Rapid laser target detection method based on alpha-shapes
Through the fast laser target detection method based on α-shapes, non-uniform Archimedes helical scanning and improved adaptive threshold α-shapes algorithm are adopted, combined with the RANSAC algorithm and the FPGA platform, the problems of laser scanning time and slow data processing are solved, and the fast, real-time positioning and efficient assembly of laser targets are achieved.
Patent Information
- Application Number
- CN202510694128.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-28
- Publication Date
- 2025-08-19
AI Technical Summary
The existing laser scanning methods take a long time to scan laser targets, and the data processing is slow, so they cannot adapt to complex and changeable production environments, affecting production efficiency and accuracy.
The fast laser target detection method based on α-shapes is adopted, including non-uniform Archimedes helical scanning, improved adaptive threshold α-shapes algorithm and RANSAC algorithm, and combined with the FPGA platform for hardware acceleration to achieve efficient boundary point extraction and center coordinate calculation.
The scanning speed and data processing efficiency of laser targets are significantly improved, the calculation amount is reduced, the rapid and real-time positioning of laser targets is achieved, and the assembly efficiency and accuracy are improved.
Smart Images
Figure CN120506882A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of laser scanning measurement, in particular to a rapid laser target detection method based on α-shapes. Background Art
[0002] With the rapid development of the manufacturing industry towards informatization and intelligence, traditional assembly positioning technologies are gradually being replaced by more advanced automation and digital technologies. Traditional manual positioning and assembly methods are prone to human errors, which affect product quality. The introduction of laser scanning laser targets can achieve digital positioning and rapid assembly of small connectors, effectively reducing human errors and ensuring assembly quality. With the rapid development of high-end equipment manufacturing, higher quality and higher efficiency requirements are being put forward for the positioning and assembly of large parts. As a new type of digital measurement and guidance system, the laser projection system is used in the manufacturing industry for high-precision laser-assisted hole positioning, digital positioning of wire harnesses, and laser-assisted positioning in various assembly sites.
[0003] Assembly positioning still faces problems such as low installation efficiency and lack of flexibility. Usually, fast laser scanning methods and data processing methods are required to achieve accurate installation and positioning. However, existing laser scanning methods usually take a long time to scan the laser target and slow data processing, which may lead to reduced production efficiency and accuracy, and require a large amount of computing resources and time. In actual applications, when dealing with complex products or production needs that require rapid adjustment, it is often necessary to redesign and adjust the assembly process, which is time-consuming and inflexible. It is impossible to fully utilize the collected data for precise positioning and processing. When processing data, it is impossible to adapt to complex and changing environments, affecting the production line.
[0004] Therefore, we propose a fast laser target detection method based on α-shapes to solve the above problems. Summary of the Invention
[0005] (1) Technical problems solved
[0006] In view of the deficiencies of the prior art, the present invention provides a rapid laser target detection method based on α-shapes, which solves the problems raised in the above background technology.
[0007] (2) Technical solution
[0008] In order to achieve the above-mentioned purpose, the present invention specifically adopts the following technical solutions:
[0009] A rapid laser target detection method based on α-shapes comprises the following steps:
[0010] S1: Scan the circular laser target based on the non-uniform Archimedean spiral scanning method, and obtain the signal through the photoelectric detection module;
[0011] S2: After analog amplification and ADC conversion of the acquired signal, threshold processing is performed to obtain data of the area with high laser target reflectivity;
[0012] S3: Use the radial distance-based method to roughly detect and extract edge points, and the improved adaptive threshold α-shapes algorithm to accurately extract boundary points;
[0013] S4: Use the RANSAC algorithm to fit the obtained boundary points into a circle and extract the coordinates of the circle center;
[0014] S5: The above boundary point extraction method and circle center extraction method are hardware accelerated and optimized in the FPGA platform to obtain the laser target position at high speed and in real time.
