JPH0520457A - Multiplex resolution processing incorporating type hough transformation method - Google Patents

Multiplex resolution processing incorporating type hough transformation method

Info

Publication number
JPH0520457A
JPH0520457A JP3172167A JP17216791A JPH0520457A JP H0520457 A JPH0520457 A JP H0520457A JP 3172167 A JP3172167 A JP 3172167A JP 17216791 A JP17216791 A JP 17216791A JP H0520457 A JPH0520457 A JP H0520457A
Authority
JP
Japan
Prior art keywords
cell
image
resolution
grid
grid point
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
JP3172167A
Other languages
Japanese (ja)
Other versions
JP2882912B2 (en
Inventor
Hiroshi Kamata
洋 鎌田
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
A T R SHICHIYOUKAKU KIKO KENKYUSHO KK
ATR AUDITORY VISUAL PERCEPTION
Original Assignee
A T R SHICHIYOUKAKU KIKO KENKYUSHO KK
ATR AUDITORY VISUAL PERCEPTION
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by A T R SHICHIYOUKAKU KIKO KENKYUSHO KK, ATR AUDITORY VISUAL PERCEPTION filed Critical A T R SHICHIYOUKAKU KIKO KENKYUSHO KK
Priority to JP3172167A priority Critical patent/JP2882912B2/en
Publication of JPH0520457A publication Critical patent/JPH0520457A/en
Application granted granted Critical
Publication of JP2882912B2 publication Critical patent/JP2882912B2/en
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Landscapes

  • Complex Calculations (AREA)
  • Image Analysis (AREA)

Abstract

PURPOSE:To supply a multiplex resolution processing incorporating type Hough information method for calculating Hough information by means of adaptively defining a cell near a picture element in accordance with picture information by incorporating a method used in a multiplex resolution processing in the definition of the cell. CONSTITUTION:When picture density in the cell is near '1', or the number of the candidates of segments selected in the cell becomes more than the number of the segments which is set and which is to be extracted, a grating whose resolution is 1/2 of an original grating is defined. A picture (g) on the 1/2 grating is calculated by a prescribed operation expression and the cell near B is defined again on the 1/2 grating. Thus, the corresponding straight lines are sequentially extracted from that with a large duplicate degree by calculating a straight candidate group (rho1, theta1) from the picture (g) in the cell.

Description

【発明の詳細な説明】Detailed Description of the Invention

【0001】[0001]

【産業上の利用分野】この発明は、多重解像度処理組込
型ハフ変換法に関し、特に、画像処理において、ハフ変
換を用いて画像の線分情報を抽出するような多重解像度
処理組込型ハフ変換法に関する。
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multi-resolution processing built-in Hough transform method, and more particularly to a multi-resolution processing built-in Hough transform method for extracting line segment information of an image using Hough transform in image processing. Regarding the conversion method.

【0002】[0002]

【従来の技術】2次元,3次元画像を認識し、理解する
ためには、その前処理および特徴抽出方法が最も重要な
課題であり、多くの方法が提案され、使われてきた。そ
の中でも、ハフ変換は画像の中から有効に直線,曲線を
抽出する方法であり、コンピュータビジョンなどの分野
で広く使われている。
2. Description of the Related Art In order to recognize and understand two-dimensional and three-dimensional images, preprocessing and feature extraction methods are the most important problems, and many methods have been proposed and used. Among them, the Hough transform is a method of effectively extracting straight lines and curves from an image, and is widely used in fields such as computer vision.

【0003】一方、文字認識の分野でも、前処理と特徴
抽出法が最も重要な課題であり、種々の方法が試みられ
ており、ハフ変換を用いる方法も提案されている。従来
のハフ変換の基本的原理は、画像点(X,Y)と検出し
たい線分の方向θを与え、次の第(2)式および第
(3)式によりハフ変換値H(ρ,θ)をすべての画像
点で加算することにある。
On the other hand, also in the field of character recognition, the preprocessing and the feature extraction method are the most important problems, various methods have been tried, and a method using Hough transform has been proposed. The basic principle of the conventional Hough transform is to give the image point (X, Y) and the direction θ of the line segment to be detected, and then use the following equations (2) and (3) to obtain the Hough transform value H (ρ, θ). ) Is to be added at all image points.

