JP3690002B2

JP3690002B2 - Image reading method, image reading apparatus, and medium on which image reading program is recorded

Info

Publication number: JP3690002B2
Application number: JP28558196A
Authority: JP
Inventors: 哲男伊與田
Original assignee: Fuji Xerox Co Ltd; Fujifilm Business Innovation Corp
Current assignee: Fujifilm Business Innovation Corp
Priority date: 1996-10-28
Filing date: 1996-10-28
Publication date: 2005-08-31
Anticipated expiration: 2016-10-28
Also published as: JPH10136196A

Description

【０００１】
【発明の属する技術分野】
本発明は、原稿画像を複数の分割画像に分割して順次読取り、各分割画像を接続して原稿画像のデ−タを得る画像読取方法及び画像読取装置に関し、各分割画像間の境界部分の継ぎ目を顕在化させることなく接続し、原稿画像に忠実な読取り画像を得る画像読取方法及び画像読取装置、さらに、画像読取プログラムを記録した媒体に関する。
【０００２】
【従来の技術】
従来、パ−ソナルコンピュ−タ，ワ−ドプロセッサ用スキャナ等に使用されている画像読取装置としては、例えば、原稿画像を原稿サイズ及び指定する解像度に応じて複数の分割画像に分割し、倍率可変の結像光学手段及び固体撮像素子を保持する読取機構を走査させて必要な倍率で順次撮像し、各分割画像を接続して任意の解像度で原稿画像のデ−タを得る首振り型の画像読取装置が提案されている（特公平８−１３０８８号公報参照）。
【０００３】
この画像読取装置によれば、装置の小型化を図ることができると同時に原稿画像を任意の解像度で読み取ることができ、また、読み取り対象が平面的な画像に限られず立体的な画像をも読み取ることができる。
しかし、上記構成の画像読取装置においては、読み取り原稿と撮像素子間の距離が大きいため、撮像素子を保持する読取機構が所望の読取り位置から僅かにずれた場合でも、読取り画像は意図した画像とは大きくずれたものとなってしまう。
【０００４】
上記読取り画像の位置ずれを防止するため、読取機構の方向や位置を制御する機械系に対して精密な位置精度を要求すると、装置の大型化を招くことになる。そこで、各分割画像を隣接する分割画像と重複させて読取り、隣接する分割画像間を接続する際に、重複部分の両分割画像の画素値をもとに実際の接続位置を決定する方法が考えられる。
具体的には、接続する二つの分割画像の一方に対して、他方を一定の範囲内で相対的に移動させながら、両分割画像の重複領域の対応する画素間の画素値の差が最小となる位置を、両分割画像の接続位置とする。
【０００５】
【発明が解決しようとする課題】
上記方法によれば、読取機構の位置精度に拘らず、隣接する分割画像の接続位置を決定することができる。
そして、接続位置における両分割画像の重複領域の各画素の画素値を決定して両分割画像を接続するが、画素値の決定方法としては、重複領域全体の各画素値をいずれか一方の分割画像の対応する画素値で全面的に置き換える方法が最も簡単である。
しかし、この方法によると、接続位置に微少な誤差が存在する場合に、重複領域における両分割画像の対応画素間の画素値に生じるずれによって、重複領域と他方の分割画像の境界が明確な継ぎ目となって認識されるという問題点があった。このため、複数の分割画像を接続することにより得られる原稿画像は、接続箇所において画質欠陥を含む画像となっていた。
【０００６】
本発明は上記の実情に鑑みてなされたもので、原稿画像を複数の分割画像に分割して順次読み取り、各分割画像間の境界部の継ぎ目を顕在化させることなく接続することにより、原稿画像に忠実で高品質な読取り画像を任意の解像度で得ることができる画像読取方法及び画像読取装置、さらに、画像読取のためのプログラムを記録した媒体を提供することを目的とする。
【０００７】
【課題を解決するための手段】
上記目的を達成するため請求項１の画像読取方法は、原稿画像を複数の分割画像に分割し、各分割画像を隣接する分割画像と重複させて順次読み取り、各分割画像を接続して原稿画像のデ−タを得る画像読取方法であって、
隣接する二つの分割画像を接続する際に基準とする方の分割画像を基準画像、他方の分割画像を接続画像とし、前記接続画像の任意の一画素を探索基準画素とし、前記基準画像に対して前記接続画像を一定の範囲内で相対的に移動させるとともに、両画像の重複部分の一定面積領域における両画像の対応画素間の濃度差の平均値を求める処理と、それに応じて両画像の接続位置を決定するに際し、先ず、前記探索基準画素の位置を中心とし前記読取機構の位置精度に応じた面積の粗探索領域内で、前記探索基準画素を数画素間隔で移動させ、前記濃度差の平均値が最小となるときの探索基準画素の位置を基準画像と接続画像との暫定接続位置とし、
次に、前記暫定接続位置を中心とし前記粗探索領域より面積の狭い密探索領域の範囲内で、前記探索基準画素を一画素ずつ移動させ、前記濃度差の平均値が最小となるときの前記探索基準画素の位置を基準画像と接続画像との接続位置とし、前記接続位置における両分割画像の重複領域の画素値を、各画素毎にいずれか一方の分割画像の当該画素の画素値とすることにより、両分割画像の接続を行うことを特徴としている。
【０００８】
請求項２の画像読取装置は、倍率を可変することが可能な結像光学手段と、この結像光学手段を用いて結像された原稿画像の光学像を撮像する固体撮像素子とを保持する読取機構部と、前記読取機構部を、原稿画像全体を撮像するよう、及び、原稿画像を複数の分割画像に分割して必要な倍率で順次撮像するよう位置制御する一方、前記読取機構部より各分割画像を取り込み、接続して任意の解像度の原稿画像のデ−タを得る制御手段と、を有する画像読取装置であって、前記制御手段は、以下の手段を備えることを特徴としている。
前記各分割画像の画素値デ−タを原稿画像上の読取り位置と関連付けてそれぞれ記憶する分割画像記憶手段。
隣接する二つの分割画像を接続する際に基準とする方の分割画像を基準画像、他方の分割画像を接続画像とし、前記接続画像の任意の一画素を探索基準画素とし、前記基準画像に対して前記接続画像を一定の範囲内で相対的に移動させるとともに、両画像の重複部分の一定面積領域における両画像の対応画素間の濃度差の平均値を求める処理と、それに応じて両画像の接続位置を決定するに際し、先ず、前記探索基準画素の位置を中心とし前記読取機構の位置精度に応じた面積の粗探索領域内で、前記探索基準画素を数画素間隔で移動させ、前記濃度差の平均値が最小となるときの探索基準画素の位置を基準画像と接続画像との暫定接続位置とし、次に、前記暫定接続位置を中心とし前記粗探索領域より面積の狭い密探索領域の範囲内で、前記探索基準画素を一画素ずつ移動させ、前記濃度差の平均値が最小となるときの前記探索基準画素の位置を基準画像と接続画像との接続位置を決定する接続位置算出手段。
前記接続位置における両分割画像の重複領域の画素値を、各画素毎にいずれか一方の分割画像の当該画素の画素値とする重複画素選択手段。
【０００９】
上記画像読取方法及び画像読取装置によれば、隣接する二つの分割画像の接続において、重複領域の各画素毎にいずれか一方の分割画像の画素値を採用するので、重複領域の端部がいずれか一方の分割画像の画素値のみを配置した直線になりにくく、重複領域の内部と外部との境界が目立たないため、決定された接続位置に微少な誤差が存在する場合においても、画素値の不連続を良好に分散し継ぎ目の顕在化を防止することができる。
【００１０】
請求項３の画像読取装置は、請求項２記載の画像読取装置であって、前記重複画素選択手段は、前記両分割画像の重複領域の各画素値を、重複領域における当該画素の位置に基づいて、いずれの分割画像の当該画素の画素値とするかを判断することを特徴としている。
【００１１】
上記画像読取装置によれば、隣接する分割画像間の各接続部分において、接続位置誤差に起因する画素値不連続の分散を一様に行うことができるので、複数の分割画像を接続して得られる読取り画像の画質欠陥を緩和することができる。
【００１２】
請求項４の画像読取装置は、請求項３記載の画像読取装置であって、前記重複画素選択手段は、前記両分割画像の重複領域の各画素値を、重複領域における当該画素の位置が、前記重複領域を斜めに二分する対角線により分割される二つの領域のうちいずれの領域に属するかに基づいて、いずれの分割画像の当該画素の画素値とするかを判断することを特徴としている。
【００１３】
上記画像読取装置によれば、隣接する二つの分割画像の接続において、重複領域における両分割画像の画素値の継ぎ目が斜めの線分となることから、目の特性としてよく知られている水平及び垂直の直線状の形状への高い感度を避けることができ、見掛け上継ぎ目の顕在化を防止することができる。
【００１６】
請求項５の画像読取方法は、原稿画像を複数の分割画像に分割し、各分割画像を隣接する分割画像と重複させて順次読み取り、各分割画像を接続して原稿画像のデ−タを得る画像読取方法であって、
隣接する二つの分割画像を接続する際に基準とする方の分割画像を基準画像、他方の分割画像を接続画像とし、前記接続画像の任意の一画素を探索基準画素とし、前記基準画像に対して前記接続画像を一定の範囲内で相対的に移動させるとともに、両画像の重複部分の一定面積領域における両画像の対応画素間の濃度差の平均値を求める処理と、それに応じて両画像の接続位置を決定するに際し、先ず、前記探索基準画素の位置を中心とし前記読取機構の位置精度に応じた面積の粗探索領域内で、前記探索基準画素を数画素間隔で移動させ、前記濃度差の平均値が最小となるときの探索基準画素の位置を基準画像と接続画像との暫定接続位置とし、
次に、前記暫定接続位置を中心とし前記粗探索領域より面積の狭い密探索領域の範囲内で、前記探索基準画素を一画素ずつ移動させ、前記濃度差の平均値が最小となるときの前記探索基準画素の位置を基準画像と接続画像との接続位置とし、
前記接続位置における両分割画像の重複領域の画素値を、各画素毎に両分割画像の当該画素の画素値を用いた演算値とすることにより、両分割画像の接続を行うことを特徴としている。
【００１７】
請求項６の画像読取装置は、倍率を可変することが可能な結像光学手段と、この結像光学手段を用いて結像された原稿画像の光学像を撮像する固体撮像素子とを保持する読取機構部と、前記読取機構部を、原稿画像全体を撮像するよう、及び、原稿画像を複数の分割画像に分割して必要な倍率で順次撮像するよう位置制御する一方、前記読取機構部より各分割画像を取り込み、接続して任意の解像度の原稿画像のデ−タを得る制御手段と、を有する画像読取装置であって、前記制御手段は、以下の手段を備えることを特徴としている。
前記各分割画像の画素値デ−タを原稿画像上の読取り位置と関連付けてそれぞれ記憶する分割画像記憶手段。
隣接する二つの分割画像を接続する際に基準とする方の分割画像を基準画像、他方の分割画像を接続画像とし、前記接続画像の任意の一画素を探索基準画素とし、前記基準画像に対して前記接続画像を一定の範囲内で相対的に移動させるとともに、両画像の重複部分の一定面積領域における両画像の対応画素間の濃度差の平均値を求める処理と、それに応じて両画像の接続位置を決定するに際し、先ず、前記探索基準画素の位置を中心とし前記読取機構の位置精度に応じた面積の粗探索領域内で、前記探索基準画素を数画素間隔で移動させ、前記濃度差の平均値が最小となるときの探索基準画素の位置を基準画像と接続画像との暫定接続位置とし、次に、前記暫定接続位置を中心とし前記粗探索領域より面積の狭い密探索領域の範囲内で、前記探索基準画素を一画素ずつ移動させ、前記濃度差の平均値が最小となるときの前記探索基準画素の位置を基準画像と接続画像との接続位置を決定する接続位置算出手段。
前記接続位置における両分割画像の重複領域の画素値を、各画素毎に両分割画像の当該画素の画素値を用いた演算値とする重複画素演算手段。
【００１８】
上記画像読取方法及び画像読取装置によれば、隣接する二つの分割画像の接続において、重複領域の画素値を、両分割画像の当該画素値を用いて何らかの演算を行った値とするので、決定された接続位置に微少な誤差が存在する場合においても、画素値の不連続を良好に分散し、継ぎ目の顕在化を防止することができる。
【００１９】
請求項７の画像読取装置は、請求項６記載の画像読取装置であって、前記重複画素演算手段は、前記両分割画像の重複領域の各画素値を、両分割画像の当該画素の画素値の重みつき平均による演算値とすることを特徴としている。
【００２０】
上記画像読取装置によれば、隣接する二つの分割画像の接続において、重複領域の各画素値を、両分割画像の当該画素の画素値の重みつき平均による演算値とするので、重複領域とその両端の分割画像との境界における画素値を連続的に変化させることにより、画素値の不連続を良好に分散し、継ぎ目の顕在化を防止することができる。
【００２３】
請求項８の画像読取装置は、倍率を可変することが可能な結像光学手段と、この結像光学手段を用いて結像された原稿画像の光学像を撮像する固体撮像素子とを保持する読取機構部と、前記読取機構部を、原稿画像全体を撮像するよう、及び、原稿画像を複数の分割画像に分割して必要な倍率で順次撮像するよう位置制御する一方、前記読取機構部より各分割画像を取り込み、接続して任意の解像度の原稿画像のデ−タを得る制御手段と、を有する画像読取装置であって、前記制御手段は、以下の手段を備えたことを特徴としている。