[0015] Furthermore, in S1, the laser scanning projection system used for laser target detection includes:
[0016] Laser scanning method: A scanning path is designed based on a non-uniform Archimedean spiral, achieving the goal of dense edge scanning points and sparse middle scanning points. The center coordinates are calculated using the relevant information of the edge scanning points to improve the scanning speed and the utilization rate of the scanning data. The formula of the non-uniform Archimedean spiral is:
[0017]
[0018] Among them, a is the offset of the starting position, b is the growth amplitude of the distance, and θ is the angular position, which determines the rotation angle and coverage range of the spiral. The larger the range of change, the larger the number of turns and coverage area of the spiral. n controls the growth curve of the spiral, so that the radial distance is growth rate.
[0019] Laser: emits laser light and generates light signals. The light path is scanned within a predetermined range through galvanometer deflection.
[0020] Galvanometer: A special oscillating motor. Its principle is that an energized coil generates torque in a magnetic field. Unlike a rotating motor, its rotor is subjected to a restoring torque through an electromagnetic field. The magnitude of the restoring torque is proportional to the angle at which the rotor deviates from the equilibrium position. When current passes through the coil, the rotor deflects to a certain angle. The electromagnetic torque is equal to the restoring torque, so it cannot rotate like an ordinary motor and can only deflect. The deflection angle is linearly related to the voltage.
[0021] Galvanometer control module: Output differential signal to control the galvanometer motor to drive the X and Y galvanometers to rotate around the axis. After secondary reflection, they can be projected onto the laser target, realizing the movement of the laser beam within the predetermined scanning range.
[0022] During the laser scanning target process, the FPGA sends a control signal to the galvanometer control module, which drives the dual-axis scanning galvanometer to control the rapid deflection of the laser beam. After secondary reflection, it is projected onto the laser target to form a scanning point. The laser target reflects the light signal to the photoelectric detection module and converts it into an electrical signal.
[0023] Furthermore, in S2, the acquired signal undergoes analog amplification, ADC conversion, and threshold processing to obtain data on areas with high laser target reflectivity. An external analog circuit amplifies the tiny voltage signal output by the photoelectric detection module. The amplified analog signal is then fed to the ADC via an FPGA to acquire data. The ADC module converts the analog voltage signal into a digital signal. After threshold preprocessing, data above the threshold is retained, representing the highly reflective areas of the laser target, while data from low-reflective areas is discarded.
[0024] Furthermore, in S3, the method based on radial distance is used to perform rough detection on the data of the high reflection area to extract edge points and remove the data points in the obvious central area. Since the scanning center coordinates are known, the calculation of each point P i (x i ,y i ) to the scanning center C(x C ,y C )’s radial distance:
[0025]
[0026] r i Reflects point P i The distance of the position relative to C. Set a radial distance threshold G i , if r i Less than G i , it is determined to belong to the center area and removed. This is a simple and effective method for distinguishing data points from the center area and the edge area, without calculating neighboring points, ensuring the acquisition of high-quality edge data points. This reduces the computational complexity of the α-shapes algorithm while preventing edge data loss.
[0027] The improved adaptive threshold α-shapes algorithm is used to precisely extract the roughly extracted data. The α-shapes algorithm is a simple, effective, and fast algorithm for extracting boundary points. It can be used to extract edges from a collection of unordered points. It overcomes the disadvantage of the point cloud boundary point shape and can quickly and accurately extract boundary points.
[0028] The initial complexity of the ordinary α-shapes algorithm is O(n 3 ), can be reduced to O(n 2log n), each pair of points needs to be checked, which is computationally intensive. The improved adaptive threshold α-shapes algorithm uses Delaunay triangulation to reduce the number of points that need to be checked, significantly reducing the computational complexity to O(n 2 ), when the point set is evenly distributed and the number of edges is small, the complexity can be close to O(nlogn). At the same time, the adaptive α is adopted to make the algorithm adaptable to point sets with different densities.
[0029] The improved adaptive threshold α-shapes algorithm significantly reduces the amount of computation and improves the accuracy of boundary point extraction through structured processing and reasonable boundary judgment, thereby improving processing efficiency and real-time performance on FPGA.