【0004】 ρ=xcosθ+ysinθ …(2) H(ρ,θ)=H(ρ,θ)+1 …(3)[0004]   ρ = xcos θ + ysin θ (2)   H (ρ, θ) = H (ρ, θ) +1 (3)

【0005】[0005]

【発明が解決しようとする課題】上述のハフ変換を用い
る方法では、本来存在しない線分まで検出してしまうお
それがある。これを解決するために、本願発明者は黒画
素の近傍のセルを定義し、セル内で画像の直線性を評価
し、加算値を重み付けする修正セグメント化ハフ変換法
を提案した(特願平3−165669)。
In the method using the above Hough transform, there is a possibility that even a line segment that does not originally exist may be detected. In order to solve this, the present inventor has defined a cell in the vicinity of a black pixel, evaluated the linearity of the image in the cell, and proposed a modified segmented Hough transform method for weighting the added value (Japanese Patent Application No. Hei 10-135242). 3-165669).

【0006】図6は修正セグメント化ハフ変換法の処理
方法を説明するためのフロー図である。この方法は、画
像上の画素近傍を含むセルを定義し、セルの画像情報f
から予め直線の角度成分θ1 を決定し、次の第(4)式
および第(5)式によりハフ変換値に加算する方法であ
る。
FIG. 6 is a flow chart for explaining the processing method of the modified segmented Hough transform method. This method defines a cell that includes a pixel neighborhood on the image, and the image information f of the cell
Is a method of previously determining the angle component θ 1 of the straight line and adding it to the Hough transform value by the following equations (4) and (5).

【0007】 ρ1 =xcosθ1 +ysinθ1 …(4) H(ρ1 ,θ1 )=H(ρ1 ,θ1 )+m1 …(5) ただし、m1 は重み付けされた加算値である。この方法
は、ゴースト成分の除去に著しい効果があり、また、耐
ノイズ性にも優れているが、セルがすべて黒画素から成
立つときは、直線性が評価できないため、結果として加
算値が線幅に依存するという問題点があった。
[0007] ρ 1 = xcosθ 1 + ysinθ 1 ... (4) H (ρ 1, θ 1) = H (ρ 1, θ 1) + m 1 ... (5) However, m 1 is the weighted sum value. This method has a remarkable effect of removing the ghost component and is also excellent in noise resistance, but when the cells are all made up of black pixels, the linearity cannot be evaluated, and as a result, the added value becomes linear. There was a problem that it depends on the width.

【0008】それゆえに、この発明の主たる目的は、多
重解像度処理で用いられる手法をセルの定義に組込むこ
とにより、画素の近傍のセルを画像情報に応じて適応的
に定義するようにした多重解像度処理組込型ハフ変換法
を提供することである。
Therefore, a main object of the present invention is to incorporate the technique used in multi-resolution processing into the definition of a cell so as to adaptively define a cell in the vicinity of a pixel according to image information. It is to provide a processing embedded Hough transform method.

【0009】[0009]

【課題を解決するための手段】請求項1に係る発明は、
画像上の格子点P(x,y)において、格子点Pを通り
座標軸と角度θで交わり、原点から下ろした垂線との交
点までの距離がρである直線(ρ,θ)を対応させ、画
像上の格子点Pを中心とする近傍のセルを定義し、格子
点Pを通りセル内のすべての画像点に最も一致する直線
を複数本予め算出し、それぞれの直線に対応するρ,θ
を算出し、予め算出した直線をすべての画像点上で加算
することにより(ρ,θ)に対し、重複度を算出し、重
複度の多い順に対応の直線を抽出するハフ変換法におい
て、セルの画像濃度dが1に近い場合、またはセル内で
選ばれた線分の候補数が設定した抽出すべき線分の数よ
りも多くなったとき、解像度が元の格子1/2である1
/2格子を定義し、この1/2格子上の画像gを
The invention according to claim 1 is
At a grid point P (x, y) on the image, a straight line (ρ, θ) that intersects the coordinate axis through the grid point P at an angle θ and has a distance ρ from the origin to the intersection point with the perpendicular is made to correspond, A neighboring cell centered on the grid point P on the image is defined, and a plurality of straight lines that pass through the grid point P and best match all the image points in the cell are calculated in advance, and ρ and θ corresponding to the respective straight lines are calculated.
In the Hough transform method, in which the degree of overlap is calculated for (ρ, θ) by adding the previously calculated straight lines on all image points and the corresponding straight lines are extracted in descending order of the degree of overlap. When the image density d of is close to 1, or when the number of line segment candidates selected in the cell is larger than the set number of line segments to be extracted, the resolution is the original lattice 1/2.
/ 2 grid is defined, and the image g on this 1/2 grid is

