Single-line laser radar solid particle preparation method

Abstract

The invention discloses a single-line laser radar solid particle preparation method, and belongs to the technical field of laser radar data processing, and the method comprises the following steps: S1, a single-line laser radar obtains original point cloud data according to the flight time, extracts feature information according to a set feature model, and sends the feature information to a particle feature analysis module for analysis; comprising the following steps: analyzing reflection intensity; point cloud density analysis; motion trail analysis; s2, particle judgment: calculating a particle probability P through weighted fusion based on the characteristic value in the step S1, and marking a particle swarm or a real obstacle according to a P value range; s3, dynamic parameter adjustment: adaptively adjusting a motion threshold value in motion trail analysis according to the particle concentration of the environment; and S4, false alarm suppression: removing the point cloud marked as the particle swarm, and outputting purified real obstacle data. By analyzing the target reflection intensity, the motion mode and the spatial distribution characteristics, the particle swarms and the real obstacles are dynamically distinguished, and the misjudgment rate is remarkably reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of laser radar data processing, in particular to a method for simulating solid particles of a single-line laser radar. BACKGROUND

[0002] Single-line laser radar is a sensor that obtains two-dimensional plane distance data by rotating scanning of a single laser beam. Single-line laser radar is widely used in AGV navigation, equipment safety protection, and personnel intrusion detection. However, in complex industrial environments such as metal processing workshops, welding areas, and oil mist environments, suspended solid particles (usually less than 5mm in diameter, such as metal debris, dust, rain, and oil stains) can reflect laser signals, which are easily misidentified as obstacles (such as personnel intrusion) by single-line laser radar, resulting in false shutdown of equipment or false obstacle avoidance of AGV. Traditional filtering algorithms cannot distinguish between dynamic particles and real obstacles.

[0003] The traditional solution has the following defects: 1. Static threshold filtering: ignores points with signal intensity below the threshold, but cannot distinguish between weakly reflective obstacles and particles. Particles randomly reflect and are identified as "instantaneous obstacles", with a high false positive rate. Moreover, fixed thresholds cannot adapt to dynamic changes in particle concentration, and have poor adaptability.

[0004] 2. Multi-frame matching method: relies on the continuity of target motion trajectory, but the random motion of particles causes the algorithm to fail.

[0005] 3. Traditional clustering algorithm: high computational power consumption, difficult to meet real-time requirements. SUMMARY

[0006] To overcome the shortcomings of the prior art, the present application provides a method for simulating solid particles of a single-line laser radar.

[0007] The present application adopts the following technical means: A method for simulating solid particles of a single-line laser radar, comprising the following steps: Step S1, the single-line laser radar acquires original point cloud data according to the time of flight, extracts feature information according to the set feature model, and sends it to the particle feature analysis module for analysis, including: Reflection intensity analysis; calculate the standard deviation of the signal intensity of all signal points in the scanning window to quantify the reflection intensity dispersion; if the reflection intensity dispersion is higher than the preset maximum dispersion, mark it as a particle group; Point cloud density analysis; detect the point cloud density mutation rate Δρ of adjacent scanning areas, and verify and mark in combination with the convex hull bounding box area S; Motion trajectory analysis; calculate the angle between adjacent frame displacement vectors to derive the motion continuity index η; Step S2, particle determination: based on the characteristic value of step S1, the particle probability P is calculated by weighted fusion, and the particle group or real obstacle is marked according to the P value range; Step S3, dynamic parameter adjustment: the motion threshold in the motion trajectory analysis is adaptively adjusted according to the particle concentration of the environment; Step S4, false alarm suppression: the point cloud marked as the particle group is removed, and the purified real obstacle data is output.

[0008] Further, the reflection intensity analysis in step S1 includes: The reflection intensity values of all signal points in the scanning window are input, and the reflection intensity mean value is calculated ; wherein N is the number of all signal points in the scanning window; i is the reflection intensity value of the i-th signal point, ranging from 0 to 255; The reflection intensity standard deviation is calculated ; The normalized reflection intensity dispersion σ is calculated; wherein σ max is the preset maximum standard deviation; If the normalized reflection intensity dispersion σ norm > 0.8, it is marked as a particle group.