[0030] Furthermore, in S4, the acquired boundary points are fitted into a circle using the RANSAC algorithm, and the coordinates of the circle center are extracted, thereby completing the detection of the laser target. The main steps of the RANSAC algorithm for fitting a circle are:
[0031] (1) The boundary point set extracted by the α-shapes algorithm is used as the data point set S of the algorithm. Fitting the circle requires randomly selecting at least three inner points in S to form a subset s1, which is used to estimate the parameters of the circle model M1 (i.e., the center and radius).
[0032] (2) Traverse all data points in the dataset S except the subset s1, calculate their distance to M1, and mark them as internal points if the distance is less than the given error threshold e, otherwise mark them as external points.
[0033] (3) The set s1 consisting of all the interior points of M1 itself * It is called the consistent set of the model. If the number of points in the consistent set meets the preset threshold T, then s1 is used * Regenerate model M1 * , end the algorithm.
[0034] (4) If the consistent set s1 * If the number of inliers is less than the threshold T, a new subset s2 is selected and steps 1-3 are repeated.
[0035] (5) After K iterations, a consistent set with the largest number of inliers is selected and used to regenerate the model, ending the algorithm.
[0036] The distance calculation formula from the data point mentioned in step (2) to the model M1 is:
[0037]
[0038] The coordinates of the boundary points are (x i ,y i ), the center of the fitted circle is (a, b).
[0039] Furthermore, in S5, after the entire system is powered on, the FPGA outputs a control signal to drive the galvanometer, causing the laser to scan the laser target along a preset path. The laser target reflects the light signal back to the photoelectric detection module, where it is amplified by the analog circuit and converted by the ADC module. After processing, the data is transmitted to the FPGA for further data processing, and finally the coordinates of the target's center are output.
[0040] The above boundary point extraction method is hardware accelerated and optimized in the FPGA platform to achieve parallel processing of data and reduce the data transmission delay between processors.
[0041] Furthermore, the FPGA platform mainly includes the following modules:
[0042] The galvanometer drive module receives instructions, outputs control signals and data streams, and controls the galvanometer to scan the laser target along a predetermined path;
[0043] ADC enable module, which outputs enable signal to drive ADC to collect data and receives digital signal converted by ADC;
[0044] The FIFO buffer module is used to temporarily store data transmitted from the ADC to solve the rate matching problem between different modules during data processing, so as to facilitate subsequent data processing;
[0045] The data preprocessing module performs threshold filtering on the collected data to remove invalid data;
[0046] The clock management module uses PLL to manage the clock requirements of different modules within the FPGA to ensure data synchronization and correct processing;
[0047] The data processing module uses algorithms to process data and accelerates the algorithm processing speed through parallel processing;
[0048] The serial communication module is used to transmit the calculated coordinate data to an external device through a serial interface.
[0049] (3) Beneficial effects
[0050] Compared with the existing technology, the present invention provides a rapid laser target detection method based on α-shapes, which has the following beneficial effects:
[0051] The present invention improves the problem of long scanning time of laser targets by traditional equally spaced spiral scanning methods, and proposes a non-uniform Archimedean spiral scanning method, which effectively avoids data redundancy in the central area and improves the sampling density in the boundary area, thereby realizing rapid scanning of laser targets.
[0052] The α-shapes algorithm is sensitive to the rolling circle radius, has high computational complexity and poor adaptability. An improved adaptive threshold α-shapes algorithm is proposed to flexibly process the obtained data, significantly reduce the amount of calculation, achieve accurate extraction of boundary points, improve data processing efficiency, and improve assembly efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] Figure 1 Schematic diagram of the process of the laser target detection method used in an embodiment of the present invention;
[0054] Figure 2 Schematic diagram of a non-uniform Archimedean spiral scanning method according to an embodiment of the present invention;
[0055] Figure 3 1 is a detection flow chart of a rapid laser target detection method based on α-shapes in an embodiment of the present invention;
[0056] Figure 4 4 is an overall block diagram of the laser target detection method in an embodiment of the present invention. DETAILED DESCRIPTION
[0057] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.