【0010】[0010]

【数2】 [Equation 2]

【0011】で算出し、1/2格子上で格子点Pの近傍
のセルを再定義し、セル内の画像gから直線候補群(ρ
1 ,θ1 )を算出する。
The cell in the vicinity of the grid point P on the 1/2 grid is redefined, and the straight line candidate group (ρ
1 , θ 1 ) is calculated.

【0012】請求項2に係る発明は、請求項1に係る発
明のセルの解像度に対し、ハフ変換の加算値m1 を反比
例させる。
The invention according to claim 2 makes the addition value m 1 of the Hough transform inversely proportional to the resolution of the cell of the invention according to claim 1.

【0013】請求項3に係る発明は、ハフ変換法におい
て、セルの画像濃度dが1に比べ十分小さい条件を満足
し、セルの画像が隣接するセルと連結せず孤立している
条件を満足するとき、解像度が元の格子の2倍である2
倍格子を定義し、予め測定した2倍解像度の画像fでセ
ルを再定義し、セル内の画像fから直線候補群(ρ1
θ1 )を算出する。
According to the third aspect of the invention, in the Hough transform method, the condition that the image density d of the cell is sufficiently smaller than 1 is satisfied, and the condition that the image of the cell is isolated without being connected to the adjacent cell is satisfied. , The resolution is twice that of the original grid
A double lattice is defined, a cell is redefined by an image f having a double resolution previously measured, and a straight line candidate group (ρ 1 ,
Calculate θ 1 ).

【0014】[0014]

【作用】セルの画像濃度dが1に近い場合、またはセル
内で選ばれた線分の候補数が設定した抽出すべき線分の
数よりも多くなったとき、解像度が元の格子の1/2で
ある1/2格子を定義し、1/2格子上の画像gを
When the image density d of the cell is close to 1, or when the number of line segment candidates selected in the cell exceeds the set number of line segments to be extracted, the resolution is 1 of the original grid. The 1/2 grid which is / 2 is defined, and the image g on the 1/2 grid is

【0015】[0015]

【数3】 [Equation 3]

【0016】で算出し、1/2格子上で格子点Pの近傍
のセルを再定義し、セル内の画像gから直線候補群(ρ
1 ,θ1 )を算出する。
The cell in the vicinity of the grid point P is redefined on the 1/2 grid, and the straight line candidate group (ρ
1 , θ 1 ) is calculated.

【0017】請求項2に係る発明は、請求項1に係る発
明のセルの解像度に対して、ハフ変換の加算値m1 を反
比例させる。
The invention according to claim 2 makes the addition value m 1 of the Hough transform inversely proportional to the resolution of the cell of the invention according to claim 1.

【0018】請求項3に係る発明は、セルの画像濃度d
が1に比べ十分小さい条件を満足し、セルの画像が隣接
するセルと連結せず孤立している条件を満足するとき、
解像度が元の格子の2倍である2倍格子を定義し、予め
測定した2倍解像度の画像fでセルを再定義し、セル内
の画像fから直線候補群(ρ1 ,θ1 )を算出する。
The invention according to claim 3 is the image density d of the cell.
Satisfies the condition that is sufficiently smaller than 1 and the condition that the image of the cell is isolated without being connected to the adjacent cells,
A double grid whose resolution is twice that of the original grid is defined, a cell is redefined with a premeasured double resolution image f, and a straight line candidate group (ρ 1 , θ 1 ) is obtained from the image f in the cell. calculate.