前記各分割画像の画素値デ−タを原稿画像上の読取り位置と関連付けてそれぞれ記憶する分割画像記憶手段。
隣接する二つの分割画像を接続する際に基準とする方の分割画像を基準画像、他方の分割画像を接続画像とし、前記接続画像の任意の一画素を探索基準画素とし、前記基準画像に対して前記接続画像を一定の範囲内で相対的に移動させるとともに、両画像の重複部分の一定面積領域における両画像の対応画素間の濃度差の平均値を求める処理と、それに応じて両画像の接続位置を決定するに際し、先ず、前記探索基準画素の位置を中心とし前記読取機構の位置精度に応じた面積の粗探索領域内で、前記探索基準画素を数画素間隔で移動させ、前記濃度差の平均値が最小となるときの探索基準画素の位置を基準画像と接続画像との暫定接続位置とし、次に、前記暫定接続位置を中心とし前記粗探索領域より面積の狭い密探索領域の範囲内で、前記探索基準画素を一画素ずつ移動させ、前記濃度差の平均値が最小となるときの前記探索基準画素の位置を基準画像と接続画像との接続位置を決定する接続位置算出手段。
前記接続位置における両分割画像の重複領域の画素値を、各画素毎にいずれか一方の分割画像の当該画素の画素値とする重複画素選択手段。
前記接続位置における両分割画像の重複領域の画素値を、各画素毎に両分割画像の当該画素の画素値を用いた演算値とする重複画素演算手段。
前記接続位置における両分割画像の重複領域の対応する画素間の画素値の差に応じて前記重複画素選択手段または重複画素演算手段を選択する合成手段選択手段。
【００２４】
上記画像読取装置によれば、隣接する二つの分割画像の接続において、重複領域の不連続性を表す対応する画素間の画素値の差に応じて、重複画素選択手段を選択して各画素毎にいずれか一方の分割画像の当該画素の画素値で置き換えるか、または、重複画素演算手段を選択して各画素毎に両分割画像の当該画素の画素値を用いた演算値とするので、分割画像間の継ぎ目の顕在化を防止するとともに、処理負担を軽減して接続処理の効率化を図ることができる。
【００２５】
【発明の実施の形態】
以下、本発明に係る画像読取方法及び画像読取装置の実施の形態について図面を参照しながら説明する。
図１は本発明に係る画像読取装置の構成例の外観図、図２は本発明に係る画像読取装置の主要部を示す制御ブロック図である。
画像読取装置は、原稿１０６の撮像を行う読取機構１００と、読取機構１００の位置制御及び駆動等の指示を行う一方、読取機構１００において撮像された画像デ−タを取り込んで処理する制御部４００と、解像度や読取り開始を指示する入力部５００とから構成される。
【００２６】
読取機構１００は、さらに、倍率可変な結像光学手段及び２次元ＣＣＤなどの光電変換素子とからなる撮像部１００ａと、撮像部１００ａを方向可変に駆動するとともに光電変換素子の電気的動作を制御する撮像制御部１００ｂと、これらを支持する支持台１００ｃとからなる。
撮像部１００ａは、撮像制御部１００ｂにより、Ｘ−Ｘ線を中心とする円周方向（矢印ｘ）、紙面表裏方向に延びる軸Ｙを中心とする円周方向（矢印ｙ）、Ｚ−Ｚ線を中心とする円周方向（矢印ｚ）、の各方向について回動自在となっている。
制御部４００は、撮像部１００ａの位置制御及び駆動等の指示を行うとともに、撮像部１００ａにおいて撮像された画像デ−タを取り込み、分割画像の接続処理等の各種画像処理を行い、具体的にはパ−ソナルコンピュ−タ４等に対応する。
また、入力部５００は、具体的にはパ−ソナルコンピュ−タに接続されるキ−ボ−ド５やマウス等に対応する。
【００２７】
撮像制御部１００ｂと制御部４００とは、シリアルインタ−フェ−ス、例えばＲＳ２３２Ｃケ−ブル４０４を介して接続されており、このケ−ブルを介して制御部４００から撮像制御部１００ｂへ画像読取り位置、結像光学手段の倍率などの指示を送出し、撮像制御部１００ｂから制御部４００へ撮像部１００ａの実際の位置情報等を送出する。
一方、撮像部１００ａと制御部４００とは例えばＮＴＳＣビデオ信号線４０３を介して接続され、撮像部１００ａにおいて撮像された画像デ−タは、ＮＴＳＣビデオ信号線４０３によって送出され、ビデオキャプチャ−カ−ド４０２により２次元画像に変換されて制御部４００へ送出される。
【００２８】
次に、上記画像読取装置の主要部を図２に示す制御ブロック図に沿って説明する。
制御部４００は、読取機構１００の制御及び読取機構１００より取り込んだ画像デ−タ処理の制御等全体的な制御を行う処理制御部４１０と、読取機構１００から送出されるＮＴＳＣビデオ信号を２次元画像に変換する画像信号変換部（ビデオキャプチャ−カ−ド）４０２と、複数の２次元画像を読取り位置と関連付けて記憶する一次画像メモリ４３０と、原稿を斜めから撮像したことによる歪みを補正する透視変換及び撮像した画像の任意の解像度への調整を行う透視変換・解像度調整部４４０と、複数の画像を接続する画像合成部４５０と、画像合成部４５０において重複領域の合成された接続画像を出力させる出力Ｉ／Ｆ４７０とから構成される。
【００２９】
上記画像読取装置による画像読取りは、以下のように行われる。
先ず、机上等に原稿１０６を設置した後、入力部５００より解像度指定及び読取り開始指示が行われると、処理制御部４１０は撮像制御部１００ｂに対して原稿１０６全体を撮像するよう指示を行い、原稿１０６が１画面内に収まるように撮像部１００ａの光学結像系が自動的にズ−ム、フォ−カシングされ、原稿１０６全体が撮像され、画像信号変換部４０２においてＮＴＳＣビデオ信号から２次元画像に変換される。
この原稿全体の画像は、必要に応じて一次画像メモリ４３０、二次画像メモリ４５４、出力Ｉ／Ｆ４７０を介してディスプレイ等に出力される。
続いて、処理制御部４１０は、原稿のサイズ演算を行い、この原稿１０６のサイズと指定された解像度とから画像の分割数を決定し、撮像制御部１００ｂに対して原稿１０６の各分割画像を順次読み取るよう位置制御及び光電変換素子の駆動の指示を送出し、各分割画像が順次撮像される。
【００３０】
撮像部１００ａにおいて撮像された各分割画像は、画像信号変換部４０２においてＮＴＳＣビデオ信号から２次元画像に変換され、画像デ−タと共に送出される撮像部１００ａの実際の読取り位置と関連付けて一次画像メモリ４３０に記憶される。
一次画像メモリ４３０に記憶されている各分割画像は撮像部１００ａによって斜めに撮像されているので、透視変換・解像度調整部４４０においてこの歪みを含んだ画像を正面から撮像した画像に変換させる透視変換を行い、同時に指定された解像度の画像となるよう調整を行う。
そして、一次画像メモリ４３０に記憶された各分割画像を後述する手順により接続して原稿１０６に対応する一つの画像とし、必要に応じて出力Ｉ／Ｆ４７０を介して外部へ出力する。
【００３１】
次に、隣接する分割画像の接続処理を行う制御部４００内の画像合成部４５０について説明する。
画像合成部４５０は、接続対象となる隣接する二つの分割画像の画像デ−タ及び位置情報を取り込み両分割画像の接続位置を算出する接続位置算出部４５２と、接続位置における両分割画像の重複領域の各画素値をそれぞれ記憶する重複領域メモリ４５５と、重複領域以外の領域の両分割画像の画素値をそれぞれ記憶する被接続画像メモリ４５１と、重複領域の画素値を各画素毎にいずれの分割画像の当該画素の画素値で置き換えるかを判断し出力する重複画素選択判断部４５３と、接続後の画像を記憶する二次画像メモリ４５４と、これらの各部を制御する合成制御部４６０とから構成される。
【００３２】
さらに、重複画素選択判断部４５３は、重複領域の各画素値をいずれの分割画像の当該画素の画素値で置き換えるかを、重複領域における当該画素の位置に基づいて判断するために、予め位置に関する判断基準を規定し、重複領域の画素値決定時に重複領域メモリ４５５より当該画素の重複領域における位置情報を取り込み、その位置に応じて出力結果を変化させる。
【００３３】
上記画像合成部４５０における画像の接続処理を、図３に示すフロ−チャ−トに沿って説明する。
先ず、接続位置算出部４５２において、一次画像メモリ４３０より接続対象となる隣接する二つの分割画像Ａ及びＢの画像デ−タ及び位置情報を取り込み（ステップ３００）、これらをもとに後述する手順により接続位置を算出する（ステップ３０１）。
次に、その接続位置における両分割画像の重複領域を算出し（ステップ３０２）、重複領域の両分割画像の各画素値をそれぞれ重複メモリ４５５へ記憶し、重複領域以外の両分割画像の各画素値をそれぞれ被接続画像メモリ４５１へ記憶する。
【００３４】
そして、重複領域に属さない画素については、接続前の分割画像の元の画素値を被接続画像メモリ４５１より二次画像メモリ４５４へ出力する（ステップ３０４）。
一方、重複領域に属す画素については、重複画素選択判断部４５３において予め規定される判断基準によって、各画素毎に分割画像Ａまたは分割画像Ｂのいずれか一方の分割画像の元の画素値を、重複領域メモリ４５５より二次画像メモリ４５４へ出力する（ステップ３０６，３０７）。
全画素についてこの処理を繰り返し（ステップ３０８）、二次画像メモリ４５４に両分割画像を接続した接続画像の各画素値が記憶される（ステップ３０９）。
【００３５】
この画像合成部４５０における各処理は、専用の画像処理回路の一部の回路としてのハ−ドウェアの形態、汎用の画像処理回路に設定するソフトウエアあるいはファームウエアの形態、あるいは汎用の計算機システムの上で稼働するソフトウエアの一部としての形態、これらの形態の組み合わせ、等種々の形態での実現が可能である。
ソフトウェアとして実現する場合は、予め上記の処理を行うプログラムが記録された媒体をコンピュ−タにより随時読み込みながら処理を行う。例えば処理プログラムを記録したフロッピ−ディスクをフロッピ−ディスクドライブ４０５にセットして随時読み込みながら処理を行う形態や、パ−ソナルコンピュ−タ４に内蔵されているハ−ドディスク（図示せず）等に処理プログラムを記録させる形態が考えられる。
【００３６】
次に、接続位置算出部４５２における接続位置の算出方法について図４〜図９を参照しながら説明する。この接続位置の算出は、粗探索及び密探索の２段階の探索により効率的に行われる。
左右に隣接する二つの分割画像１及び２は撮像部１００ａの実際の読取り位置に従って配置した状態で重複部分１０を有する。このとき、分割画像２は、分割画像１の読取り後読取機構１００を右方向へ移動させて読み取ったものであるが、読取機構１００はその位置精度に応じて予定の読取り位置とずれた位置を読み取る可能性があるため、上下方向にずれた画像として表示している。
以後、分割画像１を基準画像とし、他方の分割画像２を接続画像とする。また、接続画像２の左上端に位置する画素を探索基準画素２ａとする。
【００３７】
先ず、探索基準画素２ａを中心とする左右方向長Ｗ、上下方向長Ｈの矩形領域である粗探索領域２０内に、数画素間隔に配置した格子点２１を設定し、最も左上に位置する格子点を２１ａ、最も右下に位置する格子点を２１ｃとする（図５）。ここでは３画素間隔で格子点２１を配置し、各格子点間の長さはＤとする。次に、探索基準画素２ａが格子点２１ａと一致するよう、接続画像２全体を左上方向へ移動させ、基準画像１と接続画像２との重複部分１０の中央付近に、両画像の対応画素間の画素値の差（以後、濃度差という）の平均値を算出する濃度差算出領域１１を設定し（図６）、濃度差算出領域１１における濃度差の平均値を算出する。
【００３８】
以後、探索基準画素２ａが格子点２１ｃまでの各格子点と一致するよう接続画像２全体を順次移動させながら、基準画像１と接続画像２との重複部分１０の中央付近に設定した濃度差算出領域１１′の濃度差の平均値を算出する（図７）。そして、濃度差の平均値が最小となった格子点を検出し、暫定接続位置３０とする（図８）。
【００３９】
次に、暫定接続位置３０を用いて、基準画像１と接続画像２の接続位置をさらに細かく探索する。
先ず、暫定接続位置３０を中心とする左右方向長ｗ、上下方向長ｈの矩形領域である密探索領域４０内に、１画素間隔の格子点４１を設定し、最も左上に位置する格子点を４１ａ、最も右下に位置する格子点を４１ｃとする（図８）。
そして、粗探索領域２０の場合と同様に、探索基準画素２ａを格子点４１ａから４１ｃまでの各格子点と一致するよう接続画像２全体を順次移動させながら、基準画像１と接続画像２との重複部分の中央付近に設定した濃度差算出領域における対応画素間の濃度差の平均値を算出する。
そして、濃度差の平均値が最小となった格子点を検出し、基準画像１と接続画像２との接続位置５０とする（図９）。
【００４０】
上記接続位置の算出方法によれば、分割画像撮像時の実際の読取り位置をもとに一定の範囲内で、重複領域の濃度差にもとづいて接続位置を探索するので、読取機構１００の位置精度に応じて読取り画像の位置ずれが生じた場合においても正しい接続位置を決定することができる。
また、接続位置の算出は、粗探索及び密探索の２段階の探索により効率的に行われるので、処理時間が短縮される。
【００４１】
以下に、上記本発明に係る画像読取方法及び画像読取装置において隣接する二つの分割画像を接続する際の、重複領域の各画素値の決定方法の実施例について説明する。各実施例においては、前述の方法により接続位置が算出されたものとする。
【００４２】
（実施例１）
図１０に示すように、分割画像Ａと分割画像Ｂとを接続する場合で、重複領域Ｃの左上端に位置する画素を原点として左右方向の位置がi 、上下方向の位置がj の画素の接続後の画素値をＣ[i,j] とし、同様に位置[i,j] における分割画像Ａの画素値をＡ[i,j］、分割画像Ｂの画素値をＢ[i,j] とすると、重複画素選択判断部４５３（図２）において、ディザ配列を用いて次の式１で表されるように両分割画像の画素値が交互に採用される場合について説明する。
このとき、重複領域Ｃの左右方向にはＮ画素（i=０〜Ｎ−１）、上下方向にはＭ画素（j=０〜Ｍ−１）の画素が含まれ、Ｎ：Ｍ＝１：５程度とする。
【００４３】
【式１】

