CN102663395B - A straight line detection method based on self-adaptation multi-scale fast discrete Beamlet transform - Google Patents

Abstract

The present invention belongs to the technical field of image processing, and particularly relates to a straight line detection method based on self-adaptation multi-scale fast discrete Beamlet transform. Technical features of the method rest with following details: an edge detection image E is obtained by carrying out edge detection on an image I using a Canny operator, gradient direction angles of edge pixel points on the edge detection image E are calculated and the edge contour is divided into various linearity regions based on a gradient direction angle of two edge points, and fast discrete Beamlet transform is carried out on a current self-adaptation partition subblock through linearity region screening to detect straight lines. The method reduces calculation amount by a certain degree because the scale of an AMFDBT algorithm initial blockette is in a self-adaptation set by the region where a target point is] in the image. At the same time, the initial Beamlet transform can easily cause straight lines to be truncated while the self-adaptive change in partition sub-block makes result of the detected straight line more complete with no truncation of straight lines.

Description

Line detection method based on self-adapting multi-dimension fast discrete Beamlet conversion

Technical field

The invention belongs to image and process, relate in particular to a kind of line detection method based on self-adapting multi-dimension fast discrete Beamlet conversion.

Background technology

The extraction of line feature is a study hotspot in image processing and area of pattern recognition always, although traditional line detection algorithm can both detect straight line quickly, length information and the starting point of line segment and accurate location that endpoint information can't realize line segment because line segment is not provided all.

The Beamlet conversion is one of effective tool of multi-scale geometric analysis, be used for recovering straight line from noisy image, curve and boxed area, it is easily realized the accurate location of line segment and the multiple dimensioned of line segment is similar to, on noisy linear feature detection and fitting a straight line, obvious advantage and potentiality is arranged.But the calculated amount of traditional B eamlet conversion is very large, and many improved fast algorithms are all wasted a lot of calculated amount because will carry out a large amount of repeated works.And advance square because the blockette size of traditional Beamlet conversion is two, if it is relatively loaded down with trivial details to be applied to general rectangular image.Simultaneously, because algorithm is to advance on the basis of recurrence to realize two, the situation that easily causes straight line to be truncated.

In sum, there are the deficiencies such as calculated amount is large, testing result is imperfect in traditional Beamlet conversion.

Summary of the invention

The technical matters that solves

For fear of the deficiencies in the prior art part, the present invention proposes a kind of line detection method based on self-adapting multi-dimension fast discrete Beamlet conversion.

Technical scheme

A kind of line detection method based on self-adapting multi-dimension fast discrete Beamlet conversion is characterized in that step is as follows:

Step 1: utilize the Canny operator to carry out rim detection to image I, obtain edge-detected image E;

Step 2: the deflection of edge calculation detected image E coboundary pixel gradient:

When the gradient direction angle of two marginal points differs by more than predetermined threshold value threshold, in this this marginal point place's disconnection, edge contour is divided into a plurality of ranges of linearity and gives mark; In formula, G _xAnd G _yThe horizontal and vertical gradient component that represents respectively marginal point;

Step 3: range of linearity inward flange is counted is less than the district's deletion of 20 line, obtain through the image E ' after line district's screening, and with the marginal point after rim detection as impact point;

Step 4: take edge image E ' as initial subregion sub-block, seek the Minimum Area of impact point existence as the adaptive partition sub-block; If this sub-block is more than or equal to defined threshold t ' time execution in step 5; Otherwise execution in step 6;

Step 5: current adaptive partition sub-block is carried out fast discrete Beamlet conversion, detection of straight lines.If find the beamlet straight line, record end points and the slope information of straight line, then wipe this straight line under current sub-block; Otherwise, with the current blockette quartern, and as the initial size of next yardstick subregion sub-block, repeated execution of steps 4;

Step 6: if the quantity of beamlet straight line greater than 0, according to the slope information of all straight lines, find straight slope poor 1 the degree with all interior straight lines as parallel lines.

Beneficial effect

A kind of line detection method based on self-adapting multi-dimension fast discrete Beamlet conversion that the present invention proposes, because the yardstick of AMFDBT algorithm primary partition piece is to be set by the region adaptivity that impact point in figure exists, in the situation that the most extreme impact point just can spread all over full figure, this makes calculated amount obtain certain reducing.Simultaneously, the situation that initial Beamlet conversion easily causes straight line to be truncated, and the adaptive change of subregion sub-block makes the result of final detection of straight lines more complete, phenomenon is blocked in difficult generation.