[0058] Example
[0059] See Figure 1-4 An embodiment of the present invention provides a rapid laser target detection method based on α-shapes, which specifically includes steps S1-S5:
[0060] S1: Scan the circular laser target based on the non-uniform Archimedean spiral scanning method, and obtain the signal through the photoelectric detection module;
[0061] The laser scanning projection system used for laser target detection includes:
[0062] Laser scanning method: A scanning path is designed based on a non-uniform Archimedean spiral, achieving the goal of dense edge scanning points and sparse middle scanning points. The center coordinates are calculated using the relevant information of the edge scanning points to improve the scanning speed and the utilization rate of the scanning data. The formula of the non-uniform Archimedean spiral is:
[0063]
[0064] Among them, a is the offset of the starting position, b is the growth amplitude of the distance, and θ is the angular position, which determines the rotation angle and coverage range of the spiral. The larger the range of change, the larger the number of turns and coverage area of the spiral. n controls the growth curve of the spiral, so that the radial distance is growth rate.
[0065] Laser: emits laser light and generates light signals. The light path is scanned within a predetermined range through galvanometer deflection.
[0066] Galvanometer: A special oscillating motor. Its principle is that an energized coil generates torque in a magnetic field. Unlike a rotating motor, its rotor is subjected to a restoring torque through an electromagnetic field. The magnitude of the restoring torque is proportional to the angle at which the rotor deviates from the equilibrium position. When current passes through the coil, the rotor deflects to a certain angle. The electromagnetic torque is equal to the restoring torque, so it cannot rotate like an ordinary motor and can only deflect. The deflection angle is linearly related to the voltage.
[0067] Galvanometer control module: Output differential signal to control the galvanometer motor to drive the X and Y galvanometers to rotate around the axis. After secondary reflection, they can be projected onto the laser target, realizing the movement of the laser beam within the predetermined scanning range.
[0068] During the laser scanning target process, the FPGA sends a control signal to the galvanometer control module, which drives the dual-axis scanning galvanometer to control the rapid deflection of the laser beam. After secondary reflection, it is projected onto the laser target to form a scanning point. The laser target reflects the light signal to the photoelectric detection module and converts it into an electrical signal.
[0069] S2: After analog amplification and ADC conversion of the acquired signal, threshold processing is performed to obtain data of the area with high laser target reflectivity;
[0070] An external analog circuit is used to amplify the tiny voltage signal output by the photoelectric detection module. The amplified analog signal is used to drive the ADC for data acquisition through the FPGA. The ADC module converts the analog voltage signal into a digital signal. After threshold preprocessing, data greater than the threshold is retained, that is, the high-reflection area of the laser target, and data in the low-reflection area is discarded.
[0071] S3: Use the radial distance-based method to roughly detect and extract edge points, and the improved adaptive threshold α-shapes algorithm to accurately extract boundary points;
[0072] In S3, the method based on radial distance is used to perform rough detection on the data of the high reflection area to extract edge points and remove the data points in the obvious central area. Since the coordinates of the scanning center are known, the calculation of each point P i (x i ,y i ) to the scanning center C(x C ,y C)’s radial distance:
[0073]
[0074] r i Reflects point P i The distance of the position relative to C. Set a radial distance threshold G i , if r i Less than G i , it is determined to belong to the center area and removed. This is a simple and effective method for distinguishing data points from the center area and the edge area, without calculating neighboring points, ensuring the acquisition of high-quality edge data points. This reduces the computational complexity of the α-shapes algorithm while preventing edge data loss.
[0075] The improved adaptive threshold α-shapes algorithm is used to precisely extract the roughly extracted data. The α-shapes algorithm is a simple, effective, and fast algorithm for extracting boundary points. It can be used to extract edges from a collection of unordered points. It overcomes the disadvantage of the point cloud boundary point shape and can quickly and accurately extract boundary points.