[0009] Further, the point cloud density analysis in step S1 includes: The number of point clouds in the unit scanning area and the adjacent area is determined; The point cloud density mutation rate Δρ = |ρ k - ρ k-1 | / ρ k-1 × 100% of the unit scanning area is calculated; ρ k is the point cloud number of the k-th unit scanning area; Combined with the convex hull bounding box area S, if Δρ > 50% and the convex hull bounding box area S < 0.1 m², it is marked as a particle group.

[0010] Further, the motion trajectory analysis in step S1 includes: The same target is tracked across frames by target clustering tracking, and the displacement direction angle is recorded; The motion continuity index η is calculated ; wherein θ1 is the previous frame displacement direction angle, and θ2 is the current frame displacement direction angle; If the motion continuity index η < motion threshold, it is determined as random motion and marked as a particle group.

[0011] Further, in step S2, the reflection intensity dispersion σ norm, point cloud density change rate Δρ and motion discontinuity (1-η) are normalized by eigenvalue: Particle probability calculation: where w1, w2 and w3 are preset weights; the determination rule includes: If P ≤ 0.4, directly marked as a real obstacle; If 0.4<P ≤ 0.7, trigger cross-frame verification, if P value of continuous 3 frames are all <0.7, mark as a real obstacle; If P>0.7, directly mark as a particle group.

[0012] Further, the dynamic parameter adjustment of step S3 includes: Calculate the environmental particle concentration C 颗粒 =(number of particle marked points / total number of point cloud points)×100%; If C 颗粒 >30%, switch to a relaxed mode, reduce the motion threshold of the motion continuity index η; and return to step S1 to re-analyze the motion trajectory.

[0013] Further, the false alarm suppression of step S4 includes: Remove the point cloud marked as a particle group: delete the point marked as a particle group from the point cloud set; Output real obstacle data, including real obstacle coordinates and motion vector.

[0014] A method for solid particle simulation of single-line laser radar, which dynamically distinguishes particle groups and real obstacles by analyzing target reflection intensity, motion mode and spatial distribution characteristics, significantly reduces the misjudgment rate; it has the following effects: One, multi-feature fusion mechanism solves the traditional defects.

[0015] Traditional defects: traditional static threshold filtering cannot distinguish weak reflection obstacles and particles, and multi-frame matching method fails due to random motion of particles.

[0016] The scheme, 1, reflection intensity analysis: through reflection intensity dispersion σ norm Quantify the reflection instability of particle groups; 2, point cloud density mutation + convex hull verification: identify contourless targets; 3, motion continuity index: distinguish between random motion (η<0.3) and continuous motion (η=0.577→obstacle).

[0017] Effect: false alarm rate from 18.2% to 1.3%.

[0018] Two, dynamic threshold improves environmental adaptability.

[0019] Traditional defects: fixed threshold cannot cope with dynamic changes in particle concentration (such as instantaneous concentration rising to 33.7% caused by oil mist injection).

[0020] The scheme 1 is based on C 颗粒 Real-time adjustment of motion continuity index η threshold; 2, trigger cross-frame verification by weighted particle probability P (P = 0.427 → continuous 3-frame verification).

[0021] Effect: false positive rate fluctuation is controlled within ±0.2% (fixed threshold scheme fluctuation reaches ±15%).

[0022] Three, lightweight design guarantees real-time performance.

[0023] Traditional defects: traditional clustering algorithm consumes large amount of computing power, and it is difficult to meet the real-time demand of 25Hz scanning.

[0024] The scheme 1, the algorithm complexity is optimized to O(N); 2, embedded deployment (NVIDIA Jetson TX2), single-frame processing delay <10ms.

[0025] Effect: CPU occupancy <15%, supporting high-precision scenarios with 270° scanning angle and 0.25° resolution. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 The schematic diagram of the principle of the present application is shown in the figure; Figure 2 The flowchart of the reflection intensity analysis is shown in the figure. DETAILED DESCRIPTION

[0027] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0028] Explanation of terms.