Description of drawings

Fig. 1: based on the parallel lines overhaul flow chart of AMFDBT

Embodiment

Now in conjunction with the embodiments, the invention will be further described for accompanying drawing:

The present invention has defined the self-adaptation tile and has returned subregion (Adaptive Tile Recursive Dyadic Partitioning is called for short ATRDP) algorithm on the basis that the recursive partitioning algorithm is advanced in traditional discrete Beamlet conversion two.Here use S _{I, j, l}Represent each subregion sub-block, wherein (i, j) represents x coordinate and the y coordinate of this sub-block upper left angle point in original image, and l represents the length of side of sub-block, and k represents k subregion sub-block, and J represents to carry out for the J time the self-adaptation tile and returns subregion.The ATRDP algorithm is defined as follows:

1. ATP _JAn ATRDP.

If 2. ${\mathrm{ATP}}_J=\{S_{i_0,j_0,l_0},....S_{i_k,j_k,l_k},...,S_{i_n,j_n,l_n}\}$ An ATRDP, and If satisfy the adaptive partition condition (refer to find beamlet b line segment in this blockette and eliminate from figure the contained impact point of b line segment can change blockette in the distributed areas of impact point),

Be adjusted into by the region adaptivity that exists by its intrapartition destination punctuate

${\mathrm{ATP}}_{J+1}=\{S_{i_0,j_0,l_0},...,S_{i_k,j_k,l_k},S{\left(\mathrm{adap}\right)}_{i_k,j_k,l_k},...,S_{i_n,j_n,l_n}\}$

Also ATRDP.

3. If do not satisfy the adaptive partition condition, and

Can be subdivided into four equal subregion sub-blocks

Its yardstick is readjusted in the zone that exists by impact point in these four sub-blockettes respectively, obtains

${\mathrm{ATP}}_{J+1}=\begin{array}{c}\{S_{i_0,j_0,l_0},...,S_{i_k,j_k,l_k},...,S{\left(\mathrm{adap}\right)}_{i_k,j_k,l_k/2},S{\left(\mathrm{adap}\right)}_{i_k+l_k/2,j_k,l_k/2},{S\left(\mathrm{adap}\right)}_{i_k,j_k+l_k/2,l_k/2},\\ S{\left(\mathrm{adap}\right)}_{i_k+l_k/2,j_k+l_k/2,l_k/2...,S_{i_n,j_n,l_n}\}}\end{array}$ Also ATRDP, wherein

Be

Father's piece.

With ATRDP partition method and fast discrete Beamlet conversion (Fast Discrete Beamlet Transform is called for short FDBT) combination, definition self-adapting multi-dimension fast discrete Beamlet is transformed to:

Wherein

Be fast discrete Beamlet conversion, be defined as follows:

$G\left(f(x,y)\right)=\left\{\begin{array}{cc}1& f(x,y)\&GreaterEqual;t\\ 0& f(x,y)<t\end{array}\right.---\left(3\right)$

In formula (1), beamlet b is a line segment in subregion sub-block S, and S is ATRDP subregion sub-block.The cut set of AMFDBT has comprised all energy functions

(wherein || b|| is the element number on beamlet b) over assign thresholds t ^*Beamlet b, will no longer consider in the fine dimension sub-block that next stage decomposes at beamlet b that sub-block under the last layer large scale has existed, the decomposition of the beamlet b under large scale can be expressed as b=U _jb _jSo the cut set of AMFDBT is:

$B_{t^{*}}=\left\{b\right|E\left(b\right)>t^{*},|b\&Element;S,S\&Element;{\mathrm{ATP}}_J,b\&NotElement;{\tilde{D}}_{t^{*}}\}---\left(4\right)$

${\tilde{D}}_{t^{*}}=\left\{b_j\right|b=U_j{b}_j,E\left(b\right)>t^{*},b\&Element;P\}---\left(5\right)$

Wherein P is father's piece of S.

(2) parallel lines based on AMFDBT detect

The present invention is applied to that with AMFDBT image is carried out parallel lines and detects.

At first will carry out series of preprocessing to image before parallel detection, its concrete steps are as follows:

1. utilize the Canny operator to carry out rim detection to image I, obtain edge-detected image E.

2. range of linearity screening is rejected curve or than short lines.Due in image I, exist a large amount of complex backgrounds toward contact except target, after rim detection, their main manifestations go out a little irregular or crooked edge features, and these edge features can cause larger impact in the straight-line detection process.Therefore before straight-line detection, first edge image E carries out the edge contour screening, only stays the more range of linearity of marginal point in the same way, removes some crooked irregular or too short edge features.

The present invention is mainly the Gradient Phase information of utilizing edge pixel point, realizes the screening of the range of linearity.At first the deflection of edge calculation pixel gradient:

In formula, G _xAnd G _yThe horizontal and vertical gradient component that represents respectively marginal point.When the gradient direction angle of two marginal points differed by more than predetermined threshold value threshold, expression edge contour direction transformation was larger, in this marginal point place's disconnection.Edge contour can be divided into a lot of ranges of linearity like this, mark these ranges of linearity, statistics range of linearity inward flange is counted, and detect with further minimizing and disturb in deletion marginal point less line district.We will be designated as E ' through the image after line district's screening.

