CN106097423A - LiDAR point cloud intensity correction method based on k neighbour - Google Patents

Abstract

The invention discloses a kind of based onkThe LiDAR point cloud intensity correction method of neighbour, withkBased on nearest neighbor algorithm, by asking forkThe strength mean value method of Neighbor Points carries out a correction for cloud intensity level, and algorithm structure is very simple, and can effectively correct the strength information of a cloud so that the strength information after correction can the attribute information of the most real reflection surface.

Description

LiDAR point cloud intensity correction method based on k neighbour

Technical field

The present invention relates to a kind of LiDAR point cloud intensity correction method, particularly to a kind of LiDAR point cloud based on k neighbour Intensity correction method.

Background technology

The luminous power that laser intensity is object to the backscattering echo launching laser, echo-signal received after through inside Conversion and amplification, the intensity level in final conversion raw LiDAR data.Owing to by laser ranging value and laser light incident angle etc. being The impact of the system target variable such as variable and target reflectivity, roughness and gradient so that the intensity level of acquisition exists certain Deviation.Mostly existing method is to first pass through the functional relationship setting up between intensity level and system variable, and then analysis of system variables shadow The method rung and carry out correcting obtains the attribute information of target surface.Due to manufacturer's secrecy to key parameter, and setting up Simplification to some variable during functional relationship so that the functional relationship of foundation can not accurately eliminate system variable Impact.

Summary of the invention

The technical problem to be solved is to provide a kind of LiDAR point cloud intensity correction method based on k neighbour, makes LiDAR point cloud strength information after must correcting can the attribute information of the most accurate and real reflection surface.

The present invention solves above-mentioned technical problem by the following technical solutions:

The present invention provides a kind of LiDAR point cloud intensity correction method based on k neighbour, including step in detail below:

Step 1, reads the LiDAR point cloud data comprising strength information, obtains number n of cloud data；

Step 2, uses all of some cloud point of K-D tree tissue, calculates the normal vector of each some cloud point simultaneously；

Step 3, travels through all of some cloud point, calculates angle of incidence and the Neighbor Points thereof of each some cloud point, it is thus achieved that Mei Gedian Intensity level after cloud point correction, specifically comprises the following steps that

301: make i=1；

302: utilize formula 1 to calculate i-th point cloud point p_iAnd the distance between scanner center:

$R_i=\sqrt{{x_i}^2+{y_i}^2+{z_i}^2}---\left(1\right)$

In formula, R_iFor i-th point cloud point p_iAnd the Euclidean distance between scanner center, and i ∈ (1,2 ..., n), (x_i,y_i, z_i) it is i-th point cloud point p in the coordinate system with scanner center as initial point_iOriginal coordinates；

303: utilize formula 2 to calculate i-th point cloud point p_iLaser light incident angle θ_i:

${\mathrm{\θ}}_i=acos\frac{(x_i,y_i,z_i)\·({\mathrm{nx}}_i,{\mathrm{ny}}_i,{\mathrm{nz}}_i)}{R_i}---\left(2\right)$

In formula, θ_iFor i-th point cloud point p_iLaser light incident angle, (nx_i,ny_i,nz_i) it is i-th point cloud point p_iNormal direction Amount；

304: search i-th point cloud point p_iK Neighbor Points, calculate the average of the raw intensity values of this k Neighbor Points, should Average is i-th point cloud point p_iIntensity level after correction；

305: make i=i+1, if i is ＜ n, then returns step 302, otherwise terminate iteration, thus complete a cloud point Intensity correction.

As the further prioritization scheme of the present invention, in step 304, the value of k is chosen according to actual needs.

As the further prioritization scheme of the present invention, the value of k is 10.

As the further prioritization scheme of the present invention, the value of k is 15.

As the further prioritization scheme of the present invention, the value of k is 20.

The present invention uses above technical scheme compared with prior art, has following technical effect that what the present invention provided LiDAR point cloud strength information bearing calibration, based on k nearest neighbor algorithm, is entered by the strength mean value method asking for k Neighbor Points The correction of row point cloud intensity level, algorithm structure is very simple, and can effectively correct the strength information of a cloud so that after correction Strength information can the attribute information of the most real reflection surface.

Accompanying drawing explanation

Fig. 1 is the flow chart of the present invention.

Fig. 2 be withFor transverse axis, original LiDAR point cloud intensity level I is the scatterplot of the longitudinal axis.

Fig. 3 be withFor transverse axis, after correction, LiDAR point cloud intensity level I is the scatterplot of the longitudinal axis.

Detailed description of the invention

Below in conjunction with the accompanying drawings technical scheme is described in further detail:

As it is shown in figure 1, a kind of LiDAR point cloud intensity correction method based on k neighbour, comprise the following steps:

Step 1, reads the LiDAR point cloud data comprising strength information, obtains number n=5296 of cloud data.

Step 2, uses all of some cloud point of K-D tree tissue, calculates the normal vector of each some cloud point simultaneously.