[0029] Pseudo-algorithm: an algorithm for suppressing specific interference signals through data processing.

[0030] Point cloud density: the number of effective data points returned by the laser radar per unit angle.

[0031] Motion continuity index: a smoothness quantitative value of the motion trajectory between adjacent frames of the target.

[0032] In this scheme, the single-line lidar acquires raw point cloud data based on flight time, extracts feature information according to the set feature model, and sends it to the particle feature analysis module for analysis, including reflection intensity analysis, point cloud density analysis, and motion trajectory analysis. After comprehensive judgment, it is marked as a particle group for simulation. If it is a real obstacle, it performs corresponding protection processing (stopping / obstacle avoidance / optimizing navigation route, etc.).

[0033] Figure 1 This is a schematic diagram illustrating the principle of the present invention, such as... Figure 1 As shown, a method for fabricating solid-state particles in a single-line lidar system includes the following steps: Step S1, dynamic feature extraction.

[0034] This step extracts key features from the raw point cloud data, providing a foundation for particle identification. The input is the raw point cloud dataset acquired by LiDAR, and the output is quantified feature values: reflection intensity dispersion, point cloud density change rate, and motion continuity index. Feature extraction is based on a sliding window mechanism, with a configurable window size (e.g., a 30° sector area corresponding to 120 data points).

[0035] Step S101: Reflection intensity analysis: Calculate the standard deviation of the signal intensity of all signal points within the scanning window and quantify the reflection intensity dispersion; if the reflection intensity dispersion is higher than the preset maximum dispersion, it is marked as a particle group. Figure 2 This is a flowchart of the reflection intensity analysis, such as... Figure 2 As shown.

[0036] Solid particles, due to their varying size and material, typically exhibit discrete and unstable reflected signal intensity, while real obstacles (such as metal equipment) show more uniform reflected intensity. This solution effectively distinguishes weakly reflective targets (particles) from real obstacles through reflection intensity analysis, avoiding the limitations of static threshold filtering. It also boasts low computational complexity (only O(N) time per window), making it suitable for real-time processing.

[0037] Example: AGV protection system in a turning workshop.

[0038] Hardware configuration: LiDAR: S300-C-3 (270° scanning angle, 0.25° resolution); Processor: NVIDIA Jetson TX2 (ARM Cortex-A57); Environmental parameters: Metal scrap density: 150-200 pieces / m³ (diameter 0.5-2mm); Scanning frequency: 25Hz (40ms per frame).

[0039] In this example, the window size is a 30° sector area (corresponding to 120 data points, with a resolution of 0.25°).

[0040] Table 1 is an example of a point cloud in the window (reflection intensity value 0-255) of the present embodiment.

[0041]

[0042] A 30° sector is set to balance feature sensitivity and computational efficiency: too small a window (<10°) is susceptible to noise interference; too large a window (>60°) will smooth local features.

[0043] In combination with the embodiments, the extraction steps of the reflection intensity dispersion are as follows: Step S101a: input the reflection intensity values of all signal points in the scanning window, and calculate the reflection intensity mean μ.

[0044] The calculation formula of the reflection intensity mean μ is: ; wherein N is the number of all signal points in the scanning window, which is 120 in the present embodiment; li is the reflection intensity value of the i-th signal point, ranging from 0 to 255. i

[0045] The reflection intensity mean μ reflects the concentration trend of signal intensity in the window, and a low mean value may indicate a weakly reflective target (such as a dark obstacle or a distant particle).

[0046] Step S101b: calculate the reflection intensity standard deviation σ.

[0047] The calculation formula of the reflection intensity standard deviation σ is: .

[0048] The reflection intensity standard deviation σ quantifies the degree of dispersion of signal intensity. The particle group has a large fluctuation in reflection intensity (high σ value) due to irregular shape, while the flat obstacle (such as a wall surface) has stable reflection intensity (low σ value).

[0049] Step S101c: calculate the normalized reflection intensity dispersion σ norm .

[0050] The calculation formula of the normalized reflection intensity dispersion σ norm is: ; wherein σ max is the preset maximum standard deviation, which is preset based on experimental calibration, and σ max is 40 in the present embodiment.