【００４４】
上式によれば、重複領域の画素値は交互にいずれか一方の分割画像の画素値をそのまま採用することとなり、図１１に示すように、市松模様状に両分割画像の画素値が配置されることになる。尚、図においては、重複領域の画素値が分割画像Ａと分割画像Ｂの画素値を交互に採用していることが確認されるよう、便宜上分割画像Ａの画素値と分割画像Ｂの画素値をそれぞれ均一の濃度として表示している。
本実施例によれば、重複領域の画素値として両分割画像の画素値を規則的に交互に採用しているので、重複領域の端部において重複領域の内部と外部の境目が目立ちにくくなり、算出された接続位置に微小な誤差が存在する場合でも、画素値の不連続を良好に分散することができ、隣接する二つの画像を接続して継ぎ目の目立たない接続画像を得ることができる。
【００４５】
（実施例２）
図２の重複画素選択判断部４５３において、ランダムに０または１を発生する疑似一様乱数発生器（その出力結果をRAND[0,1] と表示）を用いて、次の式２で表されるように画素毎に発生した乱数によって採用する画素値を決定する場合について説明する。
【００４６】
【式２】

【００４７】
本実施例によれば、重複領域の画素値として両分割画像の画素値をランダムに採用しているので、重複領域の端部がいずれか一方の分割画像の画素値のみを配置した直線になりにくく、重複領域の内部と外部との境界が目立ちにくいため、算出された接続位置に微小な誤差が存在する場合でも、画素値の不連続を良好に分散することができ、隣接する二つの画像を接続して継ぎ目の目立たない接続画像を得ることができる。
【００４８】
（実施例３）
図２の重複画素選択判断部４５３において、次の式３に示すように重複領域を斜めに二分する対角線を境界としていずれの分割画像の画素値を採用するかを決定する場合について説明する。
【００４９】
【式３】