【0019】[0019]

【発明の実施例】図1はこの発明の一実施例の概略ブロ
ック図である。図1を参照して、画像読取装置1は線分
を抽出すべき画像を読取り、その画像の二値化出力をメ
モリ2に与える。メモリ2は二値化された画像データを
記憶する。ハフ変換演算部3はメモリ2に記憶されてい
る画像データをハフ変換し、メモリ4に記憶するととも
に、表示部5に表示させる。
1 is a schematic block diagram of an embodiment of the present invention. Referring to FIG. 1, the image reading device 1 reads an image from which a line segment is to be extracted, and gives a binary output of the image to a memory 2. The memory 2 stores the binarized image data. The Hough transform calculation unit 3 Hough transforms the image data stored in the memory 2, stores the image data in the memory 4, and causes the display unit 5 to display the image data.

【0020】前述の修正セグメント化ハフ変換法は、角
度分解能がセルサイズに依存するためセルサイズを変え
られないという困難な問題点を含んでいたのに対して、
本願発明は画素の近傍のセルを画像情報に応じて適応的
に定義する。このために、まずセルの大きさを変える判
定条件を以下のように設定する。
The above-mentioned modified segmented Hough transform method has a difficult problem that the cell size cannot be changed because the angular resolution depends on the cell size.
The present invention adaptively defines cells in the vicinity of pixels according to image information. For this purpose, first, the judgment conditions for changing the cell size are set as follows.

【0021】判定条件1:セル内の画像濃度dが極めて
高濃度なとき、すなわち、この条件はdを数4で求めた
とき、
Judgment condition 1: When the image density d in the cell is extremely high, that is, when d is obtained by the equation 4,

【0022】[0022]

【数4】 [Equation 4]

【0023】dが設定値dm に対して、以下の条件を満
たすときと定義する。 d≧dm …(7) dm は1に十分近い数とする。
It is defined that d satisfies the following condition with respect to the set value d m . d ≧ d m (7) d m is a number sufficiently close to 1.

【0024】判定条件2:セル内で選ばれた線分の候補
数が、予め設定した抽出すべき線分の総数よりも多いこ
と。
Judgment condition 2: The number of line segment candidates selected in the cell is larger than a preset total number of line segments to be extracted.

【0025】以上の判定条件の1つでも満足するとき
は、図2に示すように、点線の原画像から実線で示す解
像度が1/2の格子でセルを再定義し(以下、1/2セ
ルと称する)、次の数5により1/2セルの画像濃度g
を求める。
If any one of the above judgment conditions is satisfied, as shown in FIG. 2, cells are redefined from a dotted original image by a grid having a resolution of ½ (hereinafter, ½). Image density g of 1/2 cell according to the following equation 5
Ask for.

【0026】[0026]

【数5】 [Equation 5]

【0027】ただし、Lは結合が及ぶ範囲であり通常2
である。また、w(i,j)は結合係数であり、たとえ
ば、1次元格子でガウス型のピラミッド画像を構成する
ときは図3に示す値を用いる。なお、より詳細な値につ
いては、(A.Rosenfeld (ed.):Multiresolution Image
Processing and Analysis, Springer-Verlag (1984)
)に詳細に説明されている。
However, L is the range of the bond and is usually 2
Is. Further, w (i, j) is a coupling coefficient, and for example, when a Gaussian-type pyramid image is constructed with a one-dimensional lattice, the values shown in FIG. 3 are used. For more detailed values, see (A.Rosenfeld (ed.): Multiresolution Image
Processing and Analysis, Springer-Verlag (1984)
) In detail.

【0028】図2から明らかなように、解像度を1/2
にすることにより、セルがカバーする画像域は4倍とな
るが、セルの格子点の数は元の格子点と変わらない。こ
のため、元の格子に対して角度分解能を不変に保つこと
ができる。
As is apparent from FIG. 2, the resolution is reduced to 1/2.
By doing so, the image area covered by the cell is quadrupled, but the number of grid points of the cell is the same as the original grid point. Therefore, the angular resolution can be kept unchanged with respect to the original grating.

