CN104268935A - Feature-based airborne laser point cloud and image data fusion system and method - Google Patents

Abstract

The invention discloses a feature-based airborne laser point cloud and image data fusion method. The method sequentially comprises the steps of conducting SIFT image stitching on image data acquired by an airborne camera; determining feature points, corresponding to image data, in point cloud data in a scanning area; establishing the corresponding relationship between the line and row numbers of a three-dimensional point cloud and line and the row numbers of pixel points in a two-dimensional stitched image plane by means of the direct linear transformation method in the photogrammetry field; finally applying RGB color information of the pixel points to every point in the point cloud so that fusion of the two kinds of data can be realized. The feature-based airborne laser point cloud and image data fusion system and method are easy to realize, high in fusion effect and fusion accuracy, capable of well achieving point cloud mapping in a small unmanned helicopter laser scanning system, high in flexibility, and adaptable to the situation that a camera and a laser scanner acquire data separately.

Description

A kind of airborne laser point cloud of feature based and image data fusion system and method

Technical field

The present invention relates to geographical three dimensions modeling field, particularly a kind of airborne laser point cloud of feature based and image data fusion system and method.

Background technology

Airborne laser radar (LIDAR) measuring technique is a kind of emerging geospatial information acquiring technology.Along with improving constantly of hardware device integrated level, in many LIDAR systems, be all integrated with CCD (imageing sensor) digital camera, the image data of tested scene can be obtained while measuring acquisition cloud data.Can be obtained rapidly spatial information and the illumination intensity information of target by airborne lidar instrument, these information contribute to setting up accurate DTM model.Owing to being subject to the restriction of laser scanner working method, cloud data has discrete, the feature such as density differs, scrambling, and directly cannot obtain the superficial makings colouring information of spatial object, therefore utilize cloud data to identify destination object and feature extraction still has certain difficulty.Although the Reflection intensity information of some laser scanner receiving beam can be similar to the color of object, not real after all.

But the texture color information that high-resolution CCD digital camera can obtain body surface has the advantages that sampling density is high, directly perceived, easily distinguish object detail.These characteristics and discrete three-dimensional laser point cloud just in time define complementation; the CCD image information utilizing digital camera to take and laser point cloud carry out fusion treatment; can obtain realistic model of place well, the work such as this three-dimensional reconstruction for point cloud model, City Modeling, coastal zone protection has great importance.

Domestic at present some are blank also also existing based on the laser point cloud of unmanned plane low-latitude flying and the integration program of CCD image data, and the image data adopted in most integration program is all obtained by integrated camera on the scanner.

For this situation, the present invention is according to the respective feature of airborne laser point cloud data and CCD image data, and the scheme devising a kind of feature based Point matching solves the fusion problem of a cloud and image data.

Summary of the invention

The object of the invention is to overcome the shortcoming of prior art and deficiency, a kind of airborne laser point cloud and image data fusion system and method for feature based are provided.

Object of the present invention is achieved through the following technical solutions:

The airborne laser point cloud of feature based and an image data fusion system, comprise the connected some cloud of order and image data acquiring module, image data concatenation module, direct linear transformation's algoritic module and some cloud pinup picture module; Wherein

Point cloud and image data acquiring module, gather cloud data and image data;

Image data concatenation module, utilizes SIFT merging algorithm for images that the image photograph of collection is carried out splicing and obtains a width panoramic pictures;

Direct linear transformation's algoritic module, utilizes the collinearity equation in photogrammetric Theory to set up transformation relation between three-dimensional point cloud and bidimensional image data;

Point cloud pinup picture module, after establishing this transformation relation, is assigned to each point in a cloud, achieves the fusion of two kinds of data by the RGB colouring information of pixel.