【００５０】
本実施例によれば、図１２に示すように、両分割画像の継ぎ目が斜めの線分となるので、目の特性としてよく知られている水平および垂直の直線状の形状への高い感度を避けることができ、算出された接続位置に微小な誤差が存在する場合でも、隣接する二つの画像を接続して見掛上継ぎ目の目立たない接続画像を得ることができる。
本実施例の場合には、加算器あるいはソフトウェアではなく、プログラマブルなカウンターの組合せによる回路にて簡単に実現可能である。
【００５１】
次に、本発明に係る画像読取方法及び画像読取装置の他の実施の形態について説明する。図１３は、画像読取装置の主要部を示す制御ブロック図である。尚、図において図２と同様の構成には同一符号を付し、説明を省略する。また、この場合の画像読取装置の外観は、図１と同様である。
制御部４００′の画像合成部４５０′は、重複領域の両分割画像の当該画素値を用いて何らかの演算を行いその演算値を当該画素の画素値として出力する重複画素演算部４５６を備え、さらに、重複領域の両分割画像の対応する画素値を用いて重複領域における当該画素の位置に基づく重み演算を行うために重み係数生成部４５７を備え、合成制御部４６０′により、予め重み係数を発生させておき、重複領域の画素値決定時に重複画像メモリ４５５より当該画素の重複領域における位置情報を取り込み、その位置に応じて重み係数を用いて演算を行い、二次画像メモリ４５４に出力させるという点において、図２の画像合成部４５０と相違する。上記重み係数生成部４５７は重複画素演算部４５６内に設けてもよい。
【００５２】
上記画像合成部４５０′における画像の接続処理を、図１４に示すフロ−チャ−トに沿って説明する。
先ず、接続位置算出部４５２において、一次画像メモリ４３０より接続対象となる隣接する二つの分割画像Ａ及びＢの画像デ−タ及び位置情報を取り込み（ステップ３１０）、これらをもとに前述の手順により接続位置を算出する（ステップ３１１）。
次に、その接続位置における両分割画像の重複領域を算出し（ステップ３１２）、重複領域の両分割画像の各画素値をそれぞれ重複メモリ４５５へ記憶し、重複領域以外の両分割画像の各画素値をそれぞれ被接続画像メモリ４５１へ記憶する。
続いて、重み係数生成部４５７において重み係数を生成する（ステップ３１３）。
【００５３】
そして、重複領域に属さない画素については、接続前の分割画像の元の画素値を被接続画像メモリ４５１より二次画像メモリ４５４へ出力する（ステップ３１５）。
一方、重複領域に属す画素については、重複画素演算部４５６において重複領域メモリ４５５より両分割画像の当該画素の画素値を取り込み、予め生成された重み係数を用いて演算を行い、その演算値を二次画像メモリ４５４へ出力する（ステップ３１６）。
全画素についてこの処理を繰り返し（ステップ３１７）、二次画像メモリ４５４に両分割画像を接続した接続画像の各画素値が記憶される（ステップ３１８）。
この画像合成部４５０′における各処理は、画像合成部４５０（図２）と同様、ハ−ドウェア、ソフトウェア、ファ−ムウェア、等種々の形態により実現可能である。
【００５４】
以下に、上記本発明に係る画像読取方法及び画像読取装置において隣接する二つの分割画像を接続する際の、重複領域の各画素値の決定方法の実施例について説明する。各実施例においては、前述の方法により接続位置が算出されたものとする。
【００５５】
（実施例４）
図１３の重複画素演算部４５６において、次の式４〜６で表されるように両分割画像の画素値に左右方向に傾斜した重み付き演算を行った場合について説明する。
ここで、傾斜した重みとは、図１０を参照して説明すると、重複領域の画素値の算出において図面左右方向について、分割画像Ａ及び分割画像Ｂの元の画素値を用いる配分を、左端部でＡ：Ｂ＝１：０、中央部分でＡ：Ｂ＝０．５：０．５、右端部でＡ：Ｂ＝０：１となるように線形に比率が変化するようにした重みであり、ｆ1(i,j)，ｆ2(i,j)はこの比率を表す重み係数である。
【００５６】
【式４】
Ｃ[i,j] ＝Ａ[i,j] ×ｆ1(i,j)＋Ｂ[i,j] ×ｆ2(i,j)
【００５７】
【式５】
ｆ2(i,j)＝i ／（Ｎ−１）
【００５８】
【式６】
ｆ1(i,j)＝１−ｆ2(i,j)
【００５９】
本実施例によれば、重複領域の分割画像Ａ側の端部は分割画像Ａの画素値をそのまま採用し、一方分割画像Ｂ側の端部は分割画像Ｂの画素値をそのまま採用し、重複領域の左右方向の中央においては各画素値を両分割画像の画素値の平均値とするので、重複領域の左右方向で左に位置する程分割画像Ａの画素値をより重く用い、右に位置する程分割画像Ｂの画素値をより重く用い、算出された接続位置に微小な誤差が存在する場合でも、重複領域とその両端の分割画像との境界における画素値を連続的に変化させることにより画素値の不連続を良好に分散することができ、継ぎ目の目立たない接続画像を得ることができる。
尚、本実施例においてはＮ：Ｍ＝１：５程度で重複領域が縦長形状となる場合を想定しており、接続部分は左右方向の継ぎ目が特に目立つと考えられるため、左右方向についてのみ重み演算を行っているが、さらに上下方向に関して重み演算を行ってもよい。
【００６０】
（実施例５）
図１５の重複画素演算部４５６において、前記式４、式６及び次の式７及び８に示すように両分割画像の画素値に重み付き演算を行い、その重み係数にディザ配列を用いた場合について説明する。
【００６１】
【式７】

