CN113837295A - Cosine similarity relation calculation method for automatic parking of ultrasonic sensor - Google Patents
Cosine similarity relation calculation method for automatic parking of ultrasonic sensor Download PDFInfo
- Publication number
- CN113837295A CN113837295A CN202111139803.1A CN202111139803A CN113837295A CN 113837295 A CN113837295 A CN 113837295A CN 202111139803 A CN202111139803 A CN 202111139803A CN 113837295 A CN113837295 A CN 113837295A
- Authority
- CN
- China
- Prior art keywords
- type
- obstacle
- contour
- obstacles
- parking space
- 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.)
- Withdrawn
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a cosine similarity relation calculation method for automatic parking of an ultrasonic sensor, which comprises the steps of S1, establishing obstacle classification, S2, collecting data outlines of different obstacles, establishing an obstacle type library, and S3, setting outline similarity parameters and S4, matching obstacle types. The method has the advantages that the obstacle classification is established, the obstacles are divided into a plane type, a square type, a cylindrical type, a slightly pointed circular type, a circular type with a small radian and a plane type with a large radian, common obstacles in an actual scene are divided, follow-up obstacle type matching is facilitated, calculation is facilitated, the calculation similarity of the actual contour points and the type calculation similarity of an obstacle type library is directly used for matching, the obstacle classification process is simplified by using a cosine similarity calculation mode, the matched obstacle types have real boundary coordinates, the actual parking space boundary coordinates are obtained according to the types of the obstacles around the parking space, and the identification precision of the ultrasonic parking space is improved.
Description
Technical Field
The invention belongs to the technical field of automatic parking, and particularly relates to a cosine similarity relation calculation method for automatic parking of an ultrasonic sensor.
Background
An ultrasonic sensor is one of sensors commonly used in an automatic parking system, and ultrasonic waves can accurately detect the distance of an obstacle, but cannot output a specific obstacle type. The boundary of the existing ultrasonic parking space detection adopts the distance change between the jumping points of the ultrasonic detection to determine the parking space boundary, the existing ultrasonic radar obtains the shape types of obstacles at the upper edge and the lower edge of a target parking space, the shapes of the obstacles are divided into three types of square, round and unknown, the target parking space length is calculated according to the distance between the left and right obstacles of the target parking space and the compensation value of the left and right obstacles of the target parking space, the edge echo signals of the left and right obstacles of the parking space are fitted into the representation form of quadratic polynomial, the quadratic coefficient of the quadratic polynomial is taken as the characteristic value of the edge echo signal, the measurement compensation value of the edge position of the obstacle to be detected is determined according to the characteristic value, further the vehicle space length is calculated according to the initial measurement value and the measurement compensation value of the edge position of the obstacle to be detected, in the actual scene, different obstacle types are different from the reflection signals of the ultrasonic wave, the jumping points of different obstacle types are unstable, and accurate parking space boundary detection precision is difficult to obtain.
Therefore, the existing calculation method of the ultrasonic sensor has the following problems:
1. the classification of the obstacles is basically three types of square, round and unknown types, the classification is too few, the obstacles contained in each type in the actual scene are too complex, and the classification effect is not a key point for the subsequent calculation of the parking space boundary effect;
2. the parking space compensation needs to calculate compensation values according to the parking space profile characteristics each time, and the similar profiles cause unnecessary calculation;
3. the detection precision of the left and right boundaries of the ultrasonic parking space is not high.
Disclosure of Invention
The invention aims to provide a cosine similarity relation calculation method for automatic parking of an ultrasonic sensor, and aims to solve the problems that in the use process of the conventional automatic parking calculation method of the ultrasonic sensor, the obstacles are classified into three types, namely square, round and unknown types, the categories are too few, the obstacles contained in each type in an actual scene are too complex, the parking space compensation needs to calculate compensation values respectively according to the parking space profile characteristics every time, the similar profiles cause unnecessary calculation, and the detection precision of the left and right boundaries of an ultrasonic parking space is not high.
In order to achieve the purpose, the invention provides the following technical scheme: a cosine similarity relation calculation method for automatic parking of an ultrasonic sensor comprises the following steps of S1, establishing obstacle classification, S2, collecting data outlines of different obstacles, establishing an obstacle type library, S3, setting outline similarity parameters and S4, matching obstacle types, and specifically comprises the following steps:
s1, establishing obstacle classification;
according to the actual scene, the parking space obstacles comprise vehicles, pillars, walls and the like, and the parking space obstacles are classified according to the outline of the type of the obstacles into a plane type, a square type, a cylinder type, a slightly pointed circle type, a circular type with a small radian and a plane type with a large radian;
s2, acquiring different barrier data profiles and establishing a barrier type library;
acquiring ultrasonic contour point information of different types of obstacles as an obstacle type library, recording left and right boundary coordinates corresponding to the obstacles as real boundaries of the obstacle type, and establishing an obstacle type library of the different obstacle types at different vehicle speeds;
s3, setting a contour similarity parameter;
the matching degree of the ultrasonic contour and the actual contour of the representation type library and the Euclidean distance between contour points are calculated as follows:
(1) selecting a contour center point (C), aligning the coordinates (x2 (R), y2 (R) of the actual contour (C)) with the coordinates (x2 (R), y2 (R)) of the type library contour center point (C), enabling x2 (R) to be x2 and y2 (R) to be y2, and calculating the vector between the boundary points (R) and (R) to (R)
(Vector)
Similarly calculating the contour in the contour library to obtain
And
respectively calculate
And the combination of (a) and (b),
cosine similarity s1 and s2 between the two sets of vectors to
The similarity s1 is calculated as an example:
wherein the content of the first and second substances,
represents the modulo length of the vector;
(2) respectively obtaining two groups of cosine similarity s1 and s2, and calculating the sum of the two groups s1+ s 2;
setting a threshold value, setting a threshold value delta, wherein the empirical preference value is 0.2, and the single similarity is not more than 0.15;
s4, matching the types of the obstacles;
obtaining type matching by calculating different similarity parameters and corresponding thresholds, carrying out ultrasonic parking space detection, detecting obstacles on the left and right sides of a parking space to obtain actual contour point information of the obstacles, matching the actual contour point information with the types of the obstacle type library in S2, obtaining the types and boundary values of the obstacles on the two sides of the parking space when the actual contour point information is smaller than the similarity threshold, calculating accurate ultrasonic parking space information, and outputting parking space boundary coordinates;
if the actually tested contour similarity does not meet the threshold value, a statistical method is adopted to directly calculate the width and the length of the whole contour, the length delta l and the width difference delta w are calculated, D is delta l plus delta w, the type with the closest width and length in all contour types is selected, the actual obstacle coordinates are obtained according to the type in the same way, and the actual parking space boundary coordinates are obtained.
Specifically, in step S1, the wall is a flat surface, the pillars are classified into a square shape and a cylindrical shape, and the vehicle types are more, and the classification bases refer to the types of common vehicles in the market, such as a slightly sharp circle, a square shape, a flat shape with a small arc, a flat shape with a large arc, and the like.
Compared with the prior art, the invention has the beneficial effects that:
1. by establishing the obstacle classification, the obstacles are divided into a plane type, a square type, a cylindrical type, a slightly pointed circular type, a circular type with a small radian and a plane type with a large radian, common obstacles in an actual scene are divided, follow-up obstacle type matching is facilitated, and calculation is facilitated.
2. The actual contour points are directly matched with the type calculation similarity of the barrier type library, and the barrier classification process is simplified by using a mode of calculating cosine similarity relation.
3. The matched obstacle type has real boundary coordinates, and the actual parking space boundary coordinates are obtained according to the types of the left and right obstacles of the parking space, so that the identification precision of the ultrasonic parking space is improved.
Drawings
FIG. 1 is an overall flow chart of the present invention;
FIG. 2 is a diagram of the obstacle category in step S1 according to the present invention;
FIG. 3 is a schematic diagram of the acquisition in step S2 according to the present invention;
fig. 4 is a schematic diagram of the calculation of the contour matching degree in step S3 according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1-4, the present invention provides a technical solution: a cosine similarity relation calculation method for automatic parking of an ultrasonic sensor comprises the following steps of S1, establishing obstacle classification, S2, collecting data outlines of different obstacles, establishing an obstacle type library, S3, setting outline similarity parameters and S4, matching obstacle types, and specifically comprises the following steps:
s1, establishing obstacle classification;
according to the actual scene, the parking space obstacles comprise vehicles, pillars, walls and the like, and the parking space obstacles are classified according to the outline of the type of the obstacles into a plane type, a square type, a cylinder type, a slightly pointed circle type, a circular type with a small radian and a plane type with a large radian.
S2, acquiring different barrier data profiles and establishing a barrier type library;
acquiring ultrasonic contour point information of different types of obstacles as an obstacle type library, recording left and right boundary coordinates corresponding to the obstacles as real boundaries of the obstacle type, and establishing an obstacle type library of the different obstacle types at different vehicle speeds;
s3, setting a contour similarity parameter;
the matching degree of the ultrasonic contour and the actual contour of the representation type library and the Euclidean distance between contour points are calculated as follows:
(1) selecting a contour center point (C), aligning the coordinates (x2 (R), y2 (R) of the actual contour (C)) with the coordinates (x2 (R), y2 (R)) of the type library contour center point (C), enabling x2 (R) to be x2 and y2 (R) to be y2, and calculating the vector between the boundary points (R) and (R) to (R)
(Vector)
Similarly calculating the contour in the contour library to obtain
And
respectively calculate
And the combination of (a) and (b),
cosine similarity s1 and s2 between the two sets of vectors to
The similarity s1 is calculated as an example:
wherein the content of the first and second substances,
represents the modulo length of the vector;
(2) respectively obtaining two groups of cosine similarity s1 and s2, and calculating the sum of the two groups s1+ s 2;
setting a threshold value, setting a threshold value delta, wherein the empirical preference value is 0.2, and the single similarity is not more than 0.15;
s4, matching the types of the obstacles;
obtaining type matching by calculating different similarity parameters and corresponding thresholds, carrying out ultrasonic parking space detection, detecting obstacles on the left and right sides of a parking space to obtain actual contour point information of the obstacles, matching the actual contour point information with the types of the obstacle type library in S2, obtaining the types and boundary values of the obstacles on the two sides of the parking space when the actual contour point information is smaller than the similarity threshold, calculating accurate ultrasonic parking space information, and outputting parking space boundary coordinates;
if the actually tested contour similarity does not meet the threshold value, a statistical method is adopted to directly calculate the width and the length of the whole contour, the length delta l and the width difference delta w are calculated, D is delta l plus delta w, the type with the closest width and length in all contour types is selected, the actual obstacle coordinates are obtained according to the type in the same way, and the actual parking space boundary coordinates are obtained.
In step S1, the wall is a flat surface, the pillars are classified into square and cylindrical, and the vehicle types are more, and the classification bases are slightly sharp circular, square, flat with small arc, flat with large arc, and the like, referring to the common vehicle types in the market.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (2)
1. A cosine similarity relation calculation method for automatic parking of an ultrasonic sensor is characterized by comprising the following steps: comprises that
S1, establishing obstacle classification, S2, collecting different obstacle data outlines and establishing an obstacle type library, S3, setting outline similarity parameters and S4, matching obstacle types, and specifically comprising the following steps:
s1, establishing obstacle classification;
according to the actual scene, the parking space obstacles comprise vehicles, pillars, walls and the like, and the parking space obstacles are classified according to the outline of the type of the obstacles into a plane type, a square type, a cylinder type, a slightly pointed circle type, a circular type with a small radian and a plane type with a large radian;
s2, acquiring different barrier data profiles and establishing a barrier type library;
acquiring ultrasonic contour point information of different types of obstacles as an obstacle type library, recording left and right boundary coordinates corresponding to the obstacles as real boundaries of the obstacle type, and establishing an obstacle type library of the different obstacle types at different vehicle speeds;
s3, setting a contour similarity parameter;
the matching degree of the ultrasonic contour and the actual contour of the representation type library and the Euclidean distance between contour points are calculated as follows:
(1) selecting a contour center point (C), aligning the coordinates (x2 (R), y2 (R) of the actual contour (C)) with the coordinates (x2 (R), y2 (R)) of the type library contour center point (C), enabling x2 (R) to be x2 and y2 (R) to be y2, and calculating the vector between the boundary points (R) and (R) to (R)
(Vector)
Similarly calculating the contour in the contour library to obtain
And
respectively calculate
And the combination of (a) and (b),
cosine similarity s1 and s2 between the two sets of vectors to
The similarity s1 is calculated as an example:
wherein the content of the first and second substances,
represents the modulo length of the vector;
(2) respectively obtaining two groups of cosine similarity s1 and s2, and calculating the sum of the two groups s1+ s 2;
setting a threshold value, setting a threshold value delta, wherein the empirical preference value is 1.8, and the single similarity is not more than 0.85;
s4, matching the types of the obstacles;
obtaining type matching by calculating different similarity parameters and corresponding thresholds, carrying out ultrasonic parking space detection, detecting obstacles on the left and right sides of a parking space to obtain actual contour point information of the obstacles, matching the actual contour point information with the types of the obstacle type library in S2, obtaining the types and boundary values of the obstacles on the two sides of the parking space when the actual contour point information is smaller than the similarity threshold, calculating accurate ultrasonic parking space information, and outputting parking space boundary coordinates;
if the actually tested contour similarity does not meet the threshold value, a statistical method is adopted to directly calculate the width and the length of the whole contour, the length delta l and the width difference delta w are calculated, D is delta l plus delta w, the type with the closest width and length in all contour types is selected, the actual obstacle coordinates are obtained according to the type in the same way, and the actual parking space boundary coordinates are obtained.
2. The cosine similarity calculation method for automatic parking with an ultrasonic sensor according to claim 1, characterized in that: in step S1, the wall is a flat surface, the pillars are classified into square and cylindrical, and the vehicle types are more, and the classification bases are slightly sharp circular, square, flat with small arc, flat with large arc, etc. referring to the common vehicle types in the market.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111139803.1A CN113837295A (en) | 2021-09-28 | 2021-09-28 | Cosine similarity relation calculation method for automatic parking of ultrasonic sensor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111139803.1A CN113837295A (en) | 2021-09-28 | 2021-09-28 | Cosine similarity relation calculation method for automatic parking of ultrasonic sensor |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113837295A true CN113837295A (en) | 2021-12-24 |
Family
ID=78970765
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111139803.1A Withdrawn CN113837295A (en) | 2021-09-28 | 2021-09-28 | Cosine similarity relation calculation method for automatic parking of ultrasonic sensor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113837295A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113869432A (en) * | 2021-09-28 | 2021-12-31 | 英博超算(南京)科技有限公司 | Contour point distance similarity calculation method for automatic parking of ultrasonic sensor |
-
2021
- 2021-09-28 CN CN202111139803.1A patent/CN113837295A/en not_active Withdrawn
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113869432A (en) * | 2021-09-28 | 2021-12-31 | 英博超算(南京)科技有限公司 | Contour point distance similarity calculation method for automatic parking of ultrasonic sensor |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111239766B (en) | Water surface multi-target rapid identification and tracking method based on laser radar | |
| WO2021008202A1 (en) | Method for kernel support vector machine target classification based on millimeter-wave radar point cloud features | |
| CN111045008B (en) | Vehicle millimeter wave radar target identification method based on widening calculation | |
| CN108681525B (en) | Road surface point cloud intensity enhancing method based on vehicle-mounted laser scanning data | |
| CN108549084B (en) | Target detection and attitude estimation method based on sparse two-dimensional laser radar | |
| CN107632308B (en) | Method for detecting contour of obstacle in front of vehicle based on recursive superposition algorithm | |
| CN112098990B (en) | Method for detecting and tracking medium and high speed vehicle by vehicle-mounted high-resolution millimeter wave radar | |
| CN110531357A (en) | Estimate the method and radar sensing system of mobile target velocity magnitude in a horizontal plane | |
| CN109444839B (en) | Target contour acquisition method and device | |
| CN110441788B (en) | Unmanned ship environment sensing method based on single line laser radar | |
| JP7152355B2 (en) | Obstacle detection device and obstacle detection method | |
| CN112083441A (en) | Obstacle detection method and system based on deep fusion of laser radar and millimeter wave radar | |
| CN111781608A (en) | Moving target detection method and system based on FMCW laser radar | |
| CN111999726B (en) | Personnel positioning method based on millimeter wave radar | |
| CN115840205B (en) | Terrain area metering method and system based on laser radar technology | |
| CN110488273B (en) | Vehicle tracking detection method and device based on radar | |
| CN113837295A (en) | Cosine similarity relation calculation method for automatic parking of ultrasonic sensor | |
| CN114609609A (en) | Speed estimation method for extracting static point cloud by FMCW laser radar random sampling | |
| CN108693517B (en) | Vehicle positioning method and device and radar | |
| CN113869409A (en) | Length and width ratio similarity calculation method for automatic parking of ultrasonic sensor | |
| CN109635692B (en) | Scene re-identification method based on ultrasonic sensor | |
| CN113251962B (en) | Ultrasonic parking space compensation system based on machine learning | |
| JP4618506B2 (en) | Object recognition device | |
| CN115453570A (en) | Multi-feature fusion mining area dust filtering method | |
| CN112859054B (en) | Automatic detection system and detection method for external parameters of vehicle-mounted multi-line laser radar |
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 | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20211224 |