Described some cloud and image data acquiring module comprise cloud data acquisition module, image data acquiring module, and wherein said cloud data acquisition module is made up of carrying platform, dynamic difference gps system, inertial navigation system INS, laser distance measuring system,

Carrying platform is helicopter or fixed wing aircraft;

Dynamic difference gps system is made up of GPS base station and GPS movement station, GPS base station is fixed on somewhere, ground, then locator data is transmitted to for receiving the positioning signal of launching from gps satellite the GPS movement station be placed on carrying platform, obtain the positional information of carrying platform thus, inertial navigation system INS, laser distance measuring system are installed on carrying platform, therefore determine the locus coordinate of laser signal transmitted-reference point;

Inertial navigation system INS utilizes the inertial measurement component such as gyroscope and accelerometer to experience the rotational transform of carrying platform in flight course, and through integral operation, obtain real-time attitude information, comprise roll angle and the angle of pitch, crab angle is obtained by electronic compass;

Laser distance measuring system can transmit and receive laser pulse, calculates the distance between scanner and target object by the mistiming of measuring transmitted beam and receiving beam; Can also provide intensity signal, these information finally can be used for realizing intensity imaging simultaneously;

The range data of laser footpoint is obtained by laser distance measuring system, in conjunction with the attitude information that spatial positional information and the inertial navigation system of the carrying platform exported by dynamic difference gps system provide, can go out the three-dimensional coordinate of laser footpoint under local horizontal coordinates NED by Combined Calculation, the set of these coordinate points puts cloud exactly.

Described image data acquiring module, by the boat sheet of the CCD digital camera shooting of device on mobile system, provides high-resolution image data, and these data can generate digital positive photograph model DOM product or merge for a cloud.

Another object of the present invention is achieved through the following technical solutions:

The airborne laser point cloud of feature based and an image data fusion method, comprise the step of following order:

S1. cloud data and image data is gathered;

S2. utilize SIFT merging algorithm for images that the image photograph of collection is carried out splicing and obtain a width panoramic pictures;

S3. the collinearity equation in photogrammetric Theory is utilized to set up transformation relation between three-dimensional point cloud and bidimensional image data;

S4., after establishing this transformation relation, the RGB colouring information of pixel is assigned to each point in a cloud, achieves the fusion of two kinds of data.

In step S2, described SIFT merging algorithm for images comprises following steps:

A) yardstick spatial extrema point is detected: the metric space of image under different scale can use gaussian kernel Convolution Formula L (x, y, σ)=G (x, y, σ) * I (x, y) representing, in order to effectively stable key point be detected at metric space, utilizing DoG operator to carry out local extremum detection to two width images in gray scale and yardstick two spaces; Operator definitions is the difference of the gaussian kernel of two different scales, is then accurately determined position and the yardstick of key point by concrete model, removes the key point of low contrast and skirt response instability;

B) distribute key point direction: the gradient direction distribution characteristic of vertex neighborhood pixel is each key point assigned direction parameter, ensure the rotational invariance of SIFT operator; Gradient is expressed as follows:

sample in neighborhood window centered by key point, and with the gradient direction of statistics with histogram neighborhood territory pixel; Histogrammic peak value represents the principal direction of this key point place neighborhood gradient, namely as the direction of this key point;

C) key point descriptor is generated: be first the direction of key point by X-axis rotate, to guarantee rotational invariance; In actual computation process, in order to strengthen the robustness of coupling, to each key point use 4 × 4 totally 16 Seed Points describe; Just can produce 128 data for a key point like this, form the SIFT feature vector of 128 dimensions; By the length normalization method of proper vector, then can remove the impact of illumination variation further, strengthen the antimierophonic ability of algorithm, good fault-tolerance be also provided for the characteristic matching containing positioning error simultaneously;

D) proper vector coupling: after generating the SIFT feature vector of two width images, adopts the Euclidean distance of key point proper vector as the similarity determination criterion of key point in two width images, is met the SIFT matching double points of criterion; Get certain key point in image, and European nearest the first two key point in finding out itself and another image, if nearest distance is less than certain threshold value except distance near in proper order, then accept this pair match point; Reduce this proportion threshold value, SIFT match point number can reduce, but more stable; Calculate the conversion parameter of image according to the SIFT matching double points obtained, carry out splicing fusion and obtain stitching image.