【００６２】
【式８】
ｆ2(i,j)＝Ｄ［（i mod n）＋１，（j mod m）＋１］： n=4,m=4
【００６３】
本実施例によれば、重複領域の隣接する１６画素において、分割画像Ａまたは分割画像Ｂの画素値を異なる配分で用いて画素値を決定した画素同志が隣接することにより、算出された接続位置に微小な誤差が存在する場合でも、画素値の不連続を良好に分散することができ、隣接する二つの画像を接続して継ぎ目の目立たない接続画像を得ることができる。
【００６４】
次に、本発明に係る画像読取装置の他の実施の形態について説明する。図１５は、画像読取装置の主要部を示す制御ブロック図である。尚、図において図２及び図１３と同様の構成には同一符号を付し、説明を省略する。また、この場合の画像読取装置の外観は、図１と同様である。
制御部４００″の画像合成部４５０″は、重複領域の画素値を各画素毎にいずれの分割画像の当該画素の画素値で置き換えるかを判断し出力する重複画素選択判断部４５３と、重複領域の両分割画像の当該画素値を用いて何らかの演算を行い、その演算値を当該画素の画素値として出力する重複画素演算部４５６とを同時に備え、さらに、重複画素選択判断部４５３または重複画素演算部４５６のいずれか一方を選択する合成手段選択部４５８を備え、合成制御部４６０″により、重複領域の画素値決定時に、接続位置算出部４５２において二つの分割画像の接続位置を決定する際に算出された重複領域の濃度差の平均値に基づいて合成手段選択部４５８にいずれかの合成手段を選択させ、各合成手段により合成された重複領域の各画素値を二次画像メモリ４５４に出力させるという点において、図２の画像合成部４５０、あるいは、図１３の画像合成部４５０′と相違する。
この画像合成部４５０″における各処理は、画像合成部４５０（図２）と同様、ハ−ドウェア、ソフトウェア、ファ−ムウェア、等種々の形態により実現可能である。
【００６５】
以下に、上記本発明に係る画像読取装置において隣接する二つの分割画像を接続する際の、重複領域の各画素値の決定方法の実施例について説明する。
【００６６】
（実施例６）
図１５において、重複画素選択判断部４５３、重複画素演算部４５６以外に、接続する二つの分割画像の重複領域の画素値をいずれか一方の分割画像の対応する画素値で全体的に置き換える重複画素置換部４５９（図示せず）を備え、合成手段選択部４５８においてこれらの３つの手段を選択可能とした場合の上記画像合成部４５０″における画像の接続処理を、図１６に示すフロ−チャ−トに沿って説明する。
【００６７】
先ず、接続位置算出部４５２において、一次画像メモリ４３０より接続対象となる隣接する二つの分割画像Ａ及びＢの画像デ−タ及び位置情報を取り込み（ステップ３２０）、これらをもとに前述の手順により接続位置を算出する（ステップ３２１）。
次に、その接続位置における両分割画像の重複領域を算出し（ステップ３２２）、重複領域の両分割画像の各画素値をそれぞれ重複メモリ４５５へ記憶し、重複領域以外の両分割画像の各画素値をそれぞれ被接続画像メモリ４５１へ記憶する。
続いて、合成手段選択部４５８において、接続位置を決定する際に算出された濃度差の平均値に応じて重複画素選択判断部４５３または重複画素演算部４５６または重複画素置換部４５９のいずれか一つを選択する。例えば、画素値のダイナミックレンジをＤとすると、濃度差の平均値がＤ／４以下の場合（ステップ３２３，Ｙｅｓ）は重複画素置換部４５９を、濃度差の平均値がＤ／４以上かつＤ／２未満の場合（ステップ３２４，Ｙｅｓ）は重複画素選択判断部４５３を、濃度差の平均値がＤ／２以上の場合（ステップ３２４，Ｎｏ）は重複画素演算部４５６をそれぞれ選択するよう予め規定する。
【００６８】
重複画素置換部４５９が選択された場合は、重複領域の画素値をいずれの分割画像の画素値で全体的に置き換えるかを、例えばＡ画像を選択する（ステップ３２５）。
そして、重複領域に属さない画素については、接続前の分割画像の元の画素値を被接続画像メモリ４５１より二次画像メモリ４５４へ出力する（ステップ３２７）。
一方、重複領域に属する画素については、重複画素置換部４５９において、予め選択されたＡ画像の当該画素の画素値を二次画像メモリ４５４へ出力する（ステップ３２８）。
全画素についてこの処理を繰り返し（ステップ３２９）、二次画像メモリ４５４に両分割画像を接続した接続画像の各画素値が記憶される（ステップ３３０）。
【００６９】
重複画素選択判断部４５３が選択された場合は、図３中の１へジャンプして以下の処理を行い、同様に重複画素演算部４５６が選択された場合は、図１４中の２へジャンプして以下の処理を行う。重複画素選択判断部４５３の具体的な画素値決定方法としては、上記実施例１〜３のいずれかを使用し、また、重複画素演算部４５６については、上記実施例４または５のいずれかを使用することができる。
そして、重複領域の各画素値を濃度差の平均値に応じて合成し、二次画像メモリ４５４には隣接する二つの分割画像を接続した接続画像の各画素値が記憶される（図３・ステップ３０９、図１４・ステップ３１８）。
【００７０】
本実施例によれば、重複領域の濃度差の平均値が小さく分割画像の継ぎ目が比較的目立ちにくいと考えられる場合は、重複領域の各画素値をいずれか一方の分割画像の対応画素値で単純に置き換えて、接続処理の負担を軽減することができる。
一方、濃度差の平均値が大きく分割画像の継ぎ目が顕著に現れると考えられる場合は、ディザ配列重みによる平均化等を採用し、接続位置の誤差に起因する画素値の不連続を良好に分散させることができる。
また、上記のように、重複領域の合成手段を濃度差の平均値に応じて自動的に選択可能とすることにより、接続画像に応じた経済的で効果的な接続処理を行うことができる。
【００７１】
尚、合成手段選択部４５８における各合成手段を選択する際の閾値、選択された合成手段における具体的な処理内容との組み合わせは上記説明に限られない。また、指定された解像度や原稿の種類等によって各合成手段を選択可能とする方法や、自動的に各合成手段の選択を行って処理内容を切り替えることも考えられる。
【００７２】
本発明の画像読取装置の他の実施の形態として、例えば、光電変換素子を１次元ＣＣＤ素子のスキャンタイプで構成した光学結像系、光学結像系にミラーを内蔵して光軸方向可変な制御機構部を備えたもの、分割画像の接続処理と光学結像系のズーム，フォーカシング，各分割画像の読み取り位置制御を行う専用コントローラを含む装置、光電変換素子からディジタル画像へ変換するための中間信号にＮＴＳＣ信号以外のものを用いるものや、アナログ信号を用いずディジタル画像出力を利用するもの、などのバリエーションがあり得、これらの組み合わせも可能である。
【００７３】
【発明の効果】
本発明によれば、隣接する分割画像を接続する際に、重複領域の画素値をいずれか一方の分割画像の画素値で全体的に置き換えるのではなく、画素毎にいずれか一方の分割画像の当該画素値で置き換えるか、または、両分割画像の当該画素の画素値を用いた演算値により置き換えることにより重複領域を合成するので、両分割画像の接続位置に微小な誤差が存在する場合においても、画素値の不連続を良好に分散して継ぎ目の目立たない接続画像を得ることができる。従って、複数の分割画像を同様に接続することにより、画像原稿に忠実で高品質な読取り画像を任意の解像度で得ることができる。
また、上記合成手法を自動的に選択可能とすることにより、画素値の不連続を効果的に分散するとともに、処理負担を軽減して実用に耐える接続処理を行うことができる。
【図面の簡単な説明】
【図１】本発明に係る画像読取装置の外観説明図である。
【図２】本発明に係る画像読取装置の主要部を示す制御ブロック図である。
【図３】本発明に係る画像読取方法の一部を示すフロ−チャ−トである。
【図４】本発明に係る画像読取方法における接続位置算出方法の手順を示す説明図である。
【図５】本発明に係る画像読取方法における接続位置算出方法の手順を示す説明図である。
【図６】本発明に係る画像読取方法における接続位置算出方法の手順を示す説明図である。
【図７】本発明に係る画像読取方法における接続位置算出方法の手順を示す説明図である。
【図８】本発明に係る画像読取方法における接続位置算出方法の手順を示す説明図である。
【図９】本発明に係る画像読取方法における接続位置算出方法の手順を示す説明図である。
【図１０】本発明に係る画像読取方法における重複領域内の各画素の位置を示す説明図である。
【図１１】実施例１による隣接する分割画像の接続結果を示す説明図である。
【図１２】実施例３による隣接する分割画像の接続結果を示す説明図である。
【図１３】本発明に係る画像読取装置の主要部を示す制御ブロック図である。
【図１４】本発明に係る画像読取方法の一部を示すフロ−チャ−トである。
【図１５】本発明に係る画像読取装置の主要部を示す制御ブロック図である。
【図１６】本発明に係る画像読取方法の一部を示すフロ−チャ−トである。
【符号の説明】
１００…読取機構、１００ａ…撮像部、１００ｂ…撮像制御部、１００ｃ…支持台、１０６…原稿、４００…制御部、４０２…画像信号変換部（ビデオキャプチャ−カ−ド）、４０３…ＲＳ２３２Ｃケ−ブル、４０４…ＮＴＳＣビデオ信号線、４１０…処理制御部、４３０…一次画像メモリ、４４０…透視変換・解像度調整部、４５０，４５０′，４５０″…画像合成部、４５１…被接続画像メモリ、４５２…接続位置算出部、４５３…重複画素選択判断部、４５４…二次画像メモリ、４５５…重複領域メモリ、４５６…重複画素演算部、４５７…重み係数生成部、４５８…合成手段選択部、４６０，４６０′，４６０″…合成制御部、４７０…出力Ｉ／Ｆ、５００…入力部[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image reading method and an image reading apparatus for dividing an original image into a plurality of divided images, sequentially reading the images, and connecting the divided images to obtain original image data. The present invention relates to an image reading method and an image reading apparatus that connect without making a seam visible and obtain a read image faithful to a document image, and further to a medium on which an image reading program is recorded.
[0002]
[Prior art]
Conventionally, as an image reading apparatus used for a personal computer, a scanner for a word processor, etc., for example, an original image is divided into a plurality of divided images according to the original size and a specified resolution, and the magnification is A scanning mechanism that scans a variable imaging optical means and a reading mechanism that holds a solid-state imaging device, sequentially captures images at a necessary magnification, and connects divided images to obtain original image data at an arbitrary resolution. An image reading apparatus has been proposed (see Japanese Patent Publication No. 8-13088).
[0003]
According to this image reading apparatus, it is possible to reduce the size of the apparatus, and at the same time, it is possible to read a document image at an arbitrary resolution, and to read a three-dimensional image as well as a reading target is not limited to a planar image. be able to.
However, in the image reading apparatus configured as described above, since the distance between the read document and the image sensor is large, the read image is not the intended image even when the reading mechanism holding the image sensor is slightly displaced from the desired reading position. Will be greatly deviated.
[0004]
In order to prevent the positional deviation of the read image, if a precise positional accuracy is required for the mechanical system that controls the direction and position of the reading mechanism, the apparatus becomes large. Therefore, a method is considered in which each divided image is read in an overlapping manner with adjacent divided images, and when the adjacent divided images are connected, the actual connection position is determined based on the pixel values of both divided images in the overlapping portion. It is done.
Specifically, with respect to one of the two divided images to be connected, while the other is relatively moved within a certain range, the difference in pixel value between corresponding pixels in the overlapping region of both divided images is minimized. Is set as the connection position of both divided images.
[0005]
[Problems to be solved by the invention]
According to the above method, it is possible to determine the connection position of adjacent divided images regardless of the position accuracy of the reading mechanism.
Then, the pixel value of each pixel in the overlapping area of both divided images at the connection position is determined and the both divided images are connected. As a method for determining the pixel value, each pixel value in the entire overlapping area is divided into one of the divided areas. The simplest method is to completely replace the corresponding pixel value of the image.
However, according to this method, when there is a slight error in the connection position, the boundary between the overlapped area and the other divided image is clear due to a shift that occurs in the pixel value between corresponding pixels of both divided images in the overlapped area. There was a problem of being recognized. For this reason, a document image obtained by connecting a plurality of divided images is an image including an image quality defect at a connection location.
[0006]
The present invention has been made in view of the above circumstances, and reads a document image into a plurality of divided images, sequentially reads them, and connects them without revealing the boundary seams between the divided images. An image reading method and an image reading apparatus capable of obtaining a high-quality read image faithful to an arbitrary resolution, and a medium on which a program for image reading is recorded.
[0007]
[Means for Solving the Problems]
  In order to achieve the above object, an image reading method according to claim 1 divides an original image into a plurality of divided images, sequentially reads each divided image with an adjacent divided image, and connects the divided images to connect the divided image. An image reading method for obtaining the following data:
When connecting two adjacent divided images, the reference divided image is the reference image, the other divided image is the connection image, and any one pixel of the connection image is the search reference pixel, and the reference image The connected image is relatively moved within a certain range, and the average value of the density difference between the corresponding pixels of both images in a certain area of the overlapping portion of both images, In determining the connection position, first, the search reference pixel is moved at intervals of several pixels within a rough search region having an area corresponding to the position accuracy of the reading mechanism with the position of the search reference pixel as the center, and the density difference The position of the search reference pixel when the average value of is the minimum is the provisional connection position between the reference image and the connection image,
Next, the search reference pixel is moved one pixel at a time within the range of the dense search region having an area smaller than the coarse search region with the provisional connection position as the center, and the average value of the density difference is minimized. The position of the search reference pixel is the connection position between the reference image and the connection image,The two divided images are connected by setting the pixel value of the overlapping area of the two divided images at the connection position to the pixel value of one of the divided images for each pixel.
[0008]
  The image reading apparatus according to claim 2 holds an imaging optical unit capable of changing a magnification, and a solid-state imaging device that captures an optical image of a document image formed using the imaging optical unit. The position of the reading mechanism unit and the reading mechanism unit is controlled so as to capture the entire document image, and the document image is divided into a plurality of divided images and sequentially captured at a necessary magnification. And an image reading apparatus that captures and connects each divided image and obtains data of a document image having an arbitrary resolution. The control means includes the following means.
Divided image storage means for storing the pixel value data of each divided image in association with the reading position on the document image.
When connecting two adjacent divided images, the reference divided image is the reference image, the other divided image is the connection image, and any one pixel of the connection image is the search reference pixel, and the reference image The connected image is relatively moved within a certain range, and the average value of the density difference between the corresponding pixels of both images in a certain area of the overlapping portion of both images, In determining the connection position, first, the search reference pixel is moved at intervals of several pixels within a rough search region having an area corresponding to the position accuracy of the reading mechanism with the position of the search reference pixel as the center, and the density difference The position of the search reference pixel when the average value of the reference value is the minimum is the provisional connection position between the reference image and the connection image, and then the range of the dense search region that is centered on the provisional connection position and smaller in area than the coarse search region In front The search reference pixel moves by one pixel, the connection position between the reference image and the connection image the position of the search reference pixel when the average value of the density difference is minimumConnection position calculation means for determining
Overlapping pixel selection means for setting the pixel value of the overlapping area of both divided images at the connection position to the pixel value of the pixel of any one of the divided images for each pixel.
[0009]
According to the image reading method and the image reading apparatus described above, in connecting two adjacent divided images, the pixel value of one of the divided images is adopted for each pixel of the overlapping region. Since it is difficult to form a straight line in which only the pixel values of one of the divided images are arranged, and the boundary between the inside and the outside of the overlapping region is not conspicuous, even if there is a slight error in the determined connection position, Discontinuity can be well dispersed and the appearance of seams can be prevented.
[0010]
The image reading apparatus according to claim 3 is the image reading apparatus according to claim 2, wherein the overlapping pixel selection unit determines each pixel value of the overlapping area of the two divided images based on the position of the pixel in the overlapping area. Thus, it is characterized in determining which divided image the pixel value of the pixel is set to.
[0011]
According to the above-described image reading apparatus, dispersion of pixel value discontinuities due to connection position errors can be uniformly performed at each connection portion between adjacent divided images. It is possible to mitigate image quality defects in the read image.
[0012]
The image reading device according to claim 4 is the image reading device according to claim 3, wherein the overlapping pixel selection unit determines each pixel value of the overlapping region of the two divided images, and the position of the pixel in the overlapping region is: Based on which of the two areas divided by the diagonal line that bisects the overlapping area diagonally, it is determined which pixel value of the pixel of which divided image is used.
[0013]
According to the image reading apparatus, since the joint between the pixel values of both divided images in the overlapping region is an oblique line segment in the connection of two adjacent divided images, horizontal and well-known characteristics of the eye are known. High sensitivity to a vertical linear shape can be avoided, and apparent seams can be prevented from becoming apparent.
[0016]
  Claim 5In this image reading method, an original image is divided into a plurality of divided images, each divided image is overlapped with an adjacent divided image, sequentially read, and each divided image is connected to obtain original image data. Because
When connecting two adjacent divided images, the reference divided image is the reference image, the other divided image is the connection image, and any one pixel of the connection image is the search reference pixel, and the reference image The connected image is relatively moved within a certain range, and the average value of the density difference between the corresponding pixels of both images in a certain area of the overlapping portion of both images, In determining the connection position, first, the search reference pixel is moved at intervals of several pixels within a rough search region having an area corresponding to the position accuracy of the reading mechanism with the position of the search reference pixel as the center, and the density difference The position of the search reference pixel when the average value of is the minimum is the provisional connection position between the reference image and the connection image,
Next, the search reference pixel is moved one pixel at a time within the range of the dense search region having an area smaller than the coarse search region with the provisional connection position as the center, and the average value of the density difference is minimized. The position of the search reference pixel is the connection position between the reference image and the connection image,
The two divided images are connected by setting the pixel value of the overlapping area of the two divided images at the connection position to a calculated value using the pixel value of the pixel of the two divided images for each pixel. .
[0017]
  Claim 6The image reading apparatus includes a imaging mechanism that can change the magnification, and a reading mechanism unit that holds a solid-state imaging device that captures an optical image of a document image formed using the imaging optical unit. And the position of the reading mechanism section so as to capture the entire original image, and to divide the original image into a plurality of divided images and sequentially capture them at a necessary magnification, And a control means for obtaining data of a document image of an arbitrary resolution by fetching and connecting the image data. The control means comprises the following means.