[0051] The normalized reflection intensity dispersion σ norm eliminates the dimensional influence by normalization, compresses the dispersion to the interval [0, 1], and facilitates unified threshold determination. The fluctuation range of reflection intensity is significantly different under different environmental particle concentrations (such as the average intensity of oil mist environment is only 20-50, and the dust environment can reach 100-200). Normalization makes the algorithm adaptable to variable scenarios. ​

[0052] Step S101d: Determination of particle groups.

[0053] Judgment rule: If the normalized reflection intensity dispersion σ norm If the value is greater than 0.8, it is marked as a particle group.

[0054] In this embodiment, the point cloud intensity values ​​within the window are: 121, 120, 125, 78, 21, ..., 32, 200, 210, 80. Calculation: , , .

[0055] If σ norm Values ​​greater than 0.8 are labeled as particle groups. In this embodiment, the normalized reflection intensity dispersion σ norm =0.23≤0.8, no flag is triggered.

[0056] For particle group identification, a threshold of 0.8 was set, corresponding to an actual standard deviation of 32 (i.e., 40 × 0.8). In the turning workshop experiment, 32 was the critical value for distinguishing particle groups from obstacles. The reason why particle marking was not triggered in this embodiment is that there are continuous high-reflectivity points (such as 121, 120, 125) within the window, which may be the edges of metal obstacles; the discrete values ​​mainly come from a few outliers (such as 21, 32), which may be individual distant particles, but because the proportion of high values ​​is large, the overall dispersion does not exceed the standard.

[0057] The process of calculating the discreteness is shown using LaTeX format: \begin{align*} \mu&= \frac{1}{120} \sum_{i=1}^{120} I_i = \frac{121+120+125+\cdots+117}{120} = 51.3 \\ \sigma&= \sqrt{\frac{1}{120} \sum_{i=1}^{120} (I_i - 51.3)^2} = 9.2\\ \sigma_{\text{norm}}&= \frac{9.2}{40} = 0.23 \quad (\text{Default}\sigma_{\max}=40) \end{align*}.

[0058] S102, Point Cloud Density Analysis: Detect the point cloud density abrupt change rate Δρ in adjacent scanned areas, and verify and label it in conjunction with the convex hull bounding box area S: If Δρ>50% and the convex hull bounding box area S<0.1m², it is labeled as a particle group.

[0059] Due to the random distribution of particles, solid particles cause the point cloud density to change abruptly; while the real obstacle density changes smoothly. By comparing the rate of change of the point cloud density of adjacent regions, the contour is verified by the convex hull bounding box area S.

[0060] Table 2 is a point cloud quantity comparison table of the region with serial number K and its adjacent regions in this embodiment.

[0061]

[0062] Step S102a: Determine the point cloud quantity of the unit scanning region and the adjacent region.

[0063] The unit scanning region is 30°. When the laser radar scans the environment, the 360° field of view is divided into 12 30° sectors. The abrupt change of the point cloud number of the adjacent region reflects the sharp change of the local spatial distribution.

[0064] Step S102b: Calculate the point cloud density mutation rate Δρ of the unit scanning region.

[0065] The calculation formula of the point cloud density mutation rate Δρ is: Δρ = |ρ k - ρ k-1 | / ρ k-1 × 100%; ρ k is the point cloud number of the kth unit scanning region; In this embodiment, Δρ = 64.4%. The mutation rate quantifies the degree of density change. 64.4% is much higher than the typical threshold (50%) of the industrial scene, indicating that there is a non-continuous distribution target in this region.

[0066] Step S102c: Verify and mark in combination with the convex hull bounding box area S. If Δρ > 50% and the convex hull bounding box area S < 0.1 m², mark it as a particle group.

[0067] The convex hull bounding box area S < 0.1 m² indicates that there is no clustered contour. The convex hull detection area S = 0.08 m 2 ² in this embodiment. Processing the dynamic change of particle concentration avoids the failure of fixed threshold, and the convex hull detection ensures that only non-contour targets are filtered.