Step 3, travels through all of some cloud point, calculates angle of incidence and the Neighbor Points thereof of each some cloud point, it is thus achieved that Mei Gedian Intensity level after cloud point correction, specifically comprises the following steps that

401: make i=1；

402: utilize formula 1 to calculate i-th point cloud point p_iAnd the distance between scanner center:

$R_i=\sqrt{{x_i}^2+{y_i}^2+{z_i}^2}---\left(1\right)$

In formula, R_iFor i-th point cloud point p_iAnd the Euclidean distance between scanner center, and i ∈ (1,2 ..., n), (x_i,y_i, z_i) it is i-th point cloud point p in the coordinate system with scanner center as initial point_iCoordinate；

403: utilize formula 2 to calculate i-th point cloud point p_iLaser light incident angle θ_i:

${\mathrm{\θ}}_i=acos\frac{(x_i,y_i,z_i)\·({\mathrm{nx}}_i,{\mathrm{ny}}_i,{\mathrm{nz}}_i)}{R_i}---\left(2\right)$

In formula, θ_iFor i-th point cloud point p_iLaser light incident angle, (nx_i,ny_i,nz_i) it is i-th point cloud point p_iNormal direction Amount；

404: search i-th point cloud point p_iK Neighbor Points, calculate the average of the observed strength value of this k Neighbor Points, should Average is i-th point cloud point p_iIntensity level after correction；

405: make i=i+1, if i is ＜ n, then returns step 402, otherwise terminate iteration, thus complete a cloud point Intensity correction.

Follow technique scheme of the present invention, the intensity of the laser point cloud after obtaining original laser point cloud intensity level and correcting Value, as shown in Figures 2 and 3, wherein transverse axis isThe longitudinal axis is the intensity level of laser spots cloud point.

As in figure 2 it is shown, be original laser point cloud point intensity scatterplot.As it is shown on figure 3, for the present invention obtain correction after Laser point cloud intensity scatterplot.From figures 2 and 3, it will be seen that the laser spots intensity scatterplot of present invention acquisition and original laser Point cloud intensity scatterplot general trend is the most similar, but the intensity scatterplot after present invention correction is distributed the distribution of relatively green strength scatterplot More concentrate, illustrate the laser point cloud intensity after present invention correction can not only the attribute information of reflection surface, and more connect It is bordering on the actual strength value of body surface.

The above, the only detailed description of the invention in the present invention, but protection scope of the present invention is not limited thereto, and appoints What is familiar with the people of this technology in the technical scope that disclosed herein, it will be appreciated that the conversion expected or replacement, all should contain Within the scope of the comprising of the present invention, therefore, protection scope of the present invention should be as the criterion with the protection domain of claims.

Claims (5)

1. LiDAR point cloud intensity correction method based on k neighbour, it is characterised in that include step in detail below:

Step 1, reads the LiDAR point cloud data comprising strength information, obtains number n of cloud data；

Step 2, uses all of some cloud point of K-D tree tissue, calculates the normal vector of each some cloud point simultaneously；

Step 3, travels through all of some cloud point, calculates angle of incidence and the Neighbor Points thereof of each some cloud point, it is thus achieved that each some cloud point Intensity level after correction, specifically comprises the following steps that

301: make i=1；

302: utilize formula 1 to calculate i-th point cloud point p_iAnd the distance between scanner center:

$R_i=\sqrt{{x_i}^2+{y_i}^2+{z_i}^2}---\left(1\right)$

In formula, R_iFor i-th point cloud point p_iAnd the Euclidean distance between scanner center, and i ∈ (1,2 ..., n), (x_i,y_i,z_i) it is I-th point cloud point p in the coordinate system with scanner center as initial point_iOriginal coordinates；

303: utilize formula 2 to calculate i-th point cloud point p_iLaser light incident angle θ_i:

${\mathrm{\θ}}_i=acos\frac{(x_i,y_i,z_i)\·({\mathrm{nx}}_i,{\mathrm{ny}}_i,{\mathrm{nz}}_i)}{R_i}---\left(2\right)$

In formula, θ_iFor i-th point cloud point p_iLaser light incident angle, (nx_i,ny_i,nz_i) it is i-th point cloud point p_iNormal vector；

304: search i-th point cloud point p_iK Neighbor Points, calculate the average of the raw intensity values of this k Neighbor Points, this average It is i-th point cloud point p_iIntensity level after correction；

305: make i=i+1, if i is ＜ n, then returns step 302, otherwise terminate iteration, thus complete the intensity of a cloud point Correction.

LiDAR point cloud intensity correction method based on k neighbour the most according to claim 1, it is characterised in that step 304 The value of middle k is chosen according to actual needs.

LiDAR point cloud intensity correction method based on k neighbour the most according to claim 2, it is characterised in that the value of k is 10。

LiDAR point cloud intensity correction method based on k neighbour the most according to claim 2, it is characterised in that the value of k is 15。

LiDAR point cloud intensity correction method based on k neighbour the most according to claim 2, it is characterised in that the value of k is 20。