Divided image storage means for storing the pixel value data of each divided image in association with the reading position on the document image.
When connecting two adjacent divided images, the reference divided image is the reference image, the other divided image is the connection image, and any one pixel of the connection image is the search reference pixel, and the reference image The connected image is relatively moved within a certain range, and the average value of the density difference between the corresponding pixels of both images in a certain area of the overlapping portion of both images, In determining the connection position, first, the search reference pixel is moved at intervals of several pixels within a rough search region having an area corresponding to the position accuracy of the reading mechanism with the position of the search reference pixel as the center, and the density difference The position of the search reference pixel when the average value of the reference value is the minimum is the provisional connection position between the reference image and the connection image, and then the range of the dense search region that is centered on the provisional connection position and smaller in area than the coarse search region In front The search reference pixel moves by one pixel, the connection position calculating means for determining a connection position of the and position the reference image in the search reference pixel and the connection image when the average value of the density difference is minimized.
Overlapping pixel calculation means that uses the pixel value of the overlapping area of both divided images at the connection position as a calculation value using the pixel value of the pixel of both divided images for each pixel.
[0018]
According to the image reading method and the image reading apparatus described above, the pixel value of the overlapping region is determined by performing some calculation using the pixel values of both divided images when connecting two adjacent divided images. Even when there is a slight error in the connected position, the discontinuity of the pixel values can be well dispersed and the appearance of the seam can be prevented.
[0019]
  Claim 7The image reading device ofClaim 6In the image reading apparatus described above, the overlapping pixel calculation unit sets each pixel value of the overlapping area of the two divided images as a calculation value based on a weighted average of the pixel values of the pixels of the two divided images. It is said.
[0020]
According to the above image reading apparatus, since each pixel value of the overlapping area is a calculated value by weighted averaging of the pixel values of the pixels of both divided images in the connection of two adjacent divided images, the overlapping area and its By continuously changing the pixel value at the boundary with the divided images at both ends, discontinuity of the pixel value can be favorably distributed and the manifestation of the seam can be prevented.
[0023]
  Claim 8The image reading apparatus includes a imaging mechanism that can change the magnification, and a reading mechanism unit that holds a solid-state imaging device that captures an optical image of a document image formed using the imaging optical unit. And the position of the reading mechanism section so as to capture the entire original image, and to divide the original image into a plurality of divided images and sequentially capture them at a necessary magnification, And a control means for obtaining data of a document image of an arbitrary resolution by taking in and connecting the control means, wherein the control means comprises the following means.
Divided image storage means for storing the pixel value data of each divided image in association with the reading position on the document image.
When connecting two adjacent divided images, the reference divided image is the reference image, the other divided image is the connection image, and any one pixel of the connection image is the search reference pixel, and the reference image The connected image is relatively moved within a certain range, and the average value of the density difference between the corresponding pixels of both images in a certain area of the overlapping portion of both images, In determining the connection position, first, the search reference pixel is moved at intervals of several pixels within a rough search region having an area corresponding to the position accuracy of the reading mechanism with the position of the search reference pixel as the center, and the density difference The position of the search reference pixel when the average value of the reference value is the minimum is the provisional connection position between the reference image and the connection image, and then the range of the dense search region that is centered on the provisional connection position and smaller in area than the coarse search region In front The search reference pixel moves by one pixel, the connection position calculating means for determining a connection position of the and position the reference image in the search reference pixel and the connection image when the average value of the density difference is minimized.
Overlapping pixel selection means for setting the pixel value of the overlapping area of both divided images at the connection position to the pixel value of the pixel of any one of the divided images for each pixel.
Overlapping pixel calculation means that uses the pixel value of the overlapping area of both divided images at the connection position as a calculation value using the pixel value of the pixel of both divided images for each pixel.
Combining means selecting means for selecting the overlapping pixel selecting means or the overlapping pixel calculating means according to a difference in pixel value between corresponding pixels in the overlapping area of both divided images at the connection position.
[0024]
According to the image reading apparatus, in connection between two adjacent divided images, the overlapping pixel selection unit is selected for each pixel according to the difference in pixel value between corresponding pixels representing discontinuity of the overlapping region. Replace with the pixel value of the corresponding pixel of one of the divided images, or select the overlapping pixel calculation means and set the calculated value using the pixel value of the corresponding pixel of both divided images for each pixel. It is possible to prevent the appearance of the seam between the images and to reduce the processing load and improve the efficiency of the connection process.
[0025]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of an image reading method and an image reading apparatus according to the present invention will be described below with reference to the drawings.
FIG. 1 is an external view of a configuration example of an image reading apparatus according to the present invention, and FIG. 2 is a control block diagram showing a main part of the image reading apparatus according to the present invention.
The image reading apparatus instructs the reading mechanism 100 that captures an image of the original 106, and a control unit 400 that captures and processes image data captured by the reading mechanism 100 while giving instructions such as position control and driving of the reading mechanism 100. And an input unit 500 for instructing resolution and reading start.
[0026]
The reading mechanism 100 further includes an imaging unit 100a composed of an imaging optical unit with variable magnification and a photoelectric conversion element such as a two-dimensional CCD, and drives the imaging unit 100a in a variable direction and controls the electrical operation of the photoelectric conversion element. Imaging control unit 100b that performs the above and a support base 100c that supports them.
The imaging unit 100a is controlled by the imaging control unit 100b in a circumferential direction centered on the XX line (arrow x), a circumferential direction centered on the axis Y extending in the front and back direction of the paper (arrow y), and a ZZ line. It can be freely rotated in each direction of the circumferential direction (arrow z) centered on.
The control unit 400 gives instructions for position control and driving of the image capturing unit 100a, captures image data captured by the image capturing unit 100a, performs various image processing such as connection processing of divided images, and more specifically. Corresponds to a personal computer 4 or the like.
Further, the input unit 500 specifically corresponds to a keyboard 5 or a mouse connected to a personal computer.
[0027]
The imaging control unit 100b and the control unit 400 are connected via a serial interface, for example, an RS232C cable 404, and an image is read from the control unit 400 to the imaging control unit 100b via this cable. Instructions such as the position and magnification of the imaging optical means are sent out, and the actual position information of the image pickup unit 100a and the like are sent from the image pickup control unit 100b to the control unit 400.
On the other hand, the image capturing unit 100a and the control unit 400 are connected via, for example, an NTSC video signal line 403, and image data captured by the image capturing unit 100a is sent out by the NTSC video signal line 403, and is captured by a video capture car. The window 402 is converted into a two-dimensional image and sent to the control unit 400.
[0028]
Next, main parts of the image reading apparatus will be described with reference to a control block diagram shown in FIG.
The control unit 400 performs two-dimensional processing of the NTSC video signal sent from the reading mechanism 100 and a processing control unit 410 that performs overall control such as control of the reading mechanism 100 and image data processing captured from the reading mechanism 100. An image signal converting unit (video capture card) 402 for converting to an image, a primary image memory 430 for storing a plurality of two-dimensional images in association with reading positions, and correcting distortion caused by imaging a document from an oblique direction. A perspective transformation / resolution adjustment unit 440 that performs perspective transformation and adjustment of a captured image to an arbitrary resolution, an image synthesis unit 450 that connects a plurality of images, and a connection image in which overlapping regions are synthesized in the image synthesis unit 450 And an output I / F 470 to be output.
[0029]
Image reading by the image reading apparatus is performed as follows.
First, after placing the document 106 on a desk or the like, when a resolution designation and reading start instruction are given from the input unit 500, the processing control unit 410 instructs the imaging control unit 100b to image the entire document 106, and The optical imaging system of the imaging unit 100a is automatically zoomed and focused so that the document 106 can be contained within one screen, and the entire document 106 is imaged, and the image signal converter 402 performs two-dimensional conversion from the NTSC video signal. Converted to an image.
The image of the entire document is output to a display or the like via a primary image memory 430, a secondary image memory 454, and an output I / F 470 as necessary.
Subsequently, the processing control unit 410 calculates the size of the original, determines the number of divisions of the image from the size of the original 106 and the designated resolution, and outputs each divided image of the original 106 to the imaging control unit 100b. Instructions for position control and driving of the photoelectric conversion elements are sent out so as to read sequentially, and each divided image is taken sequentially.
[0030]
Each divided image picked up by the image pickup unit 100a is converted from an NTSC video signal into a two-dimensional image by an image signal conversion unit 402, and is associated with the actual reading position of the image pickup unit 100a sent together with the image data to be a primary image. Stored in the memory 430.
Since each divided image stored in the primary image memory 430 is imaged obliquely by the imaging unit 100a, the perspective transformation in which the perspective transformation / resolution adjustment unit 440 converts the image including this distortion into an image captured from the front. And make adjustments so that the image has the specified resolution.
Then, the divided images stored in the primary image memory 430 are connected by a procedure described later to form one image corresponding to the original 106, and are output to the outside via the output I / F 470 as necessary.
[0031]
Next, the image composition unit 450 in the control unit 400 that performs connection processing of adjacent divided images will be described.
The image composition unit 450 takes in image data and position information of two adjacent divided images to be connected, calculates a connection position of both divided images, and overlaps of both divided images at the connection positions. The overlapping area memory 455 that stores each pixel value of the area, the connected image memory 451 that stores the pixel values of both divided images of the area other than the overlapping area, and the pixel value of the overlapping area for each pixel. An overlapping pixel selection determining unit 453 that determines and outputs whether to replace the pixel value of the pixel of the divided image, a secondary image memory 454 that stores the connected image, and a synthesis control unit 460 that controls these units. Composed.
[0032]
Further, the overlapping pixel selection determination unit 453 relates to the position in advance in order to determine, based on the position of the pixel in the overlapping region, which pixel value of the pixel in the overlapping region to replace each pixel value in the overlapping region. A determination criterion is defined, and position information in the overlapping area of the pixel is fetched from the overlapping area memory 455 when the pixel value of the overlapping area is determined, and the output result is changed according to the position.
[0033]
The image connection processing in the image composition unit 450 will be described along the flowchart shown in FIG.
First, the connection position calculation unit 452 fetches image data and position information of two adjacent divided images A and B to be connected from the primary image memory 430 (step 300), and a procedure to be described later based on these data. To calculate the connection position (step 301).
Next, an overlapping area of both divided images at the connection position is calculated (step 302), and each pixel value of both divided images in the overlapping area is stored in the overlapping memory 455, and each pixel of both divided images other than the overlapping area is stored. Each value is stored in the connected image memory 451.
[0034]
For the pixels that do not belong to the overlapping area, the original pixel value of the divided image before connection is output from the connected image memory 451 to the secondary image memory 454 (step 304).
On the other hand, for the pixels belonging to the overlapping region, the original pixel value of one of the divided images A or B for each pixel is determined for each pixel according to the determination criterion defined in advance in the overlapping pixel selection determination unit 453. The data is output from the overlapping area memory 455 to the secondary image memory 454 (steps 306 and 307).
This process is repeated for all the pixels (step 308), and each pixel value of the connected image obtained by connecting both divided images is stored in the secondary image memory 454 (step 309).
[0035]
Each processing in the image composition unit 450 is performed in the form of hardware as a part of a circuit of a dedicated image processing circuit, the form of software or firmware set in a general-purpose image processing circuit, or a general-purpose computer system. The present invention can be realized in various forms such as a form of software running on the above, a combination of these forms, and the like.
When implemented as software, processing is performed while reading a medium on which a program for performing the above processing is recorded in advance by a computer. For example, a configuration in which a floppy disk in which a processing program is recorded is set in the floppy disk drive 405 and processing is performed while being read as needed, a hard disk (not shown) built in the personal computer 4, etc. A mode in which the processing program is recorded on the computer is conceivable.
[0036]
Next, a connection position calculation method in the connection position calculation unit 452 will be described with reference to FIGS. The calculation of the connection position is efficiently performed by a two-stage search including a coarse search and a fine search.
Two divided