【0029】図4はこの発明の一実施例の具体的な動作
を説明するためのフロー図である。まず、黒画素(x,
y)を抽出し、前述の数1によりセル内の画素数が元の
セルと同じである解像度が1/2のセルと画像濃度を定
義する。1/2セルにおいて、再度判定条件1と2を調
べ、その1つでも満たすときはさらに解像度を1/2に
することにより、1/4セルを定義する。同様の手順で
セルが判定条件を満たさなくなったとき、以下のステッ
プでハフ変換値H(ρ,θ)を計算する。すなわち、セ
ル内で種々の傾きの直線を当てはめ、最もよく一致する
方向θ1 に対し、加算値m1 に対し1を加える。ただ
し、それが複数個あるときは全部のθ1 について加算す
る。
FIG. 4 is a flow chart for explaining the specific operation of the embodiment of the present invention. First, the black pixel (x,
y) is extracted, and the image density is defined as a cell having the same number of pixels in the cell as that of the original cell and a resolution of ½ according to the above-mentioned equation 1. In 1/2 cell, the determination conditions 1 and 2 are checked again, and if any one of them is satisfied, the resolution is further reduced to 1/2 to define 1/4 cell. When the cell does not satisfy the determination condition by the same procedure, the Hough transform value H (ρ, θ) is calculated in the following steps. That is, straight lines with various inclinations are fitted in the cell, and 1 is added to the added value m 1 with respect to the direction θ 1 that best matches. However, when there is more than one , add for all θ 1 .

【0030】次に、次式によりρ1 を計算する。 ρ1 =xcosθ1 +ysinθ1 …(8) ハフ変換として次式を計算する。Next, ρ 1 is calculated by the following equation. ρ 1 = xcos θ 1 + ysin θ 1 (8) The following equation is calculated as a Hough transform.

【0031】 H(ρ1 ,θ1 )=H(ρ1 ,θ1 )+m1 0 …(9) ただし、m0 はウェイトであり、修正セグメント化ハフ
変換法では1である。上述の計算をすべての画像全体に
対して行なうことにより、H(ρ,θ)を求めることが
できる。
H (ρ 1 , θ 1 ) = H (ρ 1 , θ 1 ) + m 1 m 0 (9) where m 0 is a weight and is 1 in the modified segmented Hough transform method. H (ρ, θ) can be obtained by performing the above calculation on all the images.

【0032】前述の実施例では、線分の太さに依存せ
ず、その候補を絞り込めることができるが、セルの解像
度を変化させたとき、それに寄与する画素数を比較した
とき、解像度kが低くなるほど、その数は1/kの自乗
に比例して大きくなり、ウェイトm0 の値を固定するこ
とは不自然である。また、m0 の値を固定すると、判定
条件を満たすときに比較してそうでない場合の方が圧倒
的に多い画像に対してはセルの大きさを可変にする効果
が現われにくくなる問題を生じる。
In the above-described embodiment, the candidates can be narrowed down without depending on the thickness of the line segment, but when the cell resolution is changed and the number of pixels contributing to it is compared, the resolution k Is smaller, the number becomes larger in proportion to the square of 1 / k, and it is unnatural to fix the value of the weight m 0 . Further, if the value of m 0 is fixed, there arises a problem that the effect of changing the cell size is less likely to appear for an image that is overwhelmingly larger than that when the determination condition is satisfied. .