[0068] The convex hull bounding box area is the minimum circumscribed rectangle area of the convex hull, which is calculated using the Andrew or Graham scanning method. The specific steps are as follows: 1. Calculate the vertex coordinates of the convex hull.

[0069] Input data: coordinate set A = {(x1, y1), (x2, y2),..., (x n , y n )..., (x N , yN )}; where n is the index of the point cloud, (x n ,y n ) are the coordinates of the nth point cloud, and N is the total number of points in the input point set.

[0070] Convex Hull Generation: Extract the convex hull vertices H={h1,h2,...,h} of the point cloud using the Andrew method or Graham scan method. m ...,h M} (sorted clockwise or counterclockwise); where m is the index of the convex hull vertex, and M is the total number of vertices of the output convex hull, usually M < N.

[0071] 2. Calculate the minimum bounding rectangle of the convex hull.

[0072] The minimum bounding rectangle must satisfy the following conditions: its edges are parallel to at least one edge of the convex hull; and all vertices are located inside the rectangle.

[0073] Taking a two-dimensional plane as an example, traversing each edge of the convex hull: opposite edge h i h i+1 Calculate its direction vector =(dx,dy); h i Let be the vertex of the i-th convex hull.

[0074] Calculate the perpendicular distances from all vertices of the convex hull to the edge, and take the maximum perpendicular distance d. max (i.e., height).

[0075] The side length is ; The area S of the rectangle corresponding to the side of index i i =L×d max .

[0076] Choose the minimum area: the area S of the rectangle with all sides i Take the minimum value; take the minimum area S as the area of ​​the convex hull bounding box.

[0077] Example calculation: Assume the convex hull of the abrupt change region is a quadrilateral, and the vertex coordinates are as follows (unit: meters): h1(0,0),h2(0.4,0),h3(0.4,0.1),h4(0,0.1); Traversing edges h1h2: direction vector =(0.4,0); Large vertical distance d max =0.1m (y coordinates of vertices h3 and h4); Area S1 = 0.4 × 0.1 = 0.04 m² 2 .

[0078] Traversing edges h3h4: direction vector = (0, 0.1); Maximum vertical distance d max = 0.4m; Area S2 = 0.1 x 0.4 = 0.04m 2 .

[0079] Other sides are the same, and finally the minimum area S = 0.04m 2 .

[0080] Show the mutation rate calculation process in LaTeX format: \Delta\rho = \frac{|42 - 118|}{118} \times 100\% = 64.4\%.

[0081] Step S103, trajectory analysis: calculate the angle between adjacent frame displacement vectors, and derive the motion continuity index η.

[0082] Solid particles move randomly (Brownian motion), while real obstacles move continuously (direction changes are smooth). By target clustering tracking (such as the point group of ID:100), calculate the angle between adjacent frame displacement vectors, and derive the motion continuity index η.

[0083] Step S103a, record the displacement direction angle through target clustering tracking.

[0084] Input data: clustered point group of target ID:100 (stable target generated by point cloud clustering algorithm).

[0085] Table 3 is the data table of the clustered point group of target ID:100.

[0086]

[0087] In this embodiment, the position data of three consecutive frames (T-2, T-1, T) are used to track the same target across frames, avoiding target ID jumping caused by point cloud jitter. The displacement direction angle θ1=45° of the previous frame; the displacement direction angle θ2=45° of the current frame. Record the displacement direction angle, and quantify the continuity of target motion through the change angle of adjacent frame motion direction.

[0088] Step S103b, calculate the motion continuity index η. .

[0089] In this embodiment, θ1=45°, θ2=63.4°, .

[0090] Step S103c, decision rule: if motion continuity index η < motion threshold, then determine as random motion, mark as particle group; otherwise, do not trigger. Motion threshold is 0.3 by default.

[0091] In this embodiment, η = 0.577 > 0.3, and no trigger is generated.

[0092] This step solves the problem of failure of the multi-frame matching method under random particle motion. The algorithm is efficient (single target calculation delay < 0.1 ms) and supports real-time tracking.

[0093] Show the motion trajectory analysis and calculation process in LaTeX format: \eta = \frac{\cos45^\circ + \cos63.4^\circ}{2} = \frac{0.707 + 0.447}{2} = 0.577.

