CN100443856C - A flat leaf area nondestructive measuring system and method thereof - Google Patents

Abstract

The invention relates to the field of plane area measurement technology, in particular to the non-destructive measurement system and method of the plane blade area for a single image. The system comprises: image acquisition unit, image correction unit, blade image division unit, blade area calculation unit. The method includes: the flat blade image which is attached to the demarcating board can be gained from the image acquisition unit; the projective corresponding relationship between the image element and the object can be calculated based on the isonym point of the image and the demarcating board by the image correction unit, the orthographic correction of the image is also done; the image of the blade can be separated from the background by the blade image division unit; the area of the blade image which is after the orthographic correction can be computed based on the projective corresponding relationship by the blade area calculation unit.

Description

A kind of flat leaf area nondestructive measuring system and method thereof

Technical field

The present invention relates to area of plane field of measuring technique, particularly a kind of use single image carries out flat leaf area nondestructive measuring system and method thereof.

Background technology

Blade area is one of common counter of arable farming and breeding production operation, also is the important evaluation index of crop yield and quality.Measurement about blade area is one of focus in the agriculture field always.Existing conventional leaf area measurement method, comprise and utilize check board to measure the leaf area method, utilize leaf weight to estimate the leaf area method, utilize regression equation calculation leaf area method, exist respectively complicated operation time-consuming, destroy blade and the big shortcoming of individual blade measuring error.And the leaf area instrument that utilizes import is measured leaf area, though simple to operate quick, the result is accurate, costs an arm and a leg, and is not suitable for generally applying.

A kind of alternative method is to use scanner that blade is scanned, and calculates blade area then, but this method can only be carried out the destructiveness measurement.

Summary of the invention

The objective of the invention is to not have under the condition of scanner of special expensive, accurately and rapidly individual blade is carried out area measurement and analysis, thereby obtain relevant correlation parameters such as blade area, blade girth, length of blade, width of blade and blade length breadth ratio.Purpose of the present invention is intended to provide a low price, simple to operate, measurement result blade area measuring system accurately for the user.

The present invention uses the printed circuit board (PCB) (PCB) of design voluntarily as scaling board, by single image, handles, thereby calculates blade area.These can bring the step of error not need camera Calibration and images match.Because system constitutes used parts, make that its cost is controlled simultaneously.When guaranteeing cost, system can reach 0.5% duplicate measurements precision, has satisfied most measurement needs.

$\left(\begin{array}{cccccccc}{X}_u^{\mathrm{Ob}}& X_v^{\mathrm{<figure-callout\; id="1"\; label="Ob"\; filenames="C20051008664700131.png"\; state="\{\{state\}\}">Ob</figure-callout>}}& 1& 0& 0& 0& -X_u^{\mathrm{Ob}}\&CenterDot;x_i^{\mathrm{Im}}& -X_v^{\mathrm{Ob}}\&CenterDot;x_i^{\mathrm{Im}}\\ 0& 0& 0& X_u^{\mathrm{Ob}}& X_v^{\mathrm{Ob}}& 1& -x_j^{\mathrm{Im}}\&CenterDot;X_u^{\mathrm{Ob}}& -x_j^{\mathrm{Im}}\&CenterDot;X_v^{\mathrm{Ob}}\end{array}\right)\&CenterDot;\left(\begin{array}{c}{m}_{11}\\ m_{12}\\ m_{13}\\ m_{21}\\ m_{22}\\ m_{23}\\ m_{31}\\ m_{32}\end{array}\right)=\left(\begin{array}{c}{x}_i^{\mathrm{Im}}\\ x_j^{\mathrm{Im}}\end{array}\right)$

Image correction unit is calculated the projection matrix formula.(x wherein _u ^Ob, x _v ^Ob) object coordinates, (x _i ^Im, x _j ^Im) be corresponding pixel coordinate, formula has provided the projective rejection between these two points.Provide 4 (x _u ^Ob, x _v ^Ob), (x _i ^Im, x _j ^Im) right, just can find the solution (m ₁₁, m ₁₂, m ₁₃, m ₂₁, m ₂₂, m ₂₃, m ₃₁, m ₃₂).Consider the factor of error, the user can provide more than 4 (x _u ^Ob, x _v ^Ob), (x _i ^Im, x _j ^Im) right, image correction unit can adopt the SVD method to obtain optimum (m ₁₁, m ₁₂, m ₁₃, m ₂₁, m ₂₂, m ₂₃, m ₃₁, m ₃₂).So it is user profile that the user will set up, provide (x in therein _u ^Ob, x _v ^Ob), and in image, choose corresponding (x _i ^Im, x _j ^Im).