【0033】このため、解像度がk(k=1,1/2,
1/4…)のときのウェイトをm0 (k)で表わすと、 m0 (k/2)=4m0 (k) …(10) とすれば、セルの解像度に反比例してウェイトを増加で
きる。ただし、係数4はセルの解像度が元の格子の1/
2になったとき、セルが含む領域が元のセルに対して4
倍となるように対応させた値であり、その値は用途に応
じて変えることが可能である。上述の処理は、図4の点
線で示すステップで行なわれる。
Therefore, the resolution is k (k = 1, 1/2,
When the weight at 1/4 ...) is represented by m 0 (k), if m 0 (k / 2) = 4m 0 (k) (10), the weight increases in inverse proportion to the cell resolution. it can. However, the coefficient 4 is 1 / the cell resolution of the original grid.
When it becomes 2, the area contained by the cell is 4 with respect to the original cell.
The value corresponds to double the value, and the value can be changed according to the application. The above-mentioned processing is performed in the steps shown by the dotted line in FIG.

【0034】図5はこの発明のさらに他の実施例を示す
フロー図である。この実施例は、セルの解像度を逆に高
くするようにしたものである。すなわち、画像を処理す
るとき、最初から高い解像度で処理するより、数1で求
めた低い解像度の画像gで大部分の画像部を処理し、そ
れで対応できない部分のみ、高い解像度で処理した方が
効率がよい場合がある。低い解像度で対応できない例と
して、ダッシュや句読点などの孤立点がある。このよう
な場合に対応するため、以下の判定条件とセルの定義を
用いる。すなわち、判定条件3:セル内の画像濃度dが
1よりも十分小さくかつセル画像が隣接するいずれのセ
ルの画像とも連結していないこと。
FIG. 5 is a flow chart showing still another embodiment of the present invention. In this embodiment, the resolution of the cell is increased to the contrary. That is, when processing an image, it is better to process most of the image part with the image g having the lower resolution obtained by the equation 1 and process only the part that cannot be processed with the higher resolution than the high resolution from the beginning. It may be efficient. Dashes, punctuation, and other isolated points are examples of low resolution that cannot be accommodated. In order to deal with such a case, the following determination condition and cell definition are used. That is, determination condition 3: the image density d in the cell is sufficiently smaller than 1 and the cell image is not connected to the images of any adjacent cells.

【0035】上述の判定条件3が成立したとき、予め測
定した2倍の高い解像度でセルを再定義し、修正セグメ
ント化ハフ変換法の手順でハフ変換値H(ρ,θ)を計
算すれば、このような場合にも対応可能となる。
When the above-mentioned determination condition 3 is satisfied, the cell is redefined at a resolution twice as high as previously measured, and the Hough transform value H (ρ, θ) is calculated by the procedure of the modified segmented Hough transform method. It is possible to handle such a case.

【0036】[0036]

【発明の効果】以上のように、この発明によれば、元の
画像の解像度に依存しない線分の抽出が可能であるた
め、文書画像から文字領域と画像領域を分離したり、文
字や記号からストロークを抽出することなどの広範囲な
分野に適用することができる。
As described above, according to the present invention, it is possible to extract a line segment that does not depend on the resolution of the original image, so that the character area and the image area can be separated from the document image, or the character or symbol can be separated. It can be applied to a wide range of fields such as extracting strokes from

【図面の簡単な説明】[Brief description of drawings]

【図1】この発明の一実施例の概略ブロック図である。FIG. 1 is a schematic block diagram of an embodiment of the present invention.

【図2】この発明の一実施例における1/2セル格子の
定義を説明するための図である。
FIG. 2 is a diagram for explaining the definition of a 1/2 cell lattice in one embodiment of the present invention.

【図3】ガウス型1次元ピラミッドの構成を示す図であ
る。
FIG. 3 is a diagram showing a configuration of a Gaussian one-dimensional pyramid.

【図4】この発明の一実施例の動作を説明するためのフ
ロー図である。
FIG. 4 is a flowchart for explaining the operation of the embodiment of the present invention.

【図5】この発明のさらに他の実施例を示すフロー図で
ある。
FIG. 5 is a flow chart showing still another embodiment of the present invention.

【図6】従来の修正セグメント化ハフ変換法の処理手順
を説明するためのフロー図である。
FIG. 6 is a flowchart for explaining a processing procedure of a conventional modified segmented Hough transform method.