images

1 and 2 adjacent to the left and right have an overlapping portion 10 in a state of being arranged according to the actual reading position of the imaging unit 100a. At this time, the divided image 2 is read by moving the reading mechanism 100 after reading the divided image 1 in the right direction. Since there is a possibility of reading, the image is displayed as a vertically shifted image.
Hereinafter, the divided image 1 is set as a reference image, and the other divided image 2 is set as a connection image. A pixel located at the upper left end of the connection image 2 is set as a search reference pixel 2a.
[0037]
First, lattice points 21 arranged at intervals of several pixels are set in a rough search region 20 that is a rectangular region having a horizontal length W and a vertical length H centered on the search reference pixel 2a, and the lattice located at the upper left corner. The point is 21a, and the lattice point located at the lowermost right is 21c (FIG. 5). Here, lattice points 21 are arranged at intervals of three pixels, and the length between the lattice points is D. Next, the entire connected image 2 is moved in the upper left direction so that the search reference pixel 2a coincides with the lattice point 21a, and between the corresponding pixels of both images near the center of the overlapping portion 10 of the reference image 1 and the connected image 2 A density difference calculation area 11 for calculating an average value of pixel value differences (hereinafter referred to as density difference) is set (FIG. 6), and an average value of density differences in the density difference calculation area 11 is calculated.
[0038]
Thereafter, the density difference calculation set near the center of the overlapping portion 10 between the reference image 1 and the connection image 2 while sequentially moving the entire connection image 2 so that the search reference pixel 2a coincides with each lattice point up to the lattice point 21c. An average value of density differences in the region 11 ′ is calculated (FIG. 7). Then, the grid point where the average value of the density difference is minimized is detected and set as the temporary connection position 30 (FIG. 8).
[0039]
Next, the connection position between the reference image 1 and the connection image 2 is searched in more detail using the temporary connection position 30.
First, in the dense search area 40, which is a rectangular area having a horizontal length w and a vertical length h with the provisional connection position 30 as the center, a grid point 41 with an interval of one pixel is set, and the grid point located at the upper left is 41a, the lattice point located at the lower right is 41c (FIG. 8).
Then, as in the case of the coarse search region 20, the reference image 1 and the connection image 2 are sequentially moved while sequentially moving the entire connection image 2 so that the search reference pixel 2a coincides with each lattice point from the lattice points 41a to 41c. The average value of the density differences between the corresponding pixels in the density difference calculation region set near the center of the overlapping portion is calculated.
Then, the grid point at which the average value of the density differences is minimized is detected as the connection position 50 between the reference image 1 and the connection image 2 (FIG. 9).
[0040]
According to the connection position calculation method, since the connection position is searched based on the density difference of the overlapping area within a certain range based on the actual reading position at the time of divided image capturing, the position accuracy of the reading mechanism 100 is determined. Accordingly, the correct connection position can be determined even when the read image is misaligned.
In addition, the calculation of the connection position is efficiently performed by a two-stage search of a coarse search and a fine search, so that the processing time is shortened.
[0041]
In the following, an embodiment of a method for determining each pixel value of an overlapping area when two adjacent divided images are connected in the image reading method and the image reading apparatus according to the present invention will be described. In each embodiment, it is assumed that the connection position is calculated by the method described above.
[0042]
Example 1
As shown in FIG. 10, in the case where the divided image A and the divided image B are connected, the pixel located at the upper left corner of the overlapping area C is set as the origin, and the pixel whose horizontal position is i and whose vertical position is j The pixel value after connection is C [i, j]. Similarly, the pixel value of the divided image A at the position [i, j] is A [i, j], and the pixel value of the divided image B is B [i, j]. Then, the case where the pixel values of both divided images are alternately adopted as represented by the following Expression 1 using the dither arrangement in the overlapping pixel selection determination unit 453 (FIG. 2) will be described.
At this time, N pixels (i = 0 to N−1) are included in the left-right direction of the overlapping region C, and M pixels (j = 0 to M−1) are included in the vertical direction, and N: M = 1: Set to about 5.
[0043]
[Formula 1]

[0044]
According to the above formula, the pixel values of one of the divided images are alternately adopted as the pixel values of the overlapping area, and the pixel values of both divided images are arranged in a checkered pattern as shown in FIG. Will be. In the figure, for the sake of convenience, the pixel values of the divided image A and the pixel values of the divided image B are used so that the pixel values of the overlapping area alternately adopt the pixel values of the divided image A and the divided image B. Are displayed as uniform densities.
According to this embodiment, since the pixel values of both divided images are regularly and alternately adopted as the pixel value of the overlapping region, the boundary between the inside and the outside of the overlapping region is less noticeable at the end of the overlapping region, Even when there is a minute error in the calculated connection position, discontinuity of pixel values can be favorably distributed, and a connection image with no conspicuous joint can be obtained by connecting two adjacent images.
[0045]
(Example 2)
In the overlapping pixel selection determination unit 453 in FIG. 2, a pseudo-uniform random number generator that generates 0 or 1 at random (the output result is expressed as RAND [0,1]) is expressed by the following equation 2. A case where the pixel value to be adopted is determined based on a random number generated for each pixel will be described.
[0046]
[Formula 2]

[0047]
According to this embodiment, since the pixel values of both divided images are randomly adopted as the pixel values of the overlapping region, the end of the overlapping region is a straight line in which only the pixel values of one of the divided images are arranged. Because the boundary between the inside and outside of the overlap area is not conspicuous, even when there is a minute error in the calculated connection position, discontinuity of pixel values can be well distributed, and two adjacent images Can be connected to obtain an inconspicuous connection image.
[0048]
(Example 3)
A case will be described in which the overlapping pixel selection determination unit 453 in FIG. 2 determines which pixel value of a divided image is to be adopted with a diagonal line that bisects the overlapping region as shown in the following Expression 3.
[0049]
[Formula 3]