[0076] The initial complexity of the ordinary α-shapes algorithm is O(n 3 ), can be reduced to O(n 2 log n), each pair of points needs to be checked, which is computationally intensive. The improved adaptive threshold α-shapes algorithm uses Delaunay triangulation to reduce the number of points that need to be checked, significantly reducing the computational complexity to O(n 2 ), when the point set is evenly distributed and the number of edges is small, the complexity can be close to O(nlogn). At the same time, the adaptive α is adopted to make the algorithm adaptable to point sets with different densities.
[0077] The improved adaptive threshold α-shapes algorithm significantly reduces the amount of computation and improves the accuracy of boundary point extraction through structured processing and reasonable boundary judgment, thereby improving processing efficiency and real-time performance on FPGA.
[0078] S4: Use the RANSAC algorithm to fit the acquired boundary points into a circle and extract the coordinates of the circle center to complete the detection of the laser target.
[0079] In S4, the main steps of the RANSAC algorithm for fitting a circle are:
[0080] (1) The boundary point set extracted by the α-shapes algorithm is used as the data point set S of the algorithm. Fitting the circle requires randomly selecting at least three inner points in S to form a subset s1, which is used to estimate the parameters of the circle model M1 (i.e., the center and radius).
[0081] (2) Traverse all data points in the dataset S except the subset s1, calculate their distance to M1, and mark them as internal points if the distance is less than the given error threshold e, otherwise mark them as external points.
[0082] (3) The set s1 consisting of all the interior points of M1 itself * It is called the consistent set of the model. If the number of points in the consistent set meets the preset threshold T, then s1 is used * Regenerate model M1 * , end the algorithm.
[0083] (4) If the consistent set s1 * If the number of inliers is less than the threshold T, a new subset s2 is selected and steps 1-3 are repeated.
[0084] (5) After K iterations, a consistent set with the largest number of inliers is selected and used to regenerate the model, ending the algorithm.
[0085] The distance calculation formula from the data point mentioned in step (2) to the model M1 is:
[0086]
[0087] The coordinates of the boundary points are (x i ,y i ), the center of the fitted circle is (a, b).
[0088] S5: The above boundary point extraction method and circle center extraction method are hardware accelerated and optimized in the FPGA platform to obtain the laser target position at high speed and in real time.
[0089] See also Figure 4 , which is a general block diagram of laser target detection in an embodiment of the present invention, the method includes the following steps:
[0090] In S5, after the entire system is powered on, the FPGA control signal drives the galvanometer, causing the laser to scan the laser target along a preset path. The laser target reflects the light signal back to the photoelectric detection module, where it is amplified by the analog circuit and converted by the ADC module. After processing, the data is transmitted to the FPGA for further data processing, and finally the target center coordinates are output.
[0091] The FPGA platform mainly includes the following modules:
[0092] The galvanometer drive module receives instructions, outputs control signals and data streams, and controls the galvanometer to scan the laser target along a predetermined path;
[0093] ADC enable module, which outputs enable signal to drive ADC to collect data and receives digital signal converted by ADC;
[0094] The FIFO buffer module is used to temporarily store data transmitted from the ADC to solve the rate matching problem between different modules during data processing, so as to facilitate subsequent data processing;
[0095] The data preprocessing module performs threshold filtering on the collected data to remove invalid data;
[0096] The clock management module uses PLL to manage the clock requirements of different modules within the FPGA to ensure data synchronization and correct processing;
[0097] The data processing module uses algorithms to process data and accelerates the algorithm processing speed through parallel processing;
[0098] The serial communication module is used to transmit the calculated coordinate data to an external device through a serial interface.