Described step S3, utilize the DLT algorithm in photogrammetric field, set up the corresponding relation of three-dimensional point cloud and bidimensional image panel data, wherein direct linear transformation's (Direct Linear Transformation is called for short DLT) refers to the algorithm of the linear relationship between the object space coordinate setting up pixel coordinate and corresponding object point, can derive direct linear transformation's algorithmic formula according to the collinearity equation in photogrammetric Theory:

$\left\{\begin{array}{c}x+\frac{l_{1}X+l_{2}X+l_{3}Z+l_4}{l_{9}X+l_{10}Y+l_{11}Z+1}=0\\ y+\frac{l_{5}X+l_{6}Y+l_{7}Z+l_8}{l_{9}X+l_{10}Y+l_{11}Z+1}=0\end{array}\right.,$ As can be seen from above-mentioned equation, direct linear transformation's algorithm can represent the relational expression between the object space coordinate points (X, Y, Z) that two-dimensional pixel coordinate points (x, y) is corresponding with it.

Described direct linear transformation's algorithmic formula resolve solving mainly to parameter l, wherein resolve l parameter be equivalent to photogrammetric in rear space intersection, resolve object space coordinate and be equivalent to forward intersection, there are 11 unknown parameters in DLT formula, for each culture point reference mark, its world coordinates known and can obtain two equations as planimetric coordinates, therefore will resolve these l parameters, needs 6 to such reference mark; In order to improve the precision of the solution of parameter l, introduce picpointed coordinate observed reading correction factor, its expression formula is as follows: $\left\{\begin{array}{c}x+v_x+\mathrm{\Δx}+\frac{l_{1}X+l_{2}X+l_{3}Z+l_4}{l_{9}X+l_{10}Y+l_{11}Z+1}=0\\ y+v_y+\mathrm{\Δy}+\frac{l_{5}X+l_{6}Y+l_{7}Z+l_8}{l_{9}X+l_{10}Y+l_{11}Z+1}=0\end{array}\right.,$

Represent the picture planimetric coordinates of reference mark in photograph; Then resolving of parameter l is completed by interative computation.

Described step S4, specifically comprises following steps:

(1) in picture, first choose the reference mark with obvious characteristic, these reference mark distribute relatively more even in the picture, and any 3 can not be laid in same plane, and preferably fluctuation ratio is comparatively large, is convenient to like this obtain stable calculation result; When there is no obvious unique point in target area, need by means of artificial reference mark of laying;

(2) D coordinates value (X, Y, Z) at several groups of reference mark next selected by record, and picture planimetric coordinates (u, v) of their picture points corresponding in the picture;

(3) from these points, choose the coordinate data at 6 pairs of reference mark, comprise a cloud coordinate (X, Y, Z) and picture planimetric coordinates (u, v), be updated in the matrix of DLT equation, utilize alternative manner to calculate 11 l parameter values;

(4) according to calculating l parameter, by coordinate figure (X, the Y of each point in a cloud, Z) this equation is substituted into, can obtain each some cloud corresponding to picture planimetric coordinates (u, v), according to the rgb pixel value just can taken out as planimetric coordinates in respective coordinates; Real rgb value is assigned to the point in corresponding point cloud again, this completes the pinup picture work of a cloud.

Compared with prior art, tool has the following advantages and beneficial effect in the present invention:

1, the invention provides that a kind of the method can be applicable to depopulated helicopter survey field based on the airborne laser point cloud of small-sized depopulated helicopter and the method for image co-registration, realize the fusion of three-dimensional point cloud and image, there is high-level efficiency, precision advantages of higher.