Utilize the AMFDBT parallel detection on the basis of edge image E ', specific algorithm is described below:

1. traversing operation subgraph (initial operation subgraph is edge image E '), the zone of finding impact point in figure (being that image is through the marginal point after rim detection) to exist.If this area size is carried out next step greater than defined threshold t ' (t ' can adjust according to the image size is generally picture traverse/100); Otherwise execution in step 4..

2. the impact point zone of 1. finding with step is carried out fast discrete Beamlet conversion as the subregion sub-block to this blockette, finds out all the beamlet straight lines in current blockette, calculates its slope according to the rectilinear end dot information.

If 3. find the beamlet straight line in step in 2., extract straight line information and also wipe this straight line; Otherwise, with the current blockette quartern, and as the initial size of next yardstick subregion sub-block.Execution in step 1..

If 4. detect the beamlet straight line, according to the slope information of straight line, find out all parallel lines.

Embodiment of the present invention reference flow sheet (1), as follows based on the parallel lines detecting step of self-adapting multi-dimension Beamlet conversion:

The present invention is applied to that with AMFDBT image is carried out parallel lines and detects, and its concrete steps are as follows:

(1) image pre-service

1. utilize the Canny operator to carry out rim detection to image I, obtain edge-detected image E.

2. range of linearity screening is rejected curve or than short lines.Due in image I, exist a large amount of complex backgrounds toward contact except target, after rim detection, their main manifestations go out a little irregular or crooked edge features, and these edge features can cause larger impact in the straight-line detection process.Therefore before straight-line detection, first edge image E carries out the edge contour screening, only stays the more range of linearity of marginal point in the same way, removes some crooked irregular or too short edge features.

The present invention is mainly the Gradient Phase information of utilizing edge pixel point, realizes the screening of the range of linearity.At first the deflection of edge calculation pixel gradient:

In formula, G _xAnd G _yThe horizontal and vertical gradient component that represents respectively marginal point.When the gradient direction angle of two marginal points differed by more than predetermined threshold value threshold (threshold=20 herein), expression edge contour direction transformation was larger, in this marginal point place's disconnection.Edge contour can be divided into a lot of ranges of linearity like this, mark these ranges of linearity, statistics range of linearity inward flange is counted, and detect with further minimizing and disturb in deletion marginal point less line district.We will be designated as E ' through the image after line district's screening.

(2) parallel lines based on AMFDBT detect

1. traversing operation subgraph (initial operation subgraph is edge image E '), the zone of finding impact point in figure (being that image is through the marginal point after rim detection) to exist.If this area size is carried out next step greater than defined threshold t ' (t ' can adjust according to the image size is generally picture traverse/100); Otherwise execution in step 4..

2. the impact point zone of 1. finding with step is carried out fast discrete Beamlet conversion as the subregion sub-block to this blockette, finds out all the beamlet straight lines in current blockette, calculates its slope according to the rectilinear end dot information.

If 3. find the beamlet straight line in step in 2., extract straight line information (end points and slope) and also wipe this straight line; Otherwise, with the current blockette quartern, and as the initial size of next yardstick subregion sub-block.Execution in step 1..

If 4. detect the beamlet straight line, according to the slope information of straight line, find out all parallel lines.

Claims (1)

1. line detection method based on self-adapting multi-dimension fast discrete Beamlet conversion is characterized in that step is as follows:

Step 1: utilize the Canny operator to carry out rim detection to image I, obtain edge-detected image E;

Step 2: the deflection of edge calculation detected image E up contour point gradient:

When the gradient direction angle of two marginal points differs by more than predetermined threshold value threshold, expression edge contour direction transformation is larger, judge that these two marginal points do not belong to the same range of linearity, are divided into several ranges of linearity and separate marking with this with all marginal points; In formula, G _xAnd G _yThe horizontal and vertical gradient component that represents respectively marginal point;

Step 3: all are contained the edge count and be less than the deletion of 20 the range of linearity, obtain through the image E ' after the screening of the range of linearity, and with edge image E ' as subregion sub-block at first, with all marginal points in E ' as impact point;

Step 4: seek Minimum Area that impact point exists as the adaptive partition sub-block under current subregion sub-block, if the size of this adaptive partition sub-block during more than or equal to predetermined threshold value t', execution in step 5; Otherwise execution in step 6;

Step 5: current adaptive partition sub-block is carried out fast discrete Beamlet conversion, detection of straight lines; If find the beamlet straight line, record end points and the slope information of straight line, then wipe this straight line under current subregion sub-block; Otherwise, with the current subregion sub-block quartern, and as the initial size of next yardstick subregion sub-block, turn back to step 4;

Step 6: if the quantity of beamlet straight line greater than 0, according to the slope information of all straight lines, find straight slope poor 1 the degree with all interior straight lines as parallel lines.

Images