[0099] The embodiment of the present invention provides a fast laser target detection method based on α-shapes. To achieve this purpose, this technical solution adopts a laser scanning projection system and integrates a mathematical model to design a fast and flexible non-uniform Archimedean spiral scanning method to obtain high-quality signal data. The improved adaptive threshold α-shapes algorithm is then used to process the data to extract the boundary points of the high-reflection area of the laser target. The RANSAC algorithm fits the boundary circle based on the boundary point data information, thereby calculating the center coordinates of the laser target to complete the detection of the laser target. It helps to better locate the laser target, and then use the positional relationship between the laser target and the working device to achieve assembly. Through optimization on the FPGA platform, the laser target position is obtained at high speed and in real time to achieve detection of the laser target.
[0100] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.
Claims
1. A rapid laser target detection method based on α-shapes, characterized in that: The following steps are included: S1: Scan the circular laser target based on the non-uniform Archimedean spiral scanning method, and obtain the signal through the photoelectric detection module; S2: After analog amplification and ADC conversion of the acquired signal, threshold processing is performed to obtain data of the area with high laser target reflectivity; S3: Use the radial distance-based method to roughly detect and extract edge points, and the improved adaptive threshold α-shapes algorithm to accurately extract boundary points; S4: Use the RANSAC algorithm to fit the acquired boundary points into a circle and extract the coordinates of the circle center to complete the detection of the laser target; S5: The boundary point extraction method and circle center extraction method are hardware accelerated and optimized on the FPGA platform to obtain the laser target position at high speed and in real time.
2. The rapid laser target detection method based on α-shapes according to claim 1, characterized in that: In S1, the laser scanning projection system used for laser target detection includes: Laser scanning method: A scanning path is designed based on a non-uniform Archimedean spiral to achieve dense edge scanning points and sparse middle scanning points. The center coordinates are calculated using the relevant information of the edge scanning points to improve the scanning speed and the utilization rate of the scanning data. The formula of the non-uniform Archimedean spiral is: Among them, a is the offset of the starting position, b is the growth amplitude of the distance, and θ is the angular position, which determines the rotation angle and coverage range of the spiral. The larger the range of change, the larger the number of turns and coverage area of the spiral. n controls the growth curve of the spiral, so that the radial distance is The rate of growth; Laser: emits laser, generates light signal, and realizes scanning of light path within a predetermined range through galvanometer deflection; Galvanometer: A special oscillating motor. Its principle is that an energized coil generates torque in a magnetic field. Unlike a rotary motor, its rotor is subjected to a restoring torque through the electromagnetic field. The magnitude is proportional to the angle at which the rotor deviates from the equilibrium position. When current flows through the coil, the rotor deflects to a certain angle. The electromagnetic torque is equal to the restoring torque, so it cannot rotate like an ordinary motor, but can only deflect. The deflection angle is linearly related to the voltage. Galvanometer control module: outputs differential signals to control the galvanometer motor to drive the X and Y galvanometers to rotate around the axis. After secondary reflection, the galvanometer can be projected onto the laser target, realizing the movement of the laser beam within the predetermined scanning range. During the laser scanning target process, the FPGA sends a control signal to the galvanometer control module, which drives the dual-axis scanning galvanometer to control the rapid deflection of the laser beam. After secondary reflection, it is projected onto the laser target to form a scanning point. The laser target reflects the light signal to the photoelectric detection module and converts it into an electrical signal.
3. The rapid laser target detection method based on α-shapes according to claim 1, characterized in that: In S2, an external analog circuit is used to amplify the tiny voltage signal output by the photoelectric detection module. The amplified analog signal is used to drive the ADC through the FPGA to collect data. The ADC module converts the analog voltage signal into a digital signal. After threshold preprocessing, data greater than the threshold is retained, that is, the high-reflection area of the laser target, and data in the low-reflection area is discarded.