$x_i^{\mathrm{Im}}=\frac{m_{11}\&CenterDot;X_u^{\mathrm{Ob}}+m_{12}\&CenterDot;X_v^{\mathrm{Ob}}+m_{13}}{m_{31}\&CenterDot;X_u^{\mathrm{Ob}}+m_{32}\&CenterDot;X_v^{\mathrm{Ob}}+1}$

$x_j^{\mathrm{Im}}=\frac{m_{21}\&CenterDot;X_u^{\mathrm{Ob}}+m_{22}\&CenterDot;X_v^{\mathrm{Ob}}+m_{23}}{m_{31}\&CenterDot;X_u^{\mathrm{Ob}}+m_{32}\&CenterDot;X_v^{\mathrm{Ob}}+1}$

The relation of image coordinate and object coordinates, wherein, (x _i ^Im, x _j ^Im) be pixel coordinate, (x _u ^Ob, x _v ^Ob) be object coordinates, (m ₁₁, m ₁₂, m ₁₃, m ₂₁, m ₂₂, m ₂₃, m ₃₁, m ₃₂) be 8 constants that image correction unit calculates.For each pixel on the image, this formula has provided its true coordinate on the coordinate system on the scaling board.

Technical scheme

The objective of the invention is to be achieved through the following technical solutions:

This system comprises image acquisition units, image correction unit, leaf image cutting unit, blade area computing unit.

Obtain not having specific requirement for the digital camera of taking pictures by image collecting device attached to the plane blade image on the scaling board.

Image correction apparatus is according to the same place on image and the scaling board, calculates the pixel on the image and the projection corresponding relation of actual object.

The leaf image segmenting device is separated the image of blade from background.

The blade area calculation element is according to the projection corresponding relation, calculates the area of the leaf image after the correction.

Image capture module is made up of scaling board, camera, camera Control Software module, image correction unit, leaf image cutting unit, blade area computing unit are software module, be integrated in a unit, wherein, image acquisition units, image correction unit, leaf image cutting unit, the blade area computing unit, said units is linked in sequence successively.

A kind of flat leaf area nondestructive measuring method, its step is as follows:

Obtain attached to the plane blade image on the scaling board by image acquisition units;

Image correction unit is according to the same place on image and the scaling board, calculates the pixel on the image and the projection corresponding relation of actual object;

The leaf image cutting unit is separated the image of blade from background;

The blade area computing unit is according to the projection corresponding relation, calculates the area of the leaf image behind the ortho-rectification.

Description of drawings

Fig. 1 is scaling board and captured photo figure.

Fig. 2 is flat leaf area nondestructive measuring system figure of the present invention.

Fig. 3 is a flat leaf area nondestructive measuring method flow diagram of the present invention.

Embodiment

Fig. 1 is scaling board and captured photo, is that blade to be measured is attached on the scaling board among the figure.

Fig. 2 flat leaf area nondestructive measuring system comprises image collecting device, image correction apparatus, leaf image segmenting device, blade area calculation element.Wherein, image acquisition units, image correction unit, the leaf image cutting unit, the blade area computing unit, said units is linked in sequence successively.

The specific operation process of Fig. 3 flat leaf area nondestructive measuring method is:

(1) with blade to be measured attached on the scaling board, flatten, use digital camera to take a width of cloth scaling board image;

(2) user sets up the object coordinates file corresponding with the image same place;

(3) load map picture and corresponding same place file;

(4) user carries out cutting apart of leaf image by 3 kinds of methods

RGB (red, green, blue) color space is cut apart: the user specifies sub pixel and tolerance, and the leaf image cutting unit can be automatically from sub pixel, close with the sub pixel color and topological on have neighbouring relations pixel separation come out;

HSI (colourity, saturation degree, brightness) color space is cut apart: method is cut apart with rgb color space, and just the constituency of color space is the HSI space.

Polygon is cut apart: the user uses the polygon frame to live leaf image, and the leaf image cutting unit can be filled polygon inside automatically, is partitioned into leaf image.

(5) user carries out the image ortho-rectification by user interface of software;

(6) user carries out blade area and calculates by user interface of software.

