯科技有限公司 | Method and device for generating edge lane line, electronic device and storage medium |
CN115049997B (en) * | 2022-06-07 | 2023-03-17 | 北京百度网讯科技有限公司 | Method and device for generating edge lane line, electronic device and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN113436190B (en) | 2023-03-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113436190B (en) | Lane line quality calculation method and device based on lane line curve coefficient and automobile | |
CN111650345B (en) | Method, device, equipment and medium for processing atmospheric environmental pollution detection data | |
US20040257269A1 (en) | Method and device for determining an expectancy range for a filling level echo and a false echo | |
CN117278643B (en) | Vehicle-mounted cloud calibration data transmission system based on cloud edge cooperation | |
CN113127305A (en) | Abnormality detection method and apparatus | |
CN112325967A (en) | Intelligent Internet of things water meter working state early warning method and system | |
CN115860510A (en) | Production efficiency analysis and evaluation method based on big data | |
CN111695735B (en) | Flow calculation-based railway bow net real-time early warning method, system and device | |
CN104457911B (en) | A kind of liquid level detection system and liquid-level detecting method | |
CN113071509A (en) | Fault detection method for wheel angle sensor of intelligent vehicle | |
CN111353131B (en) | Code carrier deviation degree threshold value calculation method | |
CN111091077B (en) | Vehicle speed detection method based on image correlation and template matching | |
CN109341650B (en) | Unmanned aerial vehicle elevation error double-threshold correction method based on minimum detection cost | |
CN114662060B (en) | Vehicle-mounted nitrogen oxide sensor concentration measurement value correction method based on machine learning | |
CN116522096A (en) | Three-dimensional digital twin content intelligent manufacturing method based on motion capture | |
CN113747270B (en) | Cloud data center thing allies oneself with management and control system | |
CN112874528B (en) | Vehicle speed determination method, vehicle speed determination device, vehicle-mounted terminal and storage medium | |
CN108733812B (en) | Identification method of abnormal data points in time series data based on global information | |
CN110595669B (en) | Device and method for detecting pressure of pressure container with safety valve | |
CN109870404B (en) | Rain shed structure damage identification method and device and terminal equipment | |
CN113251962A (en) | Ultrasonic parking space compensation system based on machine learning | |
CN104965980B (en) | Method of abnormal value removing and correction and device | |
CN111322984A (en) | Altitude calculation method and device, unmanned aerial vehicle and storage medium | |
CN115985143B (en) | Flight type judging method suitable for enhanced ground proximity warning system | |
CN114527043B (en) | Particle concentration measuring method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |