CN116385471A - Laser contour line extraction method based on directional region growth - Google Patents

Laser contour line extraction method based on directional region growth Download PDF

Info

Publication number
CN116385471A
CN116385471A CN202310644666.XA CN202310644666A CN116385471A CN 116385471 A CN116385471 A CN 116385471A CN 202310644666 A CN202310644666 A CN 202310644666A CN 116385471 A CN116385471 A CN 116385471A
Authority
CN
China
Prior art keywords
column
laser
pixel
sampling
interval
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
Application number
CN202310644666.XA
Other languages
Chinese (zh)
Other versions
CN116385471B (en
Inventor
许明珠
杜萍
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhongke Weizhi Technology Co ltd
Original Assignee
Zhongke Weizhi Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Zhongke Weizhi Technology Co ltd filed Critical Zhongke Weizhi Technology Co ltd
Priority to CN202310644666.XA priority Critical patent/CN116385471B/en
Publication of CN116385471A publication Critical patent/CN116385471A/en
Application granted granted Critical
Publication of CN116385471B publication Critical patent/CN116385471B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/13—Edge detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/66—Analysis of geometric attributes of image moments or centre of gravity
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Abstract

The application discloses a laser contour line extraction method based on directional region growth, which relates to the field of image data processing, and comprises the steps of starting from a pixel point at the interval center of a laser interval in an extracted target sampling column, sequentially and directionally searching effective regions to two sides, and extracting a laser line segment formed by connecting the pixel points at the interval center of the laser interval in the target sampling column and the adjacent pixel columns at the two sides; and determining an updated target sampling column and a sampling area thereof according to the line segment end points of the extracted laser line segments, continuously extracting, and finally performing post-processing operation on all the laser line segments to extract and obtain a complete laser contour line in the laser image. The pixel point searching process of the method is more guiding, so that searching efficiency can be improved, and the laser contour line can be obtained through rapid extraction.

Description

Laser contour line extraction method based on directional region growth
Technical Field
The application relates to the field of image data processing, in particular to a laser contour line extraction method based on directional region growth.
Background
Along with the development of production technology, the automation requirement of an industrial production line is higher, and the line laser contour sensor is widely applied to the non-contact acquisition of the contour information of an object, is the basis of the three-dimensional reconstruction technology of the surface of the object, and provides basis for the detection, identification, classification and the like of the object.
The line laser contour sensor emits laser to the object and receives diffuse reflection light on the surface of the object, the diffuse reflection light is imaged on the line laser contour sensor to obtain a laser image containing a laser pattern, and the laser contour line of the laser pattern is extracted from the laser image, so that the contour information of the object can be accurately obtained. Because the line laser contour sensor emits a laser beam to the object, only one laser pattern exists in the obtained laser image, but the laser pattern in the laser image is often subjected to conditions such as segmentation and bending due to factors such as the surface morphology of the object, the environmental interference and the acquisition error, so that the laser contour is difficult to accurately and completely extract from the laser pattern, and the identification efficiency of contour information of the object is directly influenced.
Because of various interferences and errors, it is difficult to extract the laser contour line from the laser image only by the threshold segmentation method, so various algorithms for extracting the laser contour line are developed at present, some even machine learning and neural networks are introduced, but the data processing capacity of the methods is often larger, and the time for extracting the laser contour line is longer.
Disclosure of Invention
The applicant provides a laser contour line extraction method based on directional region growth aiming at the problems and the technical requirements, and the technical scheme of the application is as follows:
The laser contour line extraction method based on directional region growth comprises the following steps:
initializing a pixel column in a laser image as a target sampling column, and initializing sampling areas in the target sampling column, wherein the sampling areas in each pixel column cover a plurality of continuous pixel points in the pixel column;
sampling pixel points in a target sampling column from an initialized sampling region of the target sampling column, extracting a laser interval in the target sampling column, wherein the laser interval in each pixel column comprises pixel points with gray values reaching a gray value threshold value in the pixel column;
sequentially determining sampling areas of adjacent pixel columns at two sides of a target sampling column from a pixel point at the interval center of a laser interval in the target sampling column, and sampling the pixel points in the pixel columns from the sampling area of each pixel column until a laser line segment formed by connecting the pixel points at the interval center of the laser interval in the target sampling column and the adjacent pixel columns at two sides thereof is extracted;
determining an updated target sampling column and a sampling region thereof according to line segment endpoints of the laser line segments, and executing the step of sampling pixel points in the target sampling column from the initialized sampling region of the target sampling column again until all the laser line segments are extracted when all the pixel columns in the laser image are traversed;
And carrying out post-processing operation on all the laser line segments, and extracting to obtain the laser contour lines in the laser image.
The further technical scheme is that for any one pixel column, the method for sampling the pixel points in the pixel column from the sampling area of the pixel column to extract the laser interval in the pixel column comprises the following steps:
sampling pixel points in a sampling area of the pixel column;
when the distance between a candidate interval formed by the pixel points with gray values reaching a gray value threshold in the sampling area of the pixel column and the pixel points at the boundaries of the two sides of the sampling area along the column direction reaches a distance threshold, directly taking the candidate interval as a laser interval in the pixel column;
when there is no pixel point in the sampling area of the pixel column, the gray value of which reaches the gray value threshold value, or when the distance between a candidate interval formed by the pixel points in the sampling area of the pixel column, the gray value of which reaches the gray value threshold value, and the pixel point at the boundary of at least one side of the sampling area along the column direction does not reach the distance threshold value, the sampling area is moved along the column direction and the step of sampling the pixel points in the sampling area of the pixel column is executed again;
and de-overlapping and merging the pixel points contained in the candidate interval obtained by sampling and extracting in the plurality of sampling areas to obtain the laser interval in the pixel column.
The further technical scheme is that moving the sampling area along the column direction comprises:
moving the sampling area to a target moving side along the column direction by a preset number of pixel points to update the sampling area, wherein the sampling area at least overlaps and covers one pixel point before and after the moving updating;
when no pixel point with gray value reaching gray value threshold exists in the sampling area before the movement update, and the boundary of one side of the sampling area before the movement update is positioned at the boundary of the pixel column, determining one side far away from the boundary of the pixel column as a target movement side;
when no pixel point with gray value reaching gray value threshold exists in the sampling area before the movement update and the boundaries of two sides of the sampling area before the movement update are not positioned at the boundary of the pixel column, determining any side along the column direction as a target movement side;
when a candidate interval exists in the sampling area before the movement update, determining one side, which does not reach a distance threshold value, between the sampling area before the movement update and the extracted candidate interval as a target movement side.
The further technical scheme is that the sampling areas of the adjacent pixel columns at two sides of the target sampling column are sequentially determined, wherein the sampling areas comprise the ith pixel column which is sequentially arranged along the row direction at the same side of the target sampling column from the target sampling column:
Determining the central pixel point of a sampling area of an ith pixel column and the pixel point of the interval center of a laser interval in a candidate end point column in the same row, wherein i is an integer with a starting value of 1; when i=1, the candidate end point column of the ith pixel column is the target sampling column; when i is more than or equal to 2, the candidate endpoint column is a target sampling column and the last pixel column with the laser interval determined in the previous i-1 pixel columns.
The further technical scheme is that the extracting the laser line segment connected by the pixel points at the interval center of the laser interval in the target sampling column and the adjacent pixel columns at two sides thereof comprises the following steps:
sampling the pixel points in the ith pixel column from the sampling area of the ith pixel column, and extracting a reference interval formed by the pixel points with gray values reaching a gray value threshold value in the ith pixel column;
when the reference interval in the ith pixel column has a communication relation with the laser interval in the candidate end point column of the ith pixel column, taking the reference interval in the ith pixel column as the laser interval of the ith pixel column, taking the ith pixel column as the updated candidate end point column, enabling i=i+1, and executing the step of sampling the pixel points in the ith pixel column from the sampling area of the ith pixel column again;
When the reference interval is determined to be absent in the ith pixel column or the reference interval in the ith pixel column is determined to be not in communication with the laser interval in the candidate terminal point column of the ith pixel column, determining that the laser interval is absent in the ith pixel column;
when it is determined that the laser interval does not exist in the ith pixel column and the number of columns of the pixel columns at intervals between the ith pixel column and the candidate end point columns does not reach the column number threshold, keeping the current candidate end point column unchanged, enabling i=i+1, and executing the step of determining the center pixel point of the sampling area of the ith pixel column and the pixel point of the interval center of the laser interval in the candidate end point column in the same row again;
when it is determined that the laser section does not exist in the ith pixel column and the column number of the pixel column at the interval between the ith pixel column and the candidate end point column reaches the column number threshold, taking the pixel point at the section center of the laser section in the candidate end point column as a line segment end point at one side of the laser line segment.
The further technical scheme is that the laser contour line extraction method based on directional region growth further comprises the following steps:
when it is determined that the reference section in the ith pixel column has a communication relationship with the laser section in the candidate end point column of the ith pixel column and the candidate end point column is not the ith-1 th pixel column, adding a pixel point with a gray value reaching a gray value threshold between the laser section of the ith pixel column and the laser section of the candidate end point column of the ith pixel column to form a communication domain.
The further technical scheme is that the laser line segment formed by connecting the pixel points at the interval center of the laser interval in the target sampling column and the adjacent pixel columns at two sides thereof is extracted and obtained further comprises:
and calculating the average gray value of all pixel points in the connected domain formed by the laser intervals in the target sampling column and the adjacent pixel columns at the two sides of the target sampling column, extracting a laser line segment formed by connecting the pixel points in the interval center of the laser intervals in the target sampling column and the adjacent pixel columns at the two sides of the target sampling column when the calculated average gray value reaches a line segment gray threshold, otherwise discarding the connected domain formed by the laser intervals in the target sampling column and the adjacent pixel columns at the two sides of the target sampling column.
The further technical scheme is that an updated target sampling column and a sampling area thereof are determined according to line segment endpoints of a laser line segment:
and taking a pixel column with a plurality of columns at intervals of a line segment end point of the laser line segment at a side to be searched as an updated target sampling column, and initializing a sampling area in the updated target sampling column, wherein the side to be searched is one side of the line segment end point of the laser line segment far away from the other line segment end point.
The further technical scheme is that the post-processing operation of all laser line segments comprises the following steps:
When the line segment distance of the two laser line segments does not exceed the line segment distance threshold value, adding pixel points with gray values reaching the gray value threshold value between the line segment endpoints of the two laser line segments, which are close to each other, through an interpolation method, so that the line segment endpoints of the two laser line segments are connected.
The further technical scheme is that the line segment distance threshold comprises a first distance threshold and a second distance threshold, and determining whether the line segment distance of the two laser line segments exceeds the line segment distance threshold comprises:
when the distance between the adjacent line segment endpoints of the two laser line segments in the row direction reaches a first distance threshold, or the distance between the adjacent line segment endpoints of the two laser line segments in the column direction reaches a second distance threshold, determining that the line segment distance of the two laser line segments exceeds the line segment distance threshold, otherwise, determining that the line segment distance of the two laser line segments does not exceed the line segment distance threshold.
The beneficial technical effects of this application are:
the method takes a pixel point at the center of a section of a laser section in a target sampling column as a growth seed, searches an effective area towards pixel columns at two sides, and takes a central pixel point of a sampling region of an ith pixel column and a pixel point at the center of the section of the laser section in a candidate end point column to be in the same row in the process of directionally searching the pixel columns at two sides.
In addition, based on the method for searching the pixel columns adjacent to the two sides from the target sampling column, after one laser line segment is extracted, a plurality of pixel columns which are arranged at intervals from the line segment end points of the laser line segment at the side to be searched are used as updated target sampling columns, namely, after the current laser line segment is searched, a plurality of pixel columns can be directly skipped to search the next laser line segment without fixed column-by-column searching, and the skipped laser intervals in the plurality of pixel columns can be searched and supplemented in the process of searching the next pixel columns adjacent to the two sides from the target sampling column, so that the searching efficiency can be further improved on the basis of ensuring the searching accuracy, and the laser line is rapidly extracted.
According to the laser contour line extraction method, no matter in the process of searching and extracting one laser line segment or in the process of carrying out post-processing operation on a plurality of laser line segments, pixel points with gray values reaching a gray value threshold value are added among pixel points meeting requirements to form a connected domain, so that the pixel points missing due to external interference and errors are compensated, and the laser contour line extraction method is favorable for extracting more complete, more coherent and more accurate laser contour lines.
Drawings
Fig. 1 is a method flow diagram of a laser profile extraction method based on directional region growth in one embodiment of the present application.
Fig. 2 is a flowchart of a method for extracting a laser interval in any one pixel column according to an embodiment of the present application.
FIG. 3 is a schematic diagram of a target sample column initialized in one example of the present application and a sample area of the target sample column.
Fig. 4 is a schematic diagram of extracting laser intervals from an ith pixel column sequentially arranged on either side of a target sample column in one embodiment of the present application.
Fig. 5 is a schematic diagram of a traversal search to the left starting from the target sample column, based on the example of fig. 3.
Fig. 6 is a schematic diagram of extracting laser intervals from an ith pixel column sequentially arranged on either side of a target sample column according to another embodiment of the present application.
Fig. 7 is a schematic diagram of the pixel points that continue the traversal search to the left of the target sample column and add gray values between the laser intervals of two pixel columns to reach the gray value threshold based on the example of fig. 5.
Fig. 8 is a schematic illustration of a laser image obtained in one example of the present application.
Fig. 9 is a schematic diagram of a laser profile extracted from the laser image shown in fig. 8 according to a conventional laser profile extraction method.
Fig. 10 is a schematic view of a laser profile extracted from the laser image shown in fig. 8 according to the laser profile extraction method of the present application.
Detailed Description
The following describes the embodiments of the present application further with reference to the accompanying drawings.
The application discloses a laser contour line extraction method based on directional region growth, please refer to a flowchart shown in fig. 1, the laser contour line extraction method comprises the following steps:
step 1, initializing a pixel column in a laser image as a target sampling column, and initializing a sampling region in the target sampling column. In the application, the sampling area in each pixel column covers a plurality of continuous pixel points in the pixel column, the area size of the sampling area is fixed, namely the number of the pixel points covered by the sampling area is always unchanged, but the number of the pixel points covered by the sampling area can be adjusted in a self-defined manner.
In one embodiment, the initialized target sample column is the leftmost or rightmost pixel column of the laser image, and the sampling area in the initialized target sample column covers a number of pixels starting from the first row in the target sample column.
And 2, starting to sample the pixel points in the target sampling column from the initialized sampling area of the target sampling column, and extracting the laser interval in the target sampling column.
In the application, the extracted laser interval in each pixel column contains the pixel point in the pixel column, the gray value of which reaches the gray value threshold, the pixel point with the gray value reaching the gray value threshold is the pixel point belonging to the laser pattern, the gray value threshold is a self-defined value, the gray value threshold should ensure that the laser pattern and the background image can be separated, and the gray value of the pixel point at the edge of the laser pattern can be generally selected as the gray value threshold in actual operation.
In one embodiment, the method for extracting the laser interval in any one pixel column includes the following steps, please refer to the flowchart shown in fig. 2:
(1) And sampling the pixel points in the current sampling area of the pixel column, and determining the pixel points in the current sampling area of the pixel column, wherein the gray values of the pixel points possibly exist in the sampling area and reach the gray value threshold, or the gray values of all the pixel points in the sampling area do not reach the gray value threshold.
(2) When no pixel point with the gray value reaching the gray value threshold exists in the sampling area of the pixel column, the pixel point is moved along the column direction to update the sampling area, and the step (1) is executed again for the updated sampling area to sample.
(3) When the pixel points with the gray values reaching the gray value threshold value exist in the current sampling area of the pixel column, the pixel points with the gray values reaching the gray value threshold value in the sampling area are defined to form candidate intervals in the sampling area. Detecting whether the distance between a candidate interval in the current sampling region of the pixel column and pixel points at two side boundaries of the sampling region along the column direction reaches a distance threshold value or not, wherein the distance threshold value is a self-defined value.
(4) When the distance between the candidate interval in the current sampling area of the pixel column and the pixel points at the two side boundaries of the sampling area along the column direction reaches the distance threshold value, the candidate interval extracted is indicated to be positioned in the area with the center in the sampling area, and the candidate interval extracted at the moment can be considered to be complete, and the candidate interval extracted in the sampling area is directly taken as the laser interval in the pixel column.
(5) When the distance between a candidate interval formed by the pixel points with gray values reaching the gray value threshold value in the sampling area of the pixel column and the pixel points at the boundary of at least one side of the sampling area along the column direction does not reach the distance threshold value, the candidate interval is indicated to be close to the boundary of the sampling area, the extracted candidate interval has a high probability of being incomplete at the moment, the candidate interval moves along the column direction to update the sampling area, and the step (1) is executed again for the updated sampling area to sample.
In both the cases of the above step (2) and step (5), it is necessary to move along the column direction to update the sampling area, including: and moving the sampling area to a target moving side by a preset number of pixel points along the column direction to update the sampling area, wherein the sampling area is overlapped and covered with at least one pixel point before and after the moving and updating. The preset number of the pixel points moving along the column direction can be set in a self-defined mode, the preset number can be 1 to realize sequential sampling, or the preset number can be larger than 1 to realize interval sampling, for example, 3-5 pixel points can be moved. However, no matter how many the predetermined number is set, it is required to ensure that the predetermined number of moving pixels cannot exceed the number of pixels covered by a single sampling area, that is, it is required to ensure that the sampling area repeatedly covers at least one pixel before and after the movement update. The sampling area covers different pixels in the same pixel column before and after the moving update, but the number of the covered pixels is the same.
In the case of the above step (2), since there is no pixel point in the sampling area before the movement update where the gray value reaches the gray value threshold, two cases are included in determining the target movement side: (a) When a boundary of one side of the sampling area before the movement update is positioned at the boundary of the pixel column, determining the side far from the boundary of the pixel column as a target movement side: when the sampling area before the movement update is located at the upper boundary of the pixel column, the target movement side is determined to be the side where the lower boundary of the sampling area is located. When the sampling area before the movement update is located at the lower boundary of the pixel column, the target movement side is determined to be the side where the upper boundary of the sampling area is located. (b) When neither side boundary of the sampling area before the movement update is located at the boundary of the pixel column, either side along the column direction is determined as the target movement side. That is, when the sampling region before the movement update is located neither at the upper boundary nor at the lower boundary of the pixel column, the side on which the upper boundary or the lower boundary of the sampling region is located is randomly selected as the target movement side.
When the steps (1) and (2) are repeatedly executed, pixel points with gray values reaching a gray value threshold value are not extracted in the sampling area all the time, and the sampling area after the movement update is sequentially moved along the target movement side to reach the boundary of the pixel column, determining that no laser interval exists in the pixel column.
In the case of the step (5), if a candidate segment exists in the sampling region before the movement update, it is determined that the side of the sampling region before the movement update and the extracted candidate segment, which does not reach the distance threshold, is the target movement side. In this case, the target moving side is the side on which the upper boundary or the lower boundary of the sampling region is located. In this case, in step (5), since the number of pixels covered by the sampling region is significantly greater than the number of pixels included in the laser section in one pixel column, the distance threshold is not reached between the boundary of the sampling region and at most only one side of the candidate section in the sampling region, and the distance threshold is not reached between the boundaries of both sides.
When a candidate section at the boundary of the sampling area is extracted in one sampling area of the pixel column, after the sampling area is updated by moving, the candidate section is extracted again in the sampling area which is updated by moving, and overlapping pixel points exist between the candidate sections which are extracted before and after the sampling area is updated by moving, so that the pixel points contained in the candidate sections which are extracted by sampling in a plurality of sampling areas are subjected to de-overlapping and combined, and the laser section in the pixel column is obtained. The candidate regions extracted from different sampling regions are de-overlapped and are common techniques in the field of image processing, and detailed descriptions of the application are omitted.
It should be noted that, because the number of pixel points covered by the sampling area is far greater than that of the pixel points included in the laser interval in one pixel column, after the candidate interval located at the boundary of the sampling area is extracted in one sampling area of the pixel column, the sampling area is generally updated once by moving, so that the complete laser interval can be removed and obtained, thereby reducing the number of times of movement required and improving the efficiency.
For example, in one example, as shown in fig. 3, squares in a hatched portion in fig. 3 represent pixels in the laser image where the gray value reaches the gray value threshold, and the remaining white squares represent pixels in the laser image where the gray value does not reach the gray value threshold. In fig. 3, the gradation value of each pixel is shown in a binarized form for simplicity, but the gradation value of the pixel is more valued in actual use. In this example, the pixel column of column 1 on the leftmost side of the laser image is initialized as a target sampling column, and the sampling area of the initialized target sampling column covers 7 pixel points of 1 st row to 7 th row as indicated by an arrow in the figure, as shown by a solid line box in fig. 3. Then in the example of fig. 3, for the target sample column, the method of sequentially traversing samples from the initialized sample area to extract the laser interval comprises:
7 pixel points in the initialized sampling area are sampled firstly, and the pixel points, in which the gray value reaches the gray value threshold, are not included in the initialized sampling area are determined. At this time, the sampling area is located at the upper boundary of the target sampling column, so that the target moving side is the side where the lower boundary is located, the sampling area is moved to the lower side by 2 pixel points, and 7 pixel points from the 3 rd row to the 9 th row in the target sampling column are covered by the sampling area after the movement update.
7 pixel points in the current sampling area are sampled, and two pixel points with gray values reaching a gray value threshold value in the current sampling area are determined. However, at this time, the two gray values reach the gray value threshold value to form a candidate interval, and the pixel point of the lower boundary of the current sampling area is not separated by more than one pixel point, so the current sampling area is continuously moved to the lower side by 2 pixel points to serve as a sampling area after movement update, and 7 pixel points from the 5 th line to the 11 th line in the target sampling column are covered by the sampling area after movement update.
7 pixel points in the current sampling area are sampled, and at least one pixel point is determined, wherein two pixel points with gray values reaching a gray value threshold value exist in the current sampling area, and at the moment, a candidate interval formed by the two gray values reaching the gray value threshold value and the pixel point of the lower side boundary of the current sampling area are separated by at least one pixel point. And finishing traversing, carrying out de-duplication combination on the candidate intervals extracted twice, and finally extracting the laser interval in the target sampling column to comprise two pixel points of the 8 th row and the 9 th row in the target sampling column.
More specifically, when the leftmost pixel column of the initialized laser image is taken as a target sampling column and no laser interval is extracted in the initialized target sampling column, updating the right pixel column of the current target sampling column as a target sampling column, and repeating the above process again until the laser interval in the target sampling column is extracted. Similarly, when the rightmost pixel column of the initialized laser image is taken as the target sampling column and no laser interval is extracted in the initialized target sampling column, the left pixel column of the current target sampling column is updated as the target sampling column, and the process is repeated again until the laser interval in the target sampling column is extracted.
For example, if the pixel column of the leftmost column 1 of the laser image is initialized to be the target sampling column, but the pixel column of the left column 2 of the laser image is updated to be the target sampling column when the laser section is not extracted from the pixel columns of the column 1, if the laser section is still not extracted from the pixel columns of the column 2, the pixel column of the left column 3 of the laser image is continuously updated to be the target sampling column, and if the laser section is extracted from the pixel column of the column 3, the following step 3 is continuously executed with the pixel column of the column 3 as the target sampling column.
And 3, sequentially determining sampling areas of adjacent pixel columns at two sides of the target sampling column from a pixel point at the interval center of the laser interval in the target sampling column, and sampling the pixel points in the pixel column from the sampling area of each pixel column until a laser line segment formed by connecting the pixel points at the interval center of the laser interval in the target sampling column and the adjacent pixel columns at two sides thereof is extracted.
The pixel point at the interval center of the laser interval of one pixel column can be determined by a gray-scale gravity center method. After the laser interval in the target sampling column is determined, the pixel point in the interval center of the laser interval of the target sampling column is used as a growth seed, the laser line segment is searched from the growth seed to two sides along the distribution direction of the laser line segment, and the laser line segment is obtained through directional region growth and extraction. When the target sampling column is positioned at the leftmost side of the laser image, searching is performed to each pixel column at the right side of the target sampling column. When the target sampling column is located at the rightmost side of the laser image, a search is performed for each pixel column to the left of the target sampling column. When the target sampling column is located in the pixel columns except the leftmost side and the rightmost side in the laser image, searching is performed to each pixel column on both sides of the target sampling column, after the searching of the pixel column on one side of the target sampling column is completed from the pixel point of the section center of the laser section in the target sampling column, searching is performed to the pixel column on the other side of the target sampling column from the pixel point of the section center of the laser section in the target sampling column, the searching sequence in the left-right direction can be customized, and the searching methods on both sides are the same. In one embodiment, referring to the flowchart shown in fig. 4, the method for searching either side of the target sample column includes:
(1) Initializing a parameter i=1, initializing a target sampling column as a candidate endpoint column, and setting the parameter i as an integer.
(2) For the ith pixel column sequentially arranged in the row direction on the side of the target sample column from the target sample column, the center pixel point of the sampling region of the ith pixel column is determined to be in the same row as the pixel point at the section center of the laser section in the candidate end point column.
When initializing i=1, the first pixel column is the pixel column directly adjacent to the target sampling column, the candidate end point column when initializing i=1 is the target sampling column, and the central pixel point of the sampling area of the first pixel column is the adjacent pixel point of the pixel point in the interval center of the laser interval of the target sampling column. When i is equal to or greater than 2, the candidate endpoint column is a target sampling column and the last pixel column of the first i-1 pixel columns on the side, wherein the laser interval is determined to exist.
Since the number of the pixel points covered by the sampling area is fixed, after the position of the central pixel point of the sampling area of the ith pixel column is determined, the pixel point covered by the sampling area of the ith pixel column can be determined. For example, in the example of fig. 3, assume that when it is determined that the laser interval in the target sampling column includes two pixels of the 8 th row and the 9 th row, and the pixel of the 8 th row is selected as the pixel of the interval center of the laser interval, a process of performing a traversal search to the pixel column on the right side of the target sampling column is taken as an example: the immediately adjacent pixel column located on the right side of the target sampling column is the first pixel column, the central pixel point of the sampling area of the first pixel column is the pixel point of the 8 th row, and if the condition that the sampling area covers 7 pixel points is combined, the sampling area of the first pixel column (column 2) covers 7 pixel points from the 5 th row to the 11 th row in the pixel column, and the solid line frame in the first pixel column in fig. 5 shows the sampling area in the first pixel column.
(3) Sampling the pixel points in the ith pixel column from the sampling area of the ith pixel column, and extracting a reference section formed by the pixel points of which the gray level value in the ith pixel column reaches the gray level value threshold value. The method for extracting the reference interval in the ith pixel column in this step is similar to the method for extracting the laser interval in the target sample column in the above step 2, and this embodiment will not be repeated.
(4) And when the reference interval does not exist in the ith pixel column, namely, the pixel point with the gray value reaching the gray value threshold value does not exist in the ith pixel column, determining that the laser interval does not exist in the ith pixel column.
(5) When it is determined that the reference interval exists in the ith pixel column and the reference interval in the ith pixel column has a communication relation with the laser interval in the candidate terminal point column, taking the reference interval extracted from the ith pixel column as the laser interval of the ith pixel column, taking the ith pixel column as the updated candidate terminal point column, enabling i=i+1, and executing the steps (2) and (3) again, and performing traversal search on the next pixel column sequentially arranged in the lateral direction.
(6) When it is determined that the reference section exists in the ith pixel column, but the reference section in the ith pixel column does not have a communication relation with the laser section in the candidate end point column, it is determined that the laser section does not exist in the ith pixel column.
For example, in the example shown in fig. 5, it may be determined that a reference interval exists in the first pixel column by sampling, where the reference interval includes two pixels of the 8 th row and the 9 th row in the first pixel column, and since the reference interval in the first pixel column and the laser interval in the target sampling column have directly adjacent pixels, the two pixels have a connection relationship, and it is determined that the laser interval in the first pixel column includes the two pixels of the 8 th row and the 9 th row.
The first pixel column is updated to serve as a candidate end point column, traversing to the right is continued, the column 3 serves as a second pixel column, the sampling area of the second pixel column can be determined to cover 7 pixel points from the 5 th row to the 11 th row in the pixel column by adopting the same method, and then the pixel points in the second pixel column are sampled from the sampling area in the second pixel column.
In the steps (5) and (6), when the distance between the reference section in the ith pixel column and the closest two pixels in the laser sections in the candidate end point columns does not reach the first distance threshold in the row direction, and the distance in the column direction does not reach the second distance threshold, that is, the lateral distance in the row direction is closer and the longitudinal distance in the column direction is closer, it is determined that the reference section in the ith pixel column has a communication relationship with the laser sections in the candidate end point columns, otherwise, it is determined that the reference section in the ith pixel column does not have a communication relationship with the laser sections in the candidate end point columns.
In one embodiment, for the ith pixel column sequentially arranged along the row direction on the same side of the target sampling column from the target sampling column, the corresponding candidate end point column is the ith-1 th pixel column, that is, the previous adjacent pixel column of the ith pixel column in the traversal search order is used as the candidate end point column. In this embodiment, when it is determined in the steps (4) and (6) that the laser section does not exist in the ith pixel column, the search is directly ended, the traversal search for the side of the target sample column is completed, and the candidate end point column at this time is directly taken as the pixel column at the end point of the one line segment of the laser line segment. For example, in the example of fig. 5, the search is sequentially traversed by the method until i=15, and the column 16 is sampled as the fifteenth pixel column, and finally, it is determined that the fifteenth pixel column does not have the reference section, so that the traversing search of the pixel column on the right side of the target sampling column can be directly completed, the fourteenth pixel column is directly used as the pixel column at the end point of the right side line segment of the extracted laser line segment, and the pixel point covered by the laser section in the fourteenth pixel column is formed as the pixel point at the end point of the right side line segment of the laser line segment.
However, in practical application, due to external interference and sampling errors, partial pixel points in the complete laser contour line are easy to be lost, so that the method provided by the embodiment is easy to cause incomplete finally extracted laser contour line. Thus, in another embodiment, for the ith pixel column sequentially arranged along the row direction on the same side of the target sampling column from the target sampling column, a number of columns of pixel columns may be spaced between the corresponding candidate end point column and the ith pixel column. In this embodiment, please refer to the flowchart shown in fig. 6:
in the above steps (4) and (6), when it is determined that the laser section does not exist in the ith pixel column, it is first detected whether the number of columns of the pixel columns spaced between the ith pixel column and the candidate end point column reaches the column number threshold. And (3) when the fact that the laser interval does not exist in the ith pixel column and the column number of the pixel column at the interval between the ith pixel column and the candidate end point column does not reach the column number threshold value is determined, keeping the current candidate end point column unchanged, enabling i=i+1, and executing the step of determining the same row as the pixel point at the center of the interval center of the laser interval in the ith pixel column and the pixel point at the center of the interval center of the laser interval in the candidate end point column again, namely executing the steps (2) - (6) to continue traversing searching. When it is determined that the laser section does not exist in the ith pixel column and the column number of the pixel column at the interval between the ith pixel column and the candidate end point column reaches the column number threshold, it is determined that the traversing search on the current side is completed, the candidate end point column at this time is taken as the pixel column at the end point of the line segment on one side of the laser line segment, and the pixel point at the interval center of the laser section in the candidate end point column is taken as the end point of the line segment on one side of the laser line segment.
In the step (5), after the laser interval in the ith pixel column is extracted, if the candidate end point column at this time is the ith-1 th pixel column, and the laser interval in the ith pixel column and the laser interval in the ith-1 th pixel column form a connected domain, the ith pixel column is directly used as the updated candidate end point column, so that i=i+1 is performed again, and steps (2) and (3) are performed again, and the next pixel column sequentially arranged in the lateral direction continues to traverse the search. If the candidate end point column is not the i-1 th pixel column, a plurality of pixel columns are further spaced between the laser interval in the i-th pixel column and the laser interval in the candidate end point column, at this time, first, a pixel point with a gray value reaching a gray value threshold is added between the laser interval of the i-th pixel column and the laser interval of the candidate end point column, so that a connected domain is formed between the laser interval of the i-th pixel column and the laser interval of the candidate end point column, and the pixel point with the gray value reaching the gray value threshold can be added by adopting an interpolation method. And then taking the ith pixel column as an updated candidate terminal column, enabling i=i+1, executing the steps (2) and (3) again, and continuing traversing searching towards the next pixel column which is orderly arranged towards the side direction.
For example, based on the example of fig. 5, assume a column number threshold of 3. In the process of traversing the search from the target sampling column to the pixel column at the right side in sequence, determining the laser interval in each pixel column in sequence and taking each pixel column as the candidate end point column after updating in sequence until traversing to determine the laser interval in the fourteenth pixel column and taking the fourteenth pixel column as the candidate end point column after updating, traversing to search the fifteenth pixel column and determining that the laser interval does not exist in the fifteenth pixel column, wherein the fifteenth pixel column is directly adjacent to the candidate end point column, the column number of the pixel column at the interval does not reach the column number threshold, so that the search is not directly ended, but the fourteenth pixel column is kept as the candidate end point column, and the 17 th column at the right side is continuously searched as the sixteenth pixel column.
The sample area of the sixteenth pixel column determined by the above method is shown as a solid line box in the sixteenth pixel column in fig. 7. Sampling is started from the sampling area on the sixteenth pixel column, and the pixel points of which the reference interval in the sixteenth pixel column comprises the 12 th row can be extracted. At this time, it is determined that the two closest pixel points in the reference section in the sixteenth pixel column and the laser section in the candidate end point column are respectively: and the pixel point of the 9 th row of the 15 th column and the pixel point of the 12 th row of the 17 th column, if the distance between the two pixel points in the row direction does not reach the first distance threshold and the distance between the two pixel points in the column direction does not reach the second distance threshold, the pixel point of the 12 th row in the sixteenth pixel column is taken as a laser interval thereof, and then the pixel point with the gray value reaching the gray value threshold is added between the laser interval of the sixteenth pixel column and the laser interval of the fourteenth pixel column, as shown in the schematic diagram of fig. 7. Then, the sixteenth pixel column is used as the updated candidate terminal column, and the search is traversed again for the pixel column at the right side. As can be seen from the example of fig. 7, only intermittent line segments can be extracted according to the conventional method, but complete laser line segments can be extracted according to the method of the present application.
The laser image in the application only comprises one laser pattern, the laser pattern can be segmented, bent and has a certain width, but each pixel column in the laser image has at most one pixel point in the interval center of the laser interval. Therefore, in one embodiment, after the connected domain formed by the laser sections in the target sampling column and the adjacent pixel columns on both sides thereof is extracted by the above method, the centers of the sections in the connected domain are directly connected in sequence, so as to obtain a laser line segment.
In another embodiment, however, in order to reduce the error, first, an average gray value of all the pixels in the connected domain formed by the laser sections in the target sampling column and the adjacent pixel columns on both sides thereof is calculated, when the calculated average gray value reaches the line segment gray threshold, a laser line segment formed by the pixels in the center of the interval of the laser sections in the target sampling column and the adjacent pixel columns on both sides thereof is extracted, otherwise, the connected domain formed by the laser sections in the target sampling column and the adjacent pixel columns on both sides thereof is discarded.
And 4, determining an updated target sampling column and a sampling area thereof according to line segment endpoints of the laser line segments, and executing the step of sampling pixel points in the target sampling column from the sampling area of the target sampling column again until all the laser line segments are extracted when all the pixel columns in the laser image are traversed.
In one embodiment, a pixel column of a plurality of columns is spaced from one line end point of a laser line at a side to be searched to serve as an updated target sampling column, a sampling area in the updated target sampling column is initialized to cover a plurality of continuous pixel points in the updated target sampling column, and then the corresponding laser line is searched again according to the method provided by the application for the updated target sampling column, namely, the searching is traversed again from the step 2.
The side to be searched is one side of the line segment end point of the laser line segment far away from the other line segment end point. The number of pixel columns separated between the updated target sampling column and the pixel column where the line segment end point of the laser line segment is located can be set in a self-defined manner, and the pixel columns can be directly adjacent to each other or separated by a plurality of other pixel columns. It should be noted that, as will be understood by those skilled in the art, when one line end of the laser line is at the leftmost or rightmost side of the laser image, the side to be searched directly takes the opposite side.
For example, based on the example of fig. 7, in the process of traversing from the initialized target sampling column (i.e., column 1) to the right pixel column according to the method provided in the present application, finally, traversing to the column 27 in fig. 7 in turn according to the method provided in the present application extracts the laser section, then taking the column 27 as a candidate end point column, traversing to the columns 28, 29 and 30 in turn does not extract the laser section, when traversing to the 31 st column, determining that the number of pixel columns spaced from the column 27 as the candidate end point column at this time reaches the column number threshold 3, ending the traversing, taking the pixel point in the interval center of the laser section in the column 27 as the line segment end point of the laser line segment, at this time, searching for a laser line segment, and taking the side to be searched as the right side. The search may then continue from step 2 above with column 30 to the right of column 27 as the updated target sample column.
And 5, performing post-processing operation on all the laser line segments, and extracting to obtain the laser contour lines in the laser image. When only one laser line segment is extracted, the laser line segment is directly used as a laser contour line. When a plurality of laser line segments are extracted, for every two adjacent laser line segments:
it is detected whether the line segment distance of the two laser line segments exceeds a line segment distance threshold. In one embodiment, the line segment distance threshold comprises a first distance threshold and a second distance threshold, and when it is determined that the distance between the line segment endpoints of the two laser line segments in the row direction reaches the first distance threshold, or the distance between the line segment endpoints of the two laser line segments in the column direction reaches the second distance threshold, it is determined that the line segment distance between the two laser line segments exceeds the line segment distance threshold. The line segment distance threshold may be custom set.
When the line segment distance of the two laser line segments does not exceed the line segment distance threshold value, adding pixel points with gray values reaching the gray value threshold value between the line segment endpoints of the two laser line segments by interpolation, so that the pixel points contained in the two laser line segments are communicated into a communicated domain, and the two laser line segments are connected into one laser line segment.
When the line segment distance of the two laser line segments exceeds the line segment distance threshold, the two laser line segments are reserved as independent laser line segments.
The condition that the intermittent line segments are extracted due to the pixel point missing caused by external interference and sampling errors can be avoided through the post-processing operation, and the integrity and consistency of the extracted laser contour lines can be further improved. In an example, a laser image obtained in the embodiment is shown in fig. 8, a laser contour line extracted from the laser image in fig. 8 according to a conventional laser contour line extraction method is shown in fig. 9, a laser contour line extracted from the laser image in fig. 8 according to the laser contour line extraction method in the application is shown in fig. 10, and it can be seen from comparison between fig. 9 and fig. 10 that the integrity, consistency and accuracy of the laser contour line extracted by the laser contour line extraction method in the application are all higher. In addition, after the laser contour line is extracted according to the method, the processing such as denoising and fitting can be further performed.
In an application example, the laser contour line is extracted by using a traditional MSER algorithm, a Steger algorithm and the method of the application respectively for 100 images with the resolution of 1920×648, the average time consumption of extracting the laser contour line by using the method of the application is 0.90ms, the average time consumption of extracting the laser contour line by using the MSER algorithm is 5.45ms, the average time consumption of extracting the laser contour line by using the Steger algorithm is 2.61ms, and compared, the method of the application has less time consumption of extracting the laser contour line and high extraction efficiency.
The foregoing is merely a preferred embodiment of the present application, which is not limited to the above examples, it being understood that other modifications and variations, which may be directly derived or suggested by those skilled in the art, are deemed to be included within the scope of the present application without departing from the spirit and concept of the present application.

Claims (7)

1. The laser contour line extraction method based on the directional region growth is characterized by comprising the following steps of:
initializing a pixel column in a laser image as a target sampling column, and initializing sampling areas in the target sampling column, wherein the sampling areas in each pixel column cover a plurality of continuous pixel points in the pixel column;
sampling pixel points in the target sampling column from the initialized sampling region of the target sampling column, and extracting a laser interval in the target sampling column, wherein the laser interval in each pixel column comprises pixel points with gray values reaching a gray value threshold in the pixel column; for any one pixel column, the method for sampling the pixel points in the pixel column from the sampling area of the pixel column to extract the laser interval in the pixel column comprises the following steps: sampling pixel points in a sampling area of the pixel column, when a distance between a candidate interval formed by pixel points with gray values reaching the gray value threshold in the sampling area of the pixel column and the pixel points at two side boundaries of the sampling area along the column direction reaches a distance threshold, directly taking the candidate interval as a laser interval in the pixel column, when no pixel point with gray values reaching the gray value threshold in the sampling area of the pixel column or when a distance between the candidate interval formed by pixel points with gray values reaching the gray value threshold in the sampling area of the pixel column and the pixel points at least one side boundary of the sampling area along the column direction does not reach the distance threshold, moving the sampling area along the column direction and executing the step of sampling the pixel points in the sampling area of the pixel column again, de-overlapping the pixel points contained in the candidate intervals obtained by sampling and extracting in a plurality of sampling areas, and obtaining the laser interval in the pixel column;
Sequentially determining sampling areas of adjacent pixel columns at two sides of the target sampling column from a pixel point at the interval center of a laser interval in the target sampling column, and sampling the pixel points in the pixel columns from the sampling area of each pixel column until a laser line segment formed by connecting the pixel points at the interval center of the laser interval in the target sampling column and the adjacent pixel columns at two sides thereof is extracted, wherein the laser line segment comprises an ith pixel column which is sequentially arranged along the row direction at the same side of the target sampling column from the target sampling column: determining that a central pixel point of a sampling area of the ith pixel column and a pixel point of the interval center of a laser interval in a candidate end point column are positioned in the same row, wherein i is an integer with a starting value of 1; when i=1, the candidate end point column of the ith pixel column is the target sampling column; when i is more than or equal to 2, the candidate endpoint column is the pixel column of which the laser interval exists in the last one of the target sampling column and the previous i-1 pixel columns, and extracting the laser line segment formed by connecting the pixel points at the interval center of the laser interval in the target sampling column and the adjacent pixel columns at two sides of the target sampling column comprises: sampling the pixel points in the ith pixel column from the sampling area of the ith pixel column, extracting a reference interval formed by pixel points with gray values reaching a gray value threshold value in the ith pixel column, when the reference interval in the ith pixel column has a communication relation with a laser interval in a candidate end point column of the ith pixel column, determining the reference interval in the ith pixel column as a laser interval of the ith pixel column, taking the ith pixel column as an updated candidate end point column, enabling i=i+1 and executing the step of sampling pixel points in the ith pixel column from the sampling area of the ith pixel column again, when the reference interval is determined to be not present in the ith pixel column or when the reference interval in the ith pixel column is determined to have no communication relation with the laser interval in the candidate end point column, determining that the laser interval is not present in the ith pixel column, when the laser interval is determined to be present in the candidate end point column, and the laser interval is determined to be present in the ith pixel column, and when the laser interval is determined to be present in the candidate end point column reaches the laser interval of the ith pixel column, the step of the laser interval is determined to be not present in the ith pixel column, and when the laser interval is determined to be present in the candidate end point column, the step of the laser interval is not present in the ith pixel column, and the step of the laser interval is reached again, the step of sampling the pixel point is not in the ith pixel column is determined, and the laser interval is not present in the pixel column is determined, and the laser interval is not in the laser interval is determined, taking a pixel point at the interval center of the laser interval in the candidate end point row as a line segment end point at one side of the laser line segment;
Determining an updated target sampling column and a sampling region thereof according to line segment endpoints of the laser line segments, and executing the step of sampling pixel points in the target sampling column from the initialized sampling region of the target sampling column again until all the pixel columns in the laser image are traversed, and extracting to obtain all the laser line segments;
and carrying out post-processing operation on all the laser line segments, and extracting to obtain the laser contour lines in the laser image.
2. The laser profile extraction method based on directional region growing of claim 1, wherein the moving the sampling region along the column direction comprises:
moving the sampling area to a target moving side along the column direction by a preset number of pixel points to update the sampling area, wherein the sampling area at least overlaps one pixel point before and after the moving update;
when no pixel point with the gray value reaching the gray value threshold exists in the sampling area before the movement update, and the boundary of one side of the sampling area before the movement update is positioned at the boundary of the pixel column, determining one side far away from the boundary of the pixel column as the target movement side;
When no pixel point with the gray value reaching the gray value threshold exists in the sampling area before the movement update and the boundaries of the two sides of the sampling area before the movement update are not positioned at the boundary of the pixel column, determining any side along the column direction as the target movement side;
when a candidate interval exists in the sampling area before the movement update, determining one side, which does not reach the distance threshold, between the sampling area before the movement update and the extracted candidate interval as the target movement side.
3. The directional region growth based laser profile extraction method of claim 1, further comprising:
when it is determined that the reference section in the ith pixel column has a communication relation with the laser section in the candidate end point column of the ith pixel column, and the candidate end point column is not the ith-1 th pixel column, adding a pixel point with a gray value reaching a gray value threshold between the laser section of the ith pixel column and the laser section of the candidate end point column of the ith pixel column to form a communication domain.
4. The method for extracting a laser contour line based on directional region growing according to claim 1, wherein the extracting to obtain a laser line segment formed by connecting pixel points at the interval center of a laser interval in the target sampling column and adjacent pixel columns at two sides thereof further comprises:
And calculating the average gray value of all pixel points in the connected domain formed by the laser intervals in the target sampling column and the adjacent pixel columns at the two sides of the target sampling column, extracting a laser line segment formed by connecting the pixel points in the interval center of the laser intervals in the target sampling column and the adjacent pixel columns at the two sides of the target sampling column when the calculated average gray value reaches a line segment gray threshold, otherwise discarding the connected domain formed by the laser intervals in the target sampling column and the adjacent pixel columns at the two sides of the target sampling column.
5. The method for extracting a laser profile based on directional region growing according to claim 1, wherein the updated target sampling column and the sampling region thereof are determined according to line segment end points of the laser line segment:
and taking a pixel column with a plurality of columns at intervals of a line segment end point of the laser line segment at a side to be searched as an updated target sampling column, and initializing a sampling area in the updated target sampling column, wherein the side to be searched is one side of the line segment end point of the laser line segment far away from the other line segment end point.
6. The method of claim 1, wherein the post-processing all laser line segments comprises:
When the line segment distance of the two laser line segments does not exceed the line segment distance threshold value, adding pixel points with gray values reaching the gray value threshold value between the line segment endpoints of the two laser line segments, which are close to each other, through an interpolation method, so that the line segment endpoints of the two laser line segments are connected.
7. The method of claim 6, wherein the line segment distance threshold comprises a first distance threshold and a second distance threshold, and wherein determining whether the line segment distance of two laser line segments exceeds the line segment distance threshold comprises:
and when the distance between the adjacent line segment endpoints of the two laser line segments in the row direction reaches the first distance threshold, or the distance between the adjacent line segment endpoints of the two laser line segments in the column direction reaches the second distance threshold, determining that the line segment distance of the two laser line segments exceeds the line segment distance threshold, otherwise, determining that the line segment distance of the two laser line segments does not exceed the line segment distance threshold.
CN202310644666.XA 2023-06-02 2023-06-02 Laser contour line extraction method based on directional region growth Active CN116385471B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310644666.XA CN116385471B (en) 2023-06-02 2023-06-02 Laser contour line extraction method based on directional region growth