2, the method for the present invention's design can meet the situation that image capture device and laser scanner separate image data, has broken away from the problem that integrated camera precision is not high.Present invention employs the fusion that Direct Transform algoritic module comes process points cloud and image, without the need to calibration for cameras, shorten the cycle of fusion treatment.

3, method involved in the present invention can well be applied to the shaped area without obvious characteristic, such as wasteland, shore line etc., only needs the several unique point of artificial setting just can complete fusion treatment.

Accompanying drawing explanation

Fig. 1 is the airborne laser point cloud of a kind of feature based of the present invention and the structural representation of image data fusion system;

Fig. 2 is the product process figure of the airborne laser point cloud of a kind of feature based of the present invention and the cloud data of image data fusion method;

The process flow diagram that Fig. 3 is the airborne laser point cloud of a kind of feature based described in the present invention and the SIFT stitching algorithm of image data fusion method.

The process flow diagram of the some cloud pinup picture module of the airborne laser point cloud that Fig. 4 is a kind of feature based described in the present invention and image data fusion method.

Embodiment

Below in conjunction with embodiment and accompanying drawing, the present invention is described in further detail, but embodiments of the present invention are not limited thereto.

As shown in Figure 1, the airborne laser point cloud of feature based provided by the invention and image data fusion system comprise a cloud and image data acquiring module, image data concatenation module, direct linear transformation's algoritic module and some cloud pinup picture 4 modules; Described some cloud and image data acquiring module are mainly by airborne lidar instrument and camera sensor image data respectively; Described image data concatenation module is responsible for utilizing stitching algorithm the camera viewings of collection to be carried out splicing and is obtained a width panoramic pictures; Described direct linear transformation's algoritic module is responsible for utilizing the collinearity equation in photogrammetric Theory to set up transformation relation between three-dimensional point cloud and bidimensional image data; Described some cloud pinup picture module in charge sticks real colouring information to after establishing this transformation relation each point in some cloud.

As shown in Figure 2, cloud data generation module of the present invention is made up of carrying platform (aircraft), dynamic difference gps system, inertial navigation system (INS), laser distance measuring system.Carrying platform is used for loading laser radar range system.Dynamic difference GPS and differential Global Positioning System (Differential Global Position System, DGPS), it is made up of GPS base station and GPS movement station two parts, base station is generally fixed on somewhere, ground, be responsible for receiving the positioning signal of launching from gps satellite and then locator data be transmitted to movement station, movement station is then placed on body.Worked by both simultaneously and can determine the locus coordinate of laser signal transmitted-reference point.The attitude information that inertial navigation system INS (Inertial Navigation System) provides mobile system real-time, comprises roll angle (Roll) and the angle of pitch (Pitch).Crab angle (Yaw) is obtained by electronic compass.Laser scanner can transmit and receive laser pulse, calculates the distance between scanner and target object by the mistiming of measuring transmitted beam and receiving beam.

The range data of laser footpoint is obtained by laser scanner, in conjunction with the attitude information that spatial positional information and the inertial navigation system of the mobile system exported by DGPS provide, can go out the three-dimensional coordinate of laser footpoint under local horizontal coordinates (NED) by Combined Calculation, the set of these coordinate points puts cloud exactly.

As shown in Figure 3, the step of the SIFT merging algorithm for images in the present invention is as follows:

A) yardstick spatial extrema point is detected, the metric space of image under different scale gaussian kernel Convolution Formula L (x, y, σ)=G (x, y, σ) * I (x, y) represents, in order to effectively stable key point be detected at metric space, DoG operator is utilized to carry out local extremum detection to two width images in gray scale and yardstick two spaces.Operator definitions is the difference of the gaussian kernel of two different scales, is then accurately determined position and the yardstick of key point by concrete model, removes the key point of low contrast and skirt response instability.

B) distribute key point direction, the gradient direction distribution characteristic of vertex neighborhood pixel is each key point assigned direction parameter, ensures the rotational invariance of SIFT operator.Gradient is expressed as follows:

sample in neighborhood window centered by key point, and with the gradient direction of statistics with histogram neighborhood territory pixel.Histogrammic peak value represents the principal direction of this key point place neighborhood gradient, namely as the direction of this key point.

C) generating key point descriptor, is first the direction of key point by X-axis rotate, to guarantee rotational invariance.In actual computation process, in order to strengthen the robustness of coupling, we to each key point use 4 × 4 totally 16 Seed Points describe.Just can produce 128 data for a key point like this, form the SIFT feature vector of 128 dimensions.By the length normalization method of proper vector, then can remove the impact of illumination variation further, strengthen the antimierophonic ability of algorithm, good fault-tolerance be also provided for the characteristic matching containing positioning error simultaneously.

D) proper vector coupling, after generating the SIFT feature vector of two width images, adopts the Euclidean distance of key point proper vector as the similarity determination criterion of key point in two width images, is met the SIFT matching double points of criterion.Get certain key point in image, and European nearest the first two key point in finding out itself and another image, if nearest distance is less than certain threshold value except distance near in proper order, then accept this pair match point.Reduce this proportion threshold value, SIFT match point number can reduce, but more stable.Calculate the conversion parameter of image according to the SIFT matching double points obtained, carry out splicing fusion and obtain stitching image.

As shown in Figure 4, the key of the some cloud pinup picture in the present invention sets up the corresponding relation between some cloud and the pixel of spliced picture, the RGB color value of picture point is assigned to corresponding some cloud, thus makes a cloud have real colouring information.The specific implementation step of the some cloud pinup picture of feature based is as follows:

(1) in picture, first choose the reference mark with obvious characteristic, these reference mark generally distribute relatively more even in the picture, and any 3 can not be laid in same plane, and preferably fluctuation ratio is comparatively large, is convenient to like this obtain stable calculation result.When there is no obvious unique point in target area, need by means of artificial reference mark of laying.

(2) D coordinates value (X, Y, Z) at several groups of reference mark next selected by record, and picture planimetric coordinates (u, v) of their picture points corresponding in the picture.

(3) from these points, choose the coordinate data at 6 pairs of reference mark, comprise a cloud coordinate (X, Y, Z) and picture planimetric coordinates (u, v), be updated in the matrix of DLT equation, utilize alternative manner to calculate 11 l parameter values.

(4) according to calculating l parameter, by coordinate figure (X, the Y of each point in a cloud, Z) this equation is substituted into, can obtain each some cloud corresponding to picture planimetric coordinates (u, v), according to the rgb pixel value just can taken out as planimetric coordinates in respective coordinates.Real rgb value is assigned to the point in corresponding point cloud again, this completes the pinup picture work of a cloud.

Above-described embodiment is the present invention's preferably embodiment; but embodiments of the present invention are not restricted to the described embodiments; change, the modification done under other any does not deviate from Spirit Essence of the present invention and principle, substitute, combine, simplify; all should be the substitute mode of equivalence, be included within protection scope of the present invention.

Claims (8)

1. the airborne laser point cloud of feature based and an image data fusion system, is characterized in that: comprise some cloud that order is connected with image data acquiring module, image data concatenation module, direct linear transformation's algoritic module with put cloud pinup picture module; Wherein

Point cloud and image data acquiring module, gather cloud data and image data;

Image data concatenation module, utilizes SIFT merging algorithm for images that the image photograph of collection is carried out splicing and obtains a width panoramic pictures;

Direct linear transformation's algoritic module, utilizes the collinearity equation in photogrammetric Theory to set up transformation relation between three-dimensional point cloud and bidimensional image data;

Point cloud pinup picture module, after establishing this transformation relation, is assigned to each point in a cloud, achieves the fusion of two kinds of data by the RGB colouring information of pixel.

2. the airborne laser point cloud of feature based according to claim 1 and image data fusion system, it is characterized in that: described some cloud and image data acquiring module comprise cloud data acquisition module, image data acquiring module, wherein said cloud data acquisition module is made up of carrying platform, dynamic difference gps system, inertial navigation system INS, laser distance measuring system

Carrying platform is helicopter or fixed wing aircraft;

Dynamic difference gps system is made up of GPS base station and GPS movement station, GPS base station is fixed on somewhere, ground, then locator data is transmitted to for receiving the positioning signal of launching from gps satellite the GPS movement station be placed on carrying platform, obtain the positional information of carrying platform thus, inertial navigation system INS, laser distance measuring system are installed on carrying platform, therefore determine the locus coordinate of laser signal transmitted-reference point;

Inertial navigation system INS utilizes the inertial measurement component such as gyroscope and accelerometer to experience the rotational transform of carrying platform in flight course, and through integral operation, obtain real-time attitude information, comprise roll angle and the angle of pitch, crab angle is obtained by electronic compass;

Laser distance measuring system can transmit and receive laser pulse, calculates the distance between scanner and target object by the mistiming of measuring transmitted beam and receiving beam; Can also provide intensity signal, these information finally can be used for realizing intensity imaging simultaneously;

The range data of laser footpoint is obtained by laser distance measuring system, in conjunction with the attitude information that spatial positional information and the inertial navigation system of the carrying platform exported by dynamic difference gps system provide, can go out the three-dimensional coordinate of laser footpoint under local horizontal coordinates NED by Combined Calculation, the set of these coordinate points puts cloud exactly.

3. the airborne laser point cloud of feature based according to claim 2 and image data fusion system, it is characterized in that: described image data acquiring module, by the boat sheet of the CCD digital camera shooting of device on mobile system, there is provided high-resolution image data, these data can generate digital positive photograph model DOM product or merge for a cloud.

4. the airborne laser point cloud of feature based and an image data fusion method, is characterized in that, comprise the step of following order:

S1. cloud data and image data is gathered;

S2. utilize SIFT merging algorithm for images that the image photograph of collection is carried out splicing and obtain a width panoramic pictures;

S3. the collinearity equation in photogrammetric Theory is utilized to set up transformation relation between three-dimensional point cloud and bidimensional image data;

S4., after establishing this transformation relation, the RGB colouring information of pixel is assigned to each point in a cloud, achieves the fusion of two kinds of data.

5. the airborne laser point cloud of feature based according to claim 4 and image data fusion method, it is characterized in that, in step S2, described SIFT merging algorithm for images comprises following steps:

A) yardstick spatial extrema point is detected: the metric space of image under different scale can use gaussian kernel Convolution Formula L (x, y, σ)=G (x, y, σ) * I (x, y) represent, utilize DoG operator to carry out local extremum detection to two width images in gray scale and yardstick two spaces; Operator definitions is the difference of the gaussian kernel of two different scales, is then accurately determined position and the yardstick of key point by concrete model, removes the key point of low contrast and skirt response instability;

B) distribute key point direction: the gradient direction distribution characteristic of vertex neighborhood pixel is each key point assigned direction parameter, ensure the rotational invariance of SIFT operator; Gradient is expressed as follows:

sample in neighborhood window centered by key point, and with the gradient direction of statistics with histogram neighborhood territory pixel; Histogrammic peak value represents the principal direction of this key point place neighborhood gradient, namely as the direction of this key point;

C) key point descriptor is generated: be first the direction of key point by X-axis rotate, to guarantee rotational invariance; To each key point use 4 × 4 totally 16 Seed Points describe; Just can produce 128 data for a key point like this, form the SIFT feature vector of 128 dimensions; By the length normalization method of proper vector, then can remove the impact of illumination variation further, strengthen the antimierophonic ability of algorithm, good fault-tolerance be also provided for the characteristic matching containing positioning error simultaneously;

D) proper vector coupling: after generating the SIFT feature vector of two width images, adopts the Euclidean distance of key point proper vector as the similarity determination criterion of key point in two width images, is met the SIFT matching double points of criterion; Get certain key point in image, and European nearest the first two key point in finding out itself and another image, if nearest distance is less than certain threshold value except distance near in proper order, then accept this pair match point; Reduce this proportion threshold value, SIFT match point number can reduce, but more stable; Calculate the conversion parameter of image according to the SIFT matching double points obtained, carry out splicing fusion and obtain stitching image.

