US-12624942-B2 - Apparatus and method for quantifying the surface flatness of three-dimensional point cloud data

US12624942B2US 12624942 B2US12624942 B2US 12624942B2US-12624942-B2

Abstract

A method that quantifies the surface flatness of 3D point cloud data in which a test statistic is proposed to indicate the surface flatness based on the threshold of the allowed bump level, the confidence level of test statistics and data density. The method comprises steps of converting the LIDAR measured points to coordinates along the axes using the principal component analysis (PCA) technique; calculating a Z α value based on the coordinates and predetermined bump tolerance: comparing the Z α value with a Z score of a test statistic to perform a null hypothesis; and rejecting the null hypothesis when the Z α value is greater than the Z score.

Inventors

Hok Chuen CHENG
Chun Hei CHAN
Wang Kong LAM
Winston Sun
Kei Hin NG

Assignees

AIPhotonics Limited

Dates

Publication Date: 20260512
Application Date: 20201005
Priority Date: 20191003

Claims (8)

1 . A Light Detection and Ranging (LiDAR) system for determining surface flatness, comprising: a laser configured for generating laser light beam; a scanner configured for scanning the laser light beam along a three-dimensional (3D) target surface; a photodetector configured for obtaining measured points from detected reflected light from the target surface, wherein the measured points form a point cloud; a processor configured for: converting the point cloud to target surface sample coordinates along coordinate axes according to length, width, and thickness of the target surface using principal component analysis (PCA); computing a local mean of thickness of the target surface sample coordinates; computing a global mean of thickness of the target surface sample coordinates; computing a global standard deviation of thickness of the target surface sample coordinates; computing a Z α value from the local mean of thickness of the target surface sample coordinates, the global mean of thickness of the target surface sample coordinates, the global standard deviation of thickness of the target surface sample coordinates, number of local samples of the target surface sample coordinates, and a predetermined bump tolerance; comparing the Z α value with a Z score of a test statistic, wherein the Z score is a value that corresponds to a confidence level; and determining that a bump larger than the bump tolerance exists on the target surface and that surface flatness of the target surface is outside of the bump tolerance if the Z α value is greater than the Z score; producing three-dimensional measurements of the target surface based on the Z α value, thereby quantifying surface flatness of the target surface.
2 . The LiDAR system of claim 1 , wherein the predetermined bump tolerance is in a range of 0.5 to 1.5 centimeters.
3 . The LiDAR system of claim 1 , wherein the test statistic is determined by a null hypothesis of a one tail test that states that the surface flatness of the target surface is within the predetermined bump tolerance.
4 . The LiDAR system of claim 1 , wherein a predetermined target surface with a known bump size is used to determine the predetermined bump tolerance.
5 . The LiDAR system of claim 1 wherein the scanner is selected from a mirror, a polygonal mirror, or a MEMS device.
6 . The LiDAR system of in claim 1 , further comprising performing one or more calibrations for one or more target surfaces with different incident angles, ranges, texture and refractivity to correct detection distortion.
7 . The LiDAR system of claim 1 wherein the photodetector is selected from a silicon avalanche photodiode, a photomultiplier, a charge-couple device (CCD), or a complementary metal-oxide-semiconductor (CMOS) device.
8 . The LiDAR system of claim 1 , wherein the Z α value has the following relation: Z α = | r _ local - r _ global | - d σ r , global / N local ; and wherein r local is the local mean of thickness of the target surface sample coordinates, r global is the global mean of thickness of the target surface sample coordinates, σ r,global is the global standard deviation of thickness of the target surface sample coordinates, N local is a number of local samples of the target surface sample coordinates, and d is the predetermined bump tolerance.