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310644666.XA CN116385471B (en) 2023-06-02 2023-06-02 Laser contour line extraction method based on directional region growth

Publications (2)

Publication Number Publication Date
CN116385471A true CN116385471A (en) 2023-07-04
CN116385471B CN116385471B (en) 2023-09-01

Family

ID=86975436

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310644666.XA Active CN116385471B (en) 2023-06-02 2023-06-02 Laser contour line extraction method based on directional region growth

Country Status (1)

Country Link
CN (1) CN116385471B (en)

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150269446A1 (en) * 2014-03-24 2015-09-24 Toyota Jidosha Kabushiki Kaisha Boundary detection apparatus and boundary detection method
CN108334879A (en) * 2018-02-11 2018-07-27 深圳市朗驰欣创科技股份有限公司 A kind of method for extracting region, system and terminal device
CN109074083A (en) * 2018-06-08 2018-12-21 珊口（深圳）智能科技有限公司 Control method for movement, mobile robot and computer storage medium
CN110717874A (en) * 2019-10-10 2020-01-21 徐庆 Image contour line smoothing method
CN111709981A (en) * 2020-06-22 2020-09-25 高小翎 Registration method of laser point cloud and analog image with characteristic line fusion
US20210065376A1 (en) * 2019-02-14 2021-03-04 China Institute Of Water Resources And Hydropower Research Region Merging Image Segmentation Algorithm Based on Boundary Extraction
CN113780110A (en) * 2021-08-25 2021-12-10 中国电子科技集团公司第三研究所 Method and device for detecting weak and small targets in image sequence in real time
US20220092314A1 (en) * 2020-09-21 2022-03-24 Nanjing University Of Aeronautics And Astronautics Road Recognition Method and System Based on Seed Point
CN115082699A (en) * 2022-07-28 2022-09-20 深圳市其域创新科技有限公司 Contour shape extraction method and device, electronic equipment and storage medium

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150269446A1 (en) * 2014-03-24 2015-09-24 Toyota Jidosha Kabushiki Kaisha Boundary detection apparatus and boundary detection method
CN108334879A (en) * 2018-02-11 2018-07-27 深圳市朗驰欣创科技股份有限公司 A kind of method for extracting region, system and terminal device
CN109074083A (en) * 2018-06-08 2018-12-21 珊口（深圳）智能科技有限公司 Control method for movement, mobile robot and computer storage medium
US20210065376A1 (en) * 2019-02-14 2021-03-04 China Institute Of Water Resources And Hydropower Research Region Merging Image Segmentation Algorithm Based on Boundary Extraction
CN110717874A (en) * 2019-10-10 2020-01-21 徐庆 Image contour line smoothing method
CN111709981A (en) * 2020-06-22 2020-09-25 高小翎 Registration method of laser point cloud and analog image with characteristic line fusion
US20220092314A1 (en) * 2020-09-21 2022-03-24 Nanjing University Of Aeronautics And Astronautics Road Recognition Method and System Based on Seed Point
CN113780110A (en) * 2021-08-25 2021-12-10 中国电子科技集团公司第三研究所 Method and device for detecting weak and small targets in image sequence in real time
CN115082699A (en) * 2022-07-28 2022-09-20 深圳市其域创新科技有限公司 Contour shape extraction method and device, electronic equipment and storage medium

Also Published As

Publication number Publication date
CN116385471B (en) 2023-09-01

Similar Documents

Publication Publication Date Title
CN107633192B (en) Bar code segmentation and reading method based on machine vision under complex background
CN108921813B (en) Unmanned aerial vehicle detection bridge structure crack identification method based on machine vision
JP6099479B2 (en) Crack detection method
WO2017016448A1 (en) Qr code feature detection method and system
CN111754422A (en) Visual edge inspection method based on EDLines and LSM
CN111354026A (en) Improved train image accurate registration method and system based on SURF (speeded up robust features)
CN115082462A (en) Method and system for detecting appearance quality of fluid conveying pipe
CN109190434B (en) Bar code recognition algorithm based on sub-pixel level corner detection
CN111968079B (en) Three-dimensional pavement crack extraction method based on local extremum of section and segmentation sparsity
CN111444773B (en) Image-based multi-target segmentation identification method and system
CN112883795A (en) Rapid and automatic table extraction method based on deep neural network
CN116385471B (en) Laser contour line extraction method based on directional region growth
CN112069852B (en) Low-quality two-dimensional code information extraction method and device based on run-length coding
CN113610052A (en) Tunnel water leakage automatic identification method based on deep learning
CN116842976A (en) DataMatrix two-dimensional code identification method and system based on coarse positioning of images
CN108256517B (en) Laser radar-based insulator string identification method
CN110097065B (en) Freeman chain code-based line detection method and terminal
CN111882543B (en) Cigarette filter stick counting method based on AA R2Unet and HMM
CN109784121B (en) Dot-peep DPM code identification method and device
CN115187744A (en) Cabinet identification method based on laser point cloud
CN114359378A (en) Method for positioning inspection robot of belt conveyor
JP4731295B2 (en) Velocity measuring device, method and program
CN110728723A (en) Tile map-oriented automatic road extraction method
Ponciano et al. Detection and isolation of switches in point clouds of the German railway network
CN113221898B (en) Automatic water level gauge reading 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