[0094] In step S1, if any of the following conditions is met, it is marked as a particle group: Condition 1: the reflection intensity dispersion degree is greater than the preset maximum dispersion degree (e.g., 0.8); Condition 2: the point cloud density mutation rate Δρ > 50% and the convex hull bounding box area S < 0.1 m² (not forming a stable cluster); Condition 3: the motion continuity index < 0.3; where, the motion continuity index of 0 represents complete randomness, and the motion continuity index of 1 represents uniform straight-line motion.

[0095] Step S2: decision rule for particle group.

[0096] This step is based on the feature extraction results and comprehensively judges whether the point cloud is a solid particle and triggers the verification mechanism. The input is the output feature value of step 1, and the output is the particle probability P and the preliminary judgment. In this step, a weighted fusion mechanism is used to avoid misjudgment of a single feature. The particle probability P is the weighted sum of the reflection intensity dispersion degree σ norm , the point cloud density change rate Δρ, and the motion discontinuity (1-η).

[0097] Feature value normalization is performed on the reflection intensity dispersion degree σ norm , the point cloud density change rate Δρ, and the motion discontinuity (1-η): The normalized dispersion degree σ norm = 0.23 is used as the input value; The point cloud density mutation rate Δρ = 64.4%, and the normalized Δρ = 0.644 is obtained; The motion continuity index η = 0.577, and the motion discontinuity (1-η) = 0.423 is obtained.

[0098] Particle probability calculation: where w1, w2 and w3 are preset weights.

[0099] In this embodiment, w1=0.5, w2=0.3, w3=0.2; .

[0100] Determination logic: If P ≤ 0.4, directly marked as a real obstacle; If 0.4 < P ≤ 0.7, trigger cross-frame verification, if the P value of the last 3 frames are all <0.7, mark as a real obstacle; If P > 0.7, directly marked as a particle group.

[0101] In this embodiment, P = 0.427 ∈ (0.4, 0.7], trigger cross-frame verification; the P values of the last 3 frames [0.427, 0.402, 0.415] are all <0.7, finally marked as a real obstacle.

[0102] This step reduces the false positive rate through multi-feature fusion, and enhances the robustness through cross-frame verification. The algorithm is lightweight and supports embedded deployment.

[0103] Step S3: Dynamic parameter adjustment: input the number of particle marked points and the total number of point cloud points per unit time, calculate the particle concentration C 颗粒 If C 颗粒 > preset concentration threshold, switch to relaxed mode, reduce the motion threshold of motion continuity index η, and return to step S1 to re-analyze the motion trajectory.

[0104] This step adaptively adjusts the determination threshold according to the environmental particle concentration, improving the adaptability of the scheme.

[0105] Technical principle: relax the determination threshold (reduce the motion threshold) when the particle concentration is high to avoid excessive sensitivity; tighten the motion threshold when the concentration is low to improve accuracy.

[0106] Calculation formula: particle concentration: C 颗粒 = (number of particle marked points / total number of point cloud points) × 100%.

[0107] Motion threshold adjustment: if C 颗粒 > 30%, switch to relaxed mode, the motion threshold is reduced from 0.3 to 0.2.

[0108] In this embodiment, 10-second period: total point cloud 9500 points, particle marked 3200 points.

[0109] Calculation: .

[0110] Adjustment: C 颗粒=33.7%>30%, switch to loose mode, motion threshold reduced to 0.2.

[0111] This step realizes environmental adaptation and reduces manual parameter adjustment. In actual measurement, it can cope with the change of metal debris density of 150-200 / m³.

[0112] Step S4: False alarm suppression: the point cloud marked as particle group does not participate in subsequent clustering and trajectory tracking, and only the real obstacle data is retained; the input is the determination result, and the output is the purified real obstacle data, including real obstacle coordinates and motion vector.

[0113] This step performs final output processing and suppresses false alarm point cloud.

[0114] Processing flow: remove particle point cloud: delete the points marked as particle group from the point cloud set (for example, 1200 points are removed in this embodiment). Output real obstacle data: including ID, position, speed, contour area and other structured data.