6. the airborne laser point cloud of feature based according to claim 5 and image data fusion method, it is characterized in that, described step S3, utilize the DLT algorithm in photogrammetric field, set up the corresponding relation of three-dimensional point cloud and bidimensional image panel data, wherein direct linear transformation refers to the algorithm of the linear relationship between the object space coordinate setting up pixel coordinate and corresponding object point, can derive direct linear transformation's algorithmic formula according to the collinearity equation in photogrammetric Theory: $\left\{\begin{array}{c}x+\frac{l_{1}X+l_{2}X+l_{3}Z+l_4}{l_{9}X+l_{10}Y+l_{11}Z+1}=0\\ y+\frac{l_{5}X+l_{6}Y+l_{7}Z+l_8}{l_{9}X+l_{10}Y+l_{11}Z+1}=0\end{array}\right.,$ As can be seen from above-mentioned equation, direct linear transformation's algorithm can represent the relational expression between the object space coordinate points (X, Y, Z) that two-dimensional pixel coordinate points (x, y) is corresponding with it.

7. the airborne laser point cloud of feature based according to claim 6 and image data fusion method, it is characterized in that, described direct linear transformation's algorithmic formula resolve solving mainly to parameter l, wherein resolve l parameter be equivalent to photogrammetric in rear space intersection, resolve object space coordinate and be equivalent to forward intersection, there are 11 unknown parameters in DLT formula, for each culture point reference mark, its world coordinates known and two equations can be obtained as planimetric coordinates, therefore these l parameters will be resolved, need 6 to such reference mark, in order to improve the precision of the solution of parameter l, introduce picpointed coordinate observed reading correction factor, its expression formula is as follows: $\left\{\begin{array}{c}x+v_x+\mathrm{\Δx}+\frac{l_{1}X+l_{2}X+l_{3}Z+l_4}{l_{9}X+l_{10}Y+l_{11}Z+1}=0\\ y+v_y+\mathrm{\Δy}+\frac{l_{5}X+l_{6}Y+l_{7}Z+l_8}{l_{9}X+l_{10}Y+l_{11}Z+1}=0\end{array}\right.,$

Represent the picture planimetric coordinates of reference mark in photograph; Then resolving of parameter l is completed by interative computation.

8. the airborne laser point cloud of feature based according to claim 7 and image data fusion method, it is characterized in that, described step S4, specifically comprises following steps:

(1) in picture, first choose the reference mark with obvious characteristic, these reference mark distribute relatively more even in the picture, and any 3 can not be laid in same plane; When there is no obvious unique point in target area, need by means of artificial reference mark of laying;

(2) D coordinates value (X, Y, Z) at several groups of reference mark next selected by record, and picture planimetric coordinates (u, v) of their picture points corresponding in the picture;

(3) from these points, choose the coordinate data at 6 pairs of reference mark, comprise a cloud coordinate (X, Y, Z) and picture planimetric coordinates (u, v), be updated in the matrix of DLT equation, utilize alternative manner to calculate 11 l parameter values;

(4) according to calculating l parameter, by coordinate figure (X, the Y of each point in a cloud, Z) this equation is substituted into, can obtain each some cloud corresponding to picture planimetric coordinates (u, v), according to the rgb pixel value just can taken out as planimetric coordinates in respective coordinates; Real rgb value is assigned to the point in corresponding point cloud again, this completes the pinup picture work of a cloud.