Description

FIELD OF THE INVENTION The present invention relates to three-dimensional (3D) point cloud processing, especially to methods and apparatus for quantifying the surface flatness of a scanned object using 3D point cloud data. BACKGROUND OF THE INVENTION Light detection and ranging (LIDAR) is an optical remote sensing technique that densely samples the surfaces of sensing targets. LIDAR usually employs an active optical sensor that transmits laser beams toward the target while moving through specific survey routes. The reflection of the laser from the target is detected and analyzed by receivers in the LIDAR sensor. LIDAR apparatus typically include a laser source and a scanner that directs the laser source in different directions towards a target to be imaged. Steering of the laser beam may be performed using a rotating material, microelectromechanical systems (MEMS), solid state scanning using silicon photonics, or other devices such as a Risley prism. The incident light is reflected from the target being scanned. The received reflections form a three-dimensional (3D) point cloud of data. The data can be used in many applications, such as building reconstruction and road-marking extraction. Normal estimation is a fundamental task in 3D point cloud processing. Known normal estimation methods can be classified into regression-based methods, Vorono-based methods and deep-learning methods. The regression-based method assumes the surface of an object is smooth all around, and thus the local neighborhood of any point on the surface can be well-approximated by a plane. In general, the principal component analysis (PCA) involves a covariance matrix computation of the neighborhood points, and then organizing the information in principle components. This method is widely used because it is easy to implement and quick to perform. However, the distorted point cloud data collected by the LiDAR scanner are smeared out with a standard deviation of 6-8 mm in the range measurement. Further, as PCA is an orthogonal linear transformation, it cannot smooth out sharp features in the data. Accordingly, for different applications and purposes, many techniques were presented to improve the robustness of the method. However, the techniques often involve a nontrivial trial-and-error process in order to obtain satisfactory results. The manual selection of parameters involved is also time consuming. SUMMARY OF THE INVENTION An objective of the present invention is to provide an unbiased estimator that quantifies the bumps of a surface, such as a wall, ceiling and floor of a three-dimensional (3D) point cloud. In accordance to one aspect of the present invention, a proposed estimation of the surface flatness is provided based on the threshold of the bump level, the confidence level of test statistics and data density. According to one embodiment of the present invention, the method comprises a conversion of the LIDAR measured points to coordinates using the principal component analysis (PCA) technique; a calculation of a Zα value based on the coordinates and predetermined bump tolerance; comparing the Zα value with a Z score of a test statistic to perform a null hypothesis; and rejecting the null hypothesis when the Zα value is greater than a Z score. The calculation of the Zα value can be defined by the following relationship: ❘"\[LeftBracketingBar]"r_local-r_global❘"\[RightBracketingBar]"-dσr,global/Nlocal; wherein rlocal is a local mean of coordinates, rglobal is a global mean of coordinates, σr,global is a global standard deviation of coordinates, Nlocal is the number of local sample events, and d is the predetermined bump tolerance. Accordingly, the present invention is able to quantify the surface flatness easily by using the converted coordinates and the given bump tolerance. The results of the test statistics can be an indicator for local bumps of 3D cloud point data. BRIEF DESCRIPTION OF THE DRAWINGS The invention is presented in more details using implementation examples of the drawings below. In the attached drawings: FIG. 1 depicts a LIDAR system for quantifying surface flatness according to one aspect; FIG. 2 is a flow chart illustrating a method that quantifies the surface flatness of 3D point cloud data in accordance with an embodiment of the present invention; FIG. 3 is an exemplary diagram illustrating a flatness of a ceiling line where the wall meets the ceiling; FIG. 4 is an exemplary diagram illustrating a Null hypothesis of a one tail (right) test with a bell-shaped curve; FIG. 5 is an exemplary diagram illustrating a LIDAR system being disposed to collect data points that represents a three-dimensional shape in a room; FIGS. 6A-6C are exemplary diagrams illustrating the conversion of the LIDAR points return to coordinates along the coordinate axes; FIGS. 7A and 7B are two-dimensional (2D) plot diagrams illustrating the distributions of the values by performing the method to the coordinates of FIG. 6A in accordance w