【符号の説明】 1 画像読取装置 2,4 メモリ 3 ハフ変換演算部 5 表示部[Explanation of symbols] 1 Image reading device 2,4 memory 3 Hough transform calculation unit 5 Display

Claims (3)

【特許請求の範囲】[Claims] 【請求項1】 画像上の格子点P(x,y)において、
前記格子点Pを通り座標軸と角度θで交わり、原点から
下ろした垂線との交点までの距離がρである直線(ρ,
θ)を対応させ、画像上の格子点Pを中心とする近傍の
セルを定義し、該格子点Pを通りセル内のすべての画像
点に最も一致する直線を複数本予め算出し、それぞれの
直線に対応するρ,θを算出し、予め算出した直線をす
べての画像点上で加算することにより、(ρ,θ)に対
して重複度を算出し、重複度の多い順に対応の直線を抽
出するハフ変換法において、 前記セルの解像濃度dが1に近い場合、または前記セル
内で選ばれた線分の候補数が設定した抽出すべき線分の
数よりも多くなったとき、解像度が元の格子の1/2で
ある1/2格子を定義し、前記1/2格子上の画像gを 【数1】 で算出し、1/2格子上で格子点Pの近傍のセルを再定
義し、前記セル内の画像gから直線候補群(ρ1
θ1 )を算出するようにしたことを特徴とする、多重解
像度処理組込型ハフ変換法。
1. At a grid point P (x, y) on the image,
A straight line (ρ, which passes through the grid point P and intersects the coordinate axis at an angle θ and whose distance from the origin to the intersection of the perpendicular line is ρ
θ) corresponding to each other, defining a cell in the vicinity of the grid point P on the image, calculating a plurality of straight lines which pass through the grid point P and best match all the image points in the cell in advance, and By calculating ρ and θ corresponding to the straight line and adding the pre-calculated straight lines on all the image points, the degree of overlap is calculated for (ρ, θ), and the corresponding straight lines are arranged in descending order of the degree of overlap. In the Hough transform method for extraction, when the resolution density d of the cell is close to 1, or when the number of line segment candidates selected in the cell is larger than the set number of line segments to be extracted, A 1/2 grid whose resolution is 1/2 of the original grid is defined, and an image g on the 1/2 grid is defined as follows. And redefine the cells in the vicinity of the grid point P on the ½ grid, and from the image g in the cell, the straight line candidate group (ρ 1 ,
A multi-resolution processing embedded Hough transform method characterized by calculating θ 1 ).
【請求項2】 前記セルの解像度に対して、ハフ変換の
加算値m1 を反比例させることを特徴とする、請求項1
の多重解像度処理組込型ハフ変換法。
2. The addition value m 1 of Hough transform is inversely proportional to the resolution of the cell.
Huff transform method with multi-resolution processing.
【請求項3】 画像上の格子点P(x,y)において、
格子点Pを通り座標軸と角度θで交わり、原点から下ろ
した垂線との交点までの距離がρである直線(ρ,θ)
を対応させ、画像上の格子点Pを中心とする近傍のセル
を定義し、格子点Pを通りセル内のすべての画像点に最
も一致する直線を複数本予め算出し、それぞれの直線に
対応するρ,θを算出し、予め算出した直線をすべての
画像点上で加算することにより(ρ,θ)に対して重複
度を算出し、重複度の多い順に対応の直線を抽出するハ
フ変換法において、 前記セルの画像濃度dが1に比べ十分小さい条件を満足
し、前記セルの画像が隣接するセルと連結せず孤立して
いる条件を満足するとき、解像度が元の格子の2倍であ
る2倍格子を定義し、予め測定した2倍解像度の画像f
でセルを再定義し、セル内の画像fから直線候補群(ρ
1 ,θ1 )を算出することを特徴とする、多重解像度処
理組込型ハフ変換法。
3. At a grid point P (x, y) on the image,
A straight line (ρ, θ) that passes through the grid point P and intersects the coordinate axis at an angle θ, and the distance from the origin to the intersection with the perpendicular line is ρ
By defining the neighboring cells centered on the grid point P on the image, calculating in advance a plurality of straight lines that pass through the grid point P and best match all the image points in the cell, and correspond to each straight line. Hough transform that calculates ρ and θ and calculates the degree of overlap for (ρ, θ) by adding the pre-calculated straight lines on all image points and extracts the corresponding straight lines in descending order of the degree of overlap. In the method, when the image density d of the cell satisfies the condition that it is sufficiently smaller than 1 and the image of the cell satisfies the condition of being isolated without being connected to an adjacent cell, the resolution is twice as high as that of the original grid. Image f with double resolution previously defined by defining a double grid
, The cell is redefined, and the line candidate group (ρ
A multi-resolution processing embedded Hough transform method, which is characterized by calculating 1 , θ 1 ).
JP3172167A 1991-07-12 1991-07-12 Hough transform method with built-in multi-resolution processing Expired - Lifetime JP2882912B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP3172167A JP2882912B2 (en) 1991-07-12 1991-07-12 Hough transform method with built-in multi-resolution processing