4. The rapid laser target detection method based on α-shapes according to claim 3, characterized in that: In S3, the method based on radial distance is used to roughly detect the data of the high reflection area and extract the edge points, and remove the data points in the obvious central area; since the coordinates of the scanning center are known, the calculation of each point P is performed. i (x i ,y i ) to the scanning center C(x C ,y C )’s radial distance: r i Reflects point P i The position of C is close to the distance between the two, and a radial distance threshold G is set. i , if r i Less than G i , it is determined to belong to the central area and removed. This is simple and effective for distinguishing data points in the central area and the edge area. It does not need to calculate the nearest points, ensuring the acquisition of high-quality edge data points. While reducing the calculation amount of the α-shapes algorithm, it prevents the loss of edge data. The improved adaptive threshold α-shapes algorithm is used to accurately extract the roughly extracted data. The α-shapes algorithm is a simple and effective algorithm for quickly extracting boundary points. It can be used to extract edges from a pile of disordered point sets. It overcomes the disadvantage of the influence of the shape of the boundary points of the point cloud and can quickly and accurately extract boundary points. The initial complexity of the ordinary α-shapes algorithm is O(n 3 ), can be reduced to O(n 2 log n), each pair of points needs to be checked, which is computationally intensive. The improved adaptive threshold α-shapes algorithm uses Delaunay triangulation to reduce the number of points that need to be checked, significantly reducing the computational complexity to O(n 2 ), when the point set is evenly distributed and the number of edges is small, the complexity can be close to O(nlogn), and the adaptive α is used to make the algorithm adaptable to point sets with different densities; The improved adaptive threshold α-shapes algorithm significantly reduces the amount of computation and improves the accuracy of boundary point extraction through structured processing and reasonable boundary judgment, thereby improving processing efficiency and real-time performance on FPGA.
5. The rapid laser target detection method based on α-shapes according to claim 4, characterized in that: In S4, the main steps of the RANSAC algorithm for fitting a circle are: (1) The boundary point set extracted by the α-shapes algorithm is used as the data point set S of the algorithm. To fit the circle, at least three inner points in S must be randomly selected to form a subset s1, which is used to estimate the parameters of the circle model M1, namely the center and radius. (2) Traverse all data points in the dataset S except the subset s1, calculate their distance to M1, and mark them as internal points if the distance is less than the given error threshold e, otherwise mark them as external points; (3) The set s1 consisting of all the interior points of M1 itself * It is called the consistent set of the model. If the number of points in the consistent set meets the preset threshold T, then s1 is used * Regenerate model M1 * , end the algorithm; (4) If the consistent set s1 * If the number of inliers is less than the threshold T, a new subset s2 is selected and steps 1-3 are repeated; (5) After K iterations, a consistent set with the largest number of inliers is selected and used to regenerate the model, ending the algorithm. The distance calculation formula from the data point mentioned in step (2) to the model M1 is: The coordinates of the boundary points are (x i ,y i ), the center of the fitted circle is (a, b).
6. The rapid laser target detection method based on α-shapes according to claim 5, characterized in that: In S5, after the entire system is powered on, the FPGA outputs a control signal to drive the galvanometer, causing the laser to scan the laser target along a preset path. The laser target reflects the light signal back to the photoelectric detection module, which then amplifies the signal through the analog circuit and converts it through the ADC module. After processing, the data is transmitted to the FPGA for subsequent data processing, and finally the coordinates of the target's center are output. The above boundary point extraction method is hardware accelerated and optimized in the FPGA platform to achieve parallel processing of data and reduce the data transmission delay between processors.
7. The method of rapid laser target detection based on α-shapes according to claim 5, characterized in that: The FPGA platform mainly includes the following modules: The galvanometer drive module receives instructions, outputs control signals and data streams, and controls the galvanometer to scan the laser target along a predetermined path; ADC enable module, which outputs enable signal to drive ADC to collect data and receives digital signal converted by ADC; The FIFO buffer module is used to temporarily store data transmitted from the ADC to solve the rate matching problem between different modules during data processing, so as to facilitate subsequent data processing; The data preprocessing module performs threshold filtering on the collected data to remove invalid data; The clock management module uses PLL to manage the clock requirements of different modules within the FPGA to ensure data synchronization and correct processing; The data processing module uses algorithms to process data and accelerates the algorithm processing speed through parallel processing; The serial communication module is used to transmit the calculated coordinate data to an external device through a serial interface.