The flat leaf area nondestructive measuring method can adopt three kinds of diverse ways and combination thereof to be partitioned into leaf image, thereby obtain accurate blade area result.

The invention characteristics

Native system is by single image, carries out the system of nondestructive measurement for flat leaf area, is characterized in:

1, carry out flat leaf area by single image and measure, avoided camera Calibration and images match, Reduced error.

2, because the formation of system is not used expensive device, so the control of the hardware cost of whole system In 300 yuan.

3, use printed circuit board (PCB) and top silk screen layer thereof as the yardstick of demarcating, because the essence of himself Spend, guaranteed the precision of system.

4, adopt the Combination of Methods of multiple image segmentation to cut apart leaf image, guaranteed the precision of calculating.

5, with earlier calculating, the method for post-equalization, system has extensibility, not only can accurate Calculation put down The face blade area, and can on image, directly measure the geometric senses such as length, angle.

Claims (5)

1, a kind of flat leaf area nondestructive measuring system is characterized in that: this system comprises image acquisition units, image correction unit, and the leaf image cutting unit, the blade area computing unit,

Obtain attached to the plane blade image on the scaling board by image acquisition units,

Image correction unit is according to the same place on image and the scaling board, calculates the pixel on the image and the projection corresponding relation of actual object to be,

$\left(\begin{array}{cccccccc}{X}_u^{\mathrm{Ob}}& X_v^{\mathrm{Ob}}& 1& 0& 0& 0& -X_u^{\mathrm{Ob}}\&CenterDot;x_i^{\mathrm{Im}}& -X_v^{\mathrm{Ob}}\&CenterDot;x_i^{\mathrm{Im}}\\ 0& 0& 0& X_u^{\mathrm{Ob}}& X_v^{\mathrm{Ob}}& 1& -x_j^{\mathrm{Im}}\&CenterDot;X_u^{\mathrm{Ob}}& -x_j^{\mathrm{Im}}\&CenterDot;X_v^{\mathrm{Ob}}\end{array}\right)\&CenterDot;\left(\begin{array}{c}{m}_{11}\\ m_{12}\\ m_{13}\\ m_{21}\\ m_{22}\\ m_{23}\\ m_{31}\\ m_{32}\end{array}\right)=\left(\begin{array}{c}{x}_i^{\mathrm{Im}}\\ x_j^{\mathrm{Im}}\end{array}\right)$

(x wherein _u ^Ob, x _v ^Ob) be object coordinates, (x _i ^Im, x _j ^Im) be corresponding pixel coordinate, formula has provided the projective rejection between these two points; Provide 4 (x _u ^Ob, x _v ^Ob), (x _i ^Im, x _j ^Im) right, just can find the solution (m ₁₁, m ₁₂, m ₁₃, m ₂₁, m ₂₂, m ₂₃, m ₃₁, m ₃₂); Consider the factor of error, the user can provide more than 4 (x _u ^Ob, x _v ^Ob), (x _i ^Im, x _j ^Im) right, image correction unit can adopt the SVD method to obtain optimum (m ₁₁, m ₁₂, m ₁₃, m ₂₁, m ₂₂, m ₂₃, m ₃₁, m ₃₂); So it is user profile that the user will set up, and provides (x therein _u ^Ob, x _v ^Ob), and in image, choose corresponding (x _i ^Im, x _j ^Im);

The leaf image cutting unit is separated the image of blade from background,

The blade area computing unit is according to the projection corresponding relation, calculates the area of the leaf image behind the ortho-rectification,

Image capture module is made up of scaling board, camera, camera Control Software module, image correction unit, leaf image cutting unit, blade area computing unit are software module, be integrated in a unit, wherein, image acquisition units, image correction unit, leaf image cutting unit, the blade area computing unit, said units is linked in sequence successively.

2, a kind of flat leaf area nondestructive measuring method, its step is as follows:

Obtain attached to the plane blade image on the scaling board by image acquisition units;

Image correction unit is according to the same place on image and the scaling board, calculates the pixel on the image and the projection corresponding relation of actual object;

The leaf image cutting unit is separated the image of blade from background;

The blade area computing unit is according to the projection corresponding relation, calculates the area of the leaf image behind the ortho-rectification;

It is to calculate projection matrix to be that described image correction unit is calculated the projection corresponding relation:

$\left(\begin{array}{cccccccc}{X}_u^{\mathrm{Ob}}& X_v^{\mathrm{Ob}}& 1& 0& 0& 0& -X_u^{\mathrm{Ob}}\&CenterDot;x_i^{\mathrm{Im}}& -X_v^{\mathrm{Ob}}\&CenterDot;x_i^{\mathrm{Im}}\\ 0& 0& 0& X_u^{\mathrm{Ob}}& X_v^{\mathrm{Ob}}& 1& -x_j^{\mathrm{Im}}\&CenterDot;X_u^{\mathrm{Ob}}& -x_j^{\mathrm{Im}}\&CenterDot;X_v^{\mathrm{Ob}}\end{array}\right)\&CenterDot;\left(\begin{array}{c}{m}_{11}\\ m_{12}\\ m_{13}\\ m_{21}\\ m_{22}\\ m_{23}\\ m_{31}\\ m_{32}\end{array}\right)=\left(\begin{array}{c}{x}_i^{\mathrm{Im}}\\ x_j^{\mathrm{Im}}\end{array}\right)$

(x wherein _u ^Ob, x _v ^Ob) be object coordinates, (x _i ^Im, x _j ^Im) be corresponding pixel coordinate, formula has provided the projective rejection between these two points; Provide 4 (x _u ^Ob, x _v ^Ob), (x _i ^Im, x _j ^Im) to finding the solution (m ₁₁, m ₁₂, m ₁₃m ₂₁, m ₂₂, m ₂₃, m ₃₁, m ₃₂); Consider the factor of error, the user provides more than 4 (x _u ^Ob, x _v ^Ob), (x _i ^Im, x _j ^Im) right, image correction unit adopts the SVD method to obtain optimum (m ₁₁, m ₁₂, m ₁₃, m ₂₁, m ₂₂, m ₂₃, m ₃₁, m ₃₂); So it is user profile that the user will set up, and provides (x therein _u ^Ob, x _v ^Ob), and in image, choose corresponding (x _i ^Im, x _j ^Im).

3, according to the flat leaf area nondestructive measuring method of claim 2, its concrete steps are as follows:

(1) with blade to be measured attached on the scaling board, flatten, take a width of cloth scaling board image;

(2) user sets up the object coordinates file corresponding with the image same place;

(3) load map picture and corresponding same place file;

(4) user carries out cutting apart of leaf image by 3 kinds of methods;

(5) user carries out the image ortho-rectification by user interface of software;

(6) user carries out blade area and calculates by user interface of software;

The pass of described image coordinate and object coordinates is:

$x_i^{\mathrm{Im}}=\frac{m_{11}\&CenterDot;X_u^{\mathrm{Ob}}+m_{12}\&CenterDot;X_v^{\mathrm{Ob}}+m_{13}}{m_{31}\&CenterDot;X_u^{\mathrm{Ob}}+m_{32}\&CenterDot;X_v^{\mathrm{Ob}}+1}$

$x_j^{\mathrm{Im}}=\frac{m_{21}\&CenterDot;X_u^{\mathrm{Ob}}+m_{22}\&CenterDot;X_v^{\mathrm{Ob}}+m_{23}}{m_{31}\&CenterDot;X_u^{\mathrm{Ob}}+m_{32}\&CenterDot;X_v^{\mathrm{Ob}}+1}$

Wherein, (x _i ^Im, x _j ^Im) be pixel coordinate, (x _u ^Ob, x _v ^Ob) be object coordinates, (m ₁₁, m ₁₂, m ₁₃, m ₂₁, m ₂₂, m ₂₃, m ₃₁, m ₃₂) be 8 constants that image correction unit calculates; For each pixel on the image, image coordinate and object coordinates have provided its true coordinate on the coordinate system on the scaling board.

4, according to the flat leaf area nondestructive measuring method of claim 3, it is characterized in that, adopt three kinds of diverse ways and combination thereof to be partitioned into leaf image, thereby obtain accurate blade area result.

5, according to the flat leaf area nondestructive measuring method of claim 3 or 4, it is characterized in that, comprising the cutting apart of leaf image of step (4):

Rgb color space is cut apart: the user specifies sub pixel and tolerance, and the leaf image cutting unit can be automatically from sub pixel, close with the sub pixel color and topological on have neighbouring relations pixel separation come out;

The HSI color space is cut apart: method is cut apart with rgb color space, and just color space adopts the HSI space;

Polygon is cut apart: the user uses the polygon frame to live leaf image, and the leaf image cutting unit can be filled polygon inside automatically, is partitioned into leaf image.

Images