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP3172167A JP2882912B2 (en) 1991-07-12 1991-07-12 Hough transform method with built-in multi-resolution processing

Publications (2)

Publication Number Publication Date
JPH0520457A true JPH0520457A (en) 1993-01-29
JP2882912B2 JP2882912B2 (en) 1999-04-19

Family

ID=15936820

Family Applications (1)

Application Number Title Priority Date Filing Date
JP3172167A Expired - Lifetime JP2882912B2 (en) 1991-07-12 1991-07-12 Hough transform method with built-in multi-resolution processing

Country Status (1)

Country Link
JP (1) JP2882912B2 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1995033241A1 (en) * 1994-05-27 1995-12-07 Hitachi, Ltd. High-speed arithmetic unit for discrete cosine transform and associated operation

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1995033241A1 (en) * 1994-05-27 1995-12-07 Hitachi, Ltd. High-speed arithmetic unit for discrete cosine transform and associated operation
US6029185A (en) * 1994-05-27 2000-02-22 Hitachi, Ltd. Discrete cosine high-speed arithmetic unit and related arithmetic unit
US6223195B1 (en) 1994-05-27 2001-04-24 Hitachi, Ltd. Discrete cosine high-speed arithmetic unit and related arithmetic unit

Also Published As

Publication number Publication date
JP2882912B2 (en) 1999-04-19

Similar Documents

Publication Publication Date Title
CN107680054B (en) Multi-source image fusion method in haze environment
JP3345350B2 (en) Document image recognition apparatus, method thereof, and recording medium
US5434953A (en) Use of fast textured reduction for discrimination of document image components
CN101789122B (en) Method and system for correcting distorted document image
US6185341B1 (en) Image processing using vector data to reduce noise
CN104809731B (en) A kind of rotation Scale invariant scene matching method based on gradient binaryzation
US20060062458A1 (en) Method and system for identifying objects in an image
KR102289045B1 (en) Method and Apparatus for Providing Multi-scale Object Image Restoration
CN114529459B (en) Method, system and medium for enhancing image edge
CN110180186A (en) A kind of topographic map conversion method and system
CN101957991A (en) Remote sensing image registration method
CN112016557B (en) Method for removing form interference line
Lacroix et al. Feature extraction using the constrained gradient
Ham et al. Recognition of raised characters for automatic classification of rubber tires
JPH0520457A (en) Multiplex resolution processing incorporating type hough transformation method
JP4259949B2 (en) Image creating apparatus, image creating program, and recording medium
Lindahl Study of local binary patterns
CN112907553A (en) High-definition image target detection method based on Yolov3
US20020164087A1 (en) System and method for fast rotation of binary images using block matching method
EP0702320A1 (en) Skew detection
JPH07101455B2 (en) Line detection method and apparatus
CN114332108B (en) Method for extracting virtual-real line local area in picture
CN115588204B (en) Single character image matching recognition method based on DS evidence theory
JP2613211B2 (en) Image input device
JP4259950B2 (en) Image recognition apparatus, image recognition program, and recording medium

Legal Events

Date Code Title Description
A02 Decision of refusal

Free format text: JAPANESE INTERMEDIATE CODE: A02

Effective date: 19950725