[0115] AGV navigation or equipment protection, receive real obstacle data, and perform corresponding protection processing (shutdown / obstacle avoidance / optimize navigation route, etc.).

[0116] This step ensures that AGV navigation or equipment protection only responds to real threats. The false alarm rate is reduced from 18.2% to 1.3% (72 hours of turning data in the turning workshop).

[0117] Analysis of scene depth technology: 1. Metal debris splashing scene.

[0118] Challenge nature: turning process produces 0.5-2mm metal debris, forming a 150-200 particle / m³ high-density suspended layer, and the reflection intensity fluctuates sharply (60-220), and traditional static threshold filtering cannot distinguish weak reflection obstacles.

[0119] Innovative response: adopt double mechanisms of reflection intensity analysis (reflection intensity dispersion σ norm >0.8 trigger marking) and motion trajectory analysis (motion continuity index η<0.3 determine random motion).

[0120] Achievements: false alarm times are reduced from 127 to 9, which is equivalent to avoiding 1.2 times of non-planned shutdown per hour of production line, and the overall equipment efficiency (OEE) is improved by 13%.

[0121] 2. Oil mist injection scene.

[0122] Challenge nature: cooling oil mist generates 1-3μm ultrafine particles, reflection signal is weak (20-50) but density mutation rate is more than 70%, causing multi-frame matching method to fail due to Brownian motion.

[0123] Innovative approach: Introduce point cloud density analysis (started when point cloud density mutation rate Δρ>50%) and convex hull profile verification (convex hull bounding box area S<0.1m², confirming no solid profile).

[0124] Results: The number of false alarms was reduced from 89 to 6, ensuring the continuous operation stability of the high-pressure cooling system and extending the maintenance cycle by 40%.

[0125] 3. AGV dust pollution scenario.

[0126] The challenge lies in the fact that AGVs raise a 0.3-0.8m low-altitude dust layer as they move, which traditional clustering algorithms misjudge as "low obstacles," triggering frequent obstacle avoidance interruptions.

[0127] Innovative approach: Accurate identification is achieved through dynamic threshold adjustment (the motion threshold of the motion continuity index η decreases from 0.3 to 0.2 when the dust concentration is >25%) and multi-feature weighted fusion (particles with a probability P >0.7 are directly labeled as particle groups).

[0128] Results: The number of false alarms was reduced from 54 to 4, AGV throughput efficiency increased by 22%, and the annual logistics throughput increased by 150,000 pieces.

[0129] The advantages of this solution are as follows: 1. Multi-feature fusion mechanism: For the first time, the reflection intensity discreteness σ norm The triple feature fusion of point cloud density mutation rate Δρ and motion continuity index η solves the problem of high false positive rate of single feature. It remains stable within the dynamic range of particle concentration (150-350 particles / m³) with a false negative rate of <0.5%. In contrast, the false positive rate of single feature scheme is still as high as 8.2-12.7%.

[0130] 2. Dynamic threshold adjustment: The threshold for judgment is adaptively optimized based on real-time particle concentration to adapt to dynamic environmental changes (such as instantaneous concentration spikes caused by oil mist injection).

[0131] 3. Lightweight real-time processing: The algorithm complexity is optimized to O(n), the single-frame processing latency on the embedded platform is <10ms, and the CPU usage on the Jetson TX2 platform is <15%, meeting the real-time requirements of industrial scenarios.

[0132] In this solution, point cloud data is collected by a single-line laser radar (such as S300-C-3 model), combined with multi-feature analysis (reflection intensity, point cloud density, motion trajectory) to dynamically distinguish solid particles from real obstacles (such as personnel or AGV). The algorithm runs on an embedded platform (such as NVIDIA Jetson TX2), with a processing delay of <10ms, supporting real-time applications (scanning frequency 25Hz). The advantages of the solution include a false positive rate reduction of more than 80% (from 18.2% to 1.3%) and environmental adaptability.

[0133] In addition, it should be understood that although the present specification is described in terms of embodiments, not every embodiment contains only one independent technical solution, and the description of the specification is only for the sake of clarity, and those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can be appropriately combined to form other embodiments that those skilled in the art can understand.

Claims

1. A method of monostatic lidar solid particle simulation, characterized by, Comprising the following steps: Step S1, single-line laser radar acquires raw point cloud data according to time of flight, extracts feature information according to the set feature model, and sends it to the particle feature analysis module for analysis, including: Reflection intensity analysis; calculate the standard deviation of the signal intensity of all signal points in the scanning window, quantify the reflection intensity dispersion; if the reflection intensity dispersion is higher than the preset maximum dispersion, mark it as a particle group; Point cloud density analysis; detect the point cloud density mutation rate Δρ of adjacent scanning areas, and verify and mark in combination with the convex hull bounding box area S; Motion trajectory analysis; calculate the angle of adjacent frame displacement vectors, and deduce the motion continuity index η; Step S2, particle determination: based on the characteristic values of step S1, calculate the particle probability P through weighted fusion, and mark the particle group or real obstacle according to the P value range; Step S3, dynamic parameter adjustment: adaptively adjust the motion threshold in the motion trajectory analysis according to the particle concentration of the environment; Step S4, false alarm suppression: remove the point cloud marked as a particle group, and output the purified real obstacle data.

2. The method of claim 1, wherein the single-beam lidar solid particle fabrication method is characterized by, The reflection intensity analysis in step S1 includes: Input the reflection intensity value of all signal points in the scanning window, calculate the average value of the reflection intensity ; wherein N is the number of all signal points in the scanning window; and i is the reflection intensity value of the i-th signal point, ranging from 0 to 255. Computing the standard deviation of the reflected intensity ; Computing a normalized reflectance intensity dispersion ; wherein σ max is a predetermined maximum standard deviation; If the normalized intensity dispersion σ norm > 0.8, the particle group is marked.

3. The method of claim 2, wherein the single-beam lidar solid particle fabrication method is characterized by, The point cloud density analysis in step S1 includes: Determine the number of point clouds in a unit scanning area and adjacent areas; The point cloud density mutation rate Δρ of the unit scanning area is calculated as follows: Δρ = |ρ k - ρ k-1 | / ρ k-1 × 100%; ρ k is the point cloud number of the kth unit scanning area; In combination with the convex hull bounding box area S verification and marking, if Δρ > 50% and the convex hull bounding box area S < 0.1 m², mark it as a particle group.

4. The method of claim 3, wherein the single-beam lidar solid particle fabrication method is characterized by, The motion trajectory analysis in step S1 includes: Through target clustering tracking, the same target is tracked across frames, and the displacement direction angle is recorded; Computing motion continuity index ; wherein θ1 is the previous frame displacement direction included angle, θ2 is the current frame displacement direction included angle; If the motion continuity index η < motion threshold, it is determined as random motion, and marked as a particle group.

5. The method of claim 4, wherein the single-beam lidar solid particle fabrication method is characterized by, In step S2, the reflection intensity dispersion σ norm characteristic value normalization of the point cloud density variation rate Δρ and the motion discontinuity (1-η): Particle probability calculation: where w1, w2, and w3 are preset weights. The determination rules include: If P ≤ 0.4, directly mark it as a real obstacle; If 0.4 < P ≤ 0.7, trigger cross-frame verification, if the P value of 3 consecutive frames is < 0.7, mark it as a real obstacle; If P > 0.7, directly mark it as a particle group.

6. The method of claim 5, wherein the single-beam lidar solid particle fabrication method is characterized by, The dynamic parameter adjustment of step S3 includes: Computing the particle concentration C 颗粒 = (number of particle marker points / total number of point cloud points) x 100%; If C 颗粒 > 30%, switch to a relaxed mode, reduce the motion threshold of the motion continuity index η; and return to step S1 to re-analyze the motion trajectory.

7. The method of claim 6, wherein the single-beam lidar solid particle fabrication method is characterized by, The false alarm suppression of step S4 includes: Remove the point cloud marked as a particle group: delete the point cloud marked as a particle group from the point cloud set; Output real obstacle data, including real obstacle coordinates and motion vector.

Images