[0050]
According to the present embodiment, as shown in FIG. 12, since the joint between both divided images is an oblique line segment, high sensitivity to horizontal and vertical linear shapes well known as eye characteristics is obtained. Even if there is a minute error in the calculated connection position, it is possible to connect two adjacent images and obtain a connection image that is apparently inconspicuous.
In the case of the present embodiment, it can be easily realized by a circuit using a combination of programmable counters instead of an adder or software.
[0051]
Next, another embodiment of the image reading method and the image reading apparatus according to the present invention will be described. FIG. 13 is a control block diagram illustrating a main part of the image reading apparatus. In the figure, the same components as those in FIG. Further, the appearance of the image reading apparatus in this case is the same as that in FIG.
The image synthesis unit 450 ′ of the control unit 400 ′ includes an overlap pixel calculation unit 456 that performs some calculation using the pixel values of both divided images of the overlap region and outputs the calculation value as the pixel value of the pixel. A weight coefficient generation unit 457 for performing a weight calculation based on the position of the pixel in the overlap area using the corresponding pixel values of both divided images of the overlap area, and generating a weight coefficient in advance by the synthesis control section 460 ′ In addition, when determining the pixel value of the overlapping area, the position information of the pixel in the overlapping area is fetched from the overlapping image memory 455, the calculation is performed using the weighting coefficient according to the position, and the result is output to the secondary image memory 454. This is different from the image composition unit 450 in FIG. The weight coefficient generation unit 457 may be provided in the overlapping pixel calculation unit 456.
[0052]
Image connection processing in the image composition unit 450 'will be described with reference to the flowchart shown in FIG.
First, the connection position calculation unit 452 fetches image data and position information of two adjacent divided images A and B to be connected from the primary image memory 430 (step 310), and based on these, the above-described procedure is performed. To calculate the connection position (step 311).
Next, an overlapping area of both divided images at the connection position is calculated (step 312), and each pixel value of both divided images in the overlapping area is stored in the overlapping memory 455, and each pixel of both divided images other than the overlapping area is stored. Each value is stored in the connected image memory 451.
Subsequently, the weighting factor generator 457 generates a weighting factor (step 313).
[0053]
For the pixels that do not belong to the overlapping area, the original pixel value of the divided image before connection is output from the connected image memory 451 to the secondary image memory 454 (step 315).
On the other hand, for the pixels belonging to the overlapping area, the overlapping pixel calculation unit 456 takes in the pixel values of the pixels of both divided images from the overlapping area memory 455, performs the calculation using the weight coefficient generated in advance, and calculates the calculated value. The data is output to the secondary image memory 454 (step 316).
This process is repeated for all the pixels (step 317), and each pixel value of the connected image obtained by connecting both divided images is stored in the secondary image memory 454 (step 318).
Each processing in the image synthesizing unit 450 ′ can be realized by various forms such as hardware, software, and firmware as in the case of the image synthesizing unit 450 (FIG. 2).
[0054]
In the following, an embodiment of a method for determining each pixel value of an overlapping area when two adjacent divided images are connected in the image reading method and the image reading apparatus according to the present invention will be described. In each embodiment, it is assumed that the connection position is calculated by the method described above.
[0055]
Example 4
A case will be described in which the overlapped pixel calculation unit 456 of FIG. 13 performs weighted calculation inclined in the left-right direction on the pixel values of both divided images as represented by the following equations 4-6.
Here, the inclined weight will be described with reference to FIG. 10. In the calculation of the pixel value of the overlapping region, the distribution using the original pixel values of the divided image A and the divided image B in the horizontal direction of the drawing is represented by the left end portion. A: B = 1: 0, A: B = 0.5: 0.5 at the center portion, and A: B = 0: 1 at the right end portion. , F1 (i, j) and f2 (i, j) are weighting coefficients representing this ratio.
[0056]
[Formula 4]
C [i, j] = A [i, j] * f1 (i, j) + B [i, j] * f2 (i, j)
[0057]
[Formula 5]
f2 (i, j) = i / (N-1)
[0058]
[Formula 6]
f1 (i, j) = 1-f2 (i, j)
[0059]
According to this embodiment, the end of the overlapping area on the divided image A side adopts the pixel value of the divided image A as it is, while the end of the divided image B side adopts the pixel value of the divided image B as it is. Since each pixel value is an average value of the pixel values of both divided images at the center in the horizontal direction of the region, the pixel value of the divided image A is used more heavily as it is located to the left in the horizontal direction of the overlapping region. As the pixel values of the divided image B are heavier, the pixel values at the boundary between the overlapping region and the divided images at both ends thereof are continuously changed even when there is a minute error at the calculated connection position. The discontinuity of pixel values can be dispersed well, and a connection image with no conspicuous joint can be obtained.
In the present embodiment, it is assumed that N: M = 1: 5 and the overlapping region has a vertically long shape, and the connecting portion is considered to be particularly conspicuous in the horizontal direction. Although the calculation is performed, the weight calculation may be further performed in the vertical direction.
[0060]
(Example 5)
When the overlapped pixel calculation unit 456 in FIG. 15 performs weighted calculation on the pixel values of both divided images and uses a dither array as the weighting coefficient as shown in the above expressions 4 and 6 and the following expressions 7 and 8. Will be described.
[0061]
[Formula 7]

[0062]
[Formula 8]
f2 (i, j) = D [(i mod n) +1, (j mod m) +1]: n = 4, m = 4
[0063]
According to the present embodiment, the connection position calculated by the adjacent pixels having the pixel values determined by using the pixel values of the divided image A or the divided image B in different distributions in the adjacent 16 pixels of the overlapping region. Even if there is a minute error, discontinuity of pixel values can be well dispersed, and a connection image in which the adjacent two images are connected to each other and the joint is not conspicuous can be obtained.
[0064]
Next, another embodiment of the image reading apparatus according to the present invention will be described. FIG. 15 is a control block diagram illustrating a main part of the image reading apparatus. In the figure, the same components as those in FIG. 2 and FIG. Further, the appearance of the image reading apparatus in this case is the same as that in FIG.
The image composition unit 450 ″ of the control unit 400 ″ includes an overlapping pixel selection determining unit 453 that determines and outputs which pixel value of the divided image the pixel value of the overlapping region is replaced for each pixel, and the overlapping region. And an overlapping pixel calculation unit 456 that performs some calculation using the pixel values of both divided images and outputs the calculation value as the pixel value of the pixel, and further includes an overlapping pixel selection determination unit 453 or an overlapping pixel calculation. A combining unit selecting unit 458 that selects one of the units 456, and the combining control unit 460 ″ determines the connection position of the two divided images in the connection position calculation unit 452 when determining the pixel value of the overlapping region. Based on the calculated average value of the density differences of the overlapping areas, the synthesizing means selecting unit 458 selects one of the synthesizing means, and each pixel value of the overlapping area synthesized by each synthesizing means is determined. In that is output to the next image memory 454, image synthesizing unit 450 of FIG. 2, or differs from the image synthesizing unit 450 of FIG. 13 '.
Each processing in the image synthesizing unit 450 ″ can be realized by various forms such as hardware, software, and firmware as in the case of the image synthesizing unit 450 (FIG. 2).
[0065]
In the following, an embodiment of a method for determining each pixel value of an overlapping area when two adjacent divided images are connected in the image reading apparatus according to the present invention will be described.
[0066]
(Example 6)
In FIG. 15, in addition to the overlapping pixel selection determination unit 453 and the overlapping pixel calculation unit 456, the overlapping pixel that totally replaces the pixel value of the overlapping region of two connected divided images with the corresponding pixel value of one of the divided images. The image connection processing in the image composition unit 450 ″ when a replacement unit 459 (not shown) is provided and these three units can be selected by the composition unit selection unit 458 is shown in FIG. This will be explained along the line.
[0067]
First, the connection position calculation unit 452 fetches image data and position information of two adjacent divided images A and B to be connected from the primary image memory 430 (step 320), and based on these, the above-described procedure is performed. To calculate the connection position (step 321).
Next, an overlapping area of both divided images at the connection position is calculated (step 322), and each pixel value of both divided images in the overlapping area is stored in the overlapping memory 455, and each pixel of both divided images other than the overlapping area is stored. Each value is stored in the connected image memory 451.
Subsequently, any one of the overlapping pixel selection determination unit 453, the overlapping pixel calculation unit 456, or the overlapping pixel replacement unit 459 is selected according to the average value of the density differences calculated when the connection position is determined in the combining unit selection unit 458. Select one. For example, when the dynamic range of the pixel value is D, if the average value of the density difference is D / 4 or less (step 323, Yes), the overlap pixel replacement unit 459 is used, and the average value of the density difference is D / 4 or more and D If it is less than / 2 (step 324, Yes), the overlap pixel selection determination unit 453 is selected in advance, and if the average value of density differences is D / 2 or more (step 324, No), the overlap pixel calculation unit 456 is selected in advance. Stipulate.
[0068]
When the overlapping pixel replacement unit 459 is selected, for example, an A image is selected as to which pixel value of the overlapping image the pixel value of the overlapping area is to be replaced as a whole (step 325).
For pixels that do not belong to the overlapping area, the original pixel value of the divided image before connection is output from the connected image memory 451 to the secondary image memory 454 (step 327).
On the other hand, for the pixels belonging to the overlapping area, the overlapping pixel replacement unit 459 outputs the pixel value of the pixel of the A image selected in advance to the secondary image memory 454 (step 328).
This process is repeated for all pixels (step 329), and each pixel value of the connected image obtained by connecting both divided images is stored in the secondary image memory 454 (step 330).
[0069]
When the overlapping pixel selection determination unit 453 is selected, the process jumps to 1 in FIG. 3 and performs the following processing. When the overlapping pixel calculation unit 456 is selected in the same manner, the process jumps to 2 in FIG. The following processing is performed. As a specific pixel value determination method of the overlapping pixel selection determination unit 453, any one of the above-described first to third embodiments is used, and for the overlapping pixel calculation unit 456, any one of the above-described fourth or fifth embodiment is used. Can be used.
Then, the respective pixel values of the overlapping region are synthesized according to the average value of the density difference, and the secondary image memory 454 stores each pixel value of the connection image obtained by connecting two adjacent divided images (FIG. 3). Step 309, FIG. 14, step 318).
[0070]
According to the present embodiment, when it is considered that the average value of the density difference of the overlapping region is small and the joint of the divided images is relatively inconspicuous, each pixel value of the overlapping region is set to the corresponding pixel value of one of the divided images. By simply replacing it, the burden of connection processing can be reduced.
On the other hand, when the average value of the density difference is large and it seems that the seam of the divided images appears prominently, averaging by dither array weighting etc. is adopted, and discontinuity of pixel values due to connection position error is well dispersed Can be made.
Further, as described above, it is possible to perform economical and effective connection processing according to the connection image by automatically selecting the overlapping region combining means according to the average value of the density differences.
[0071]
Note that the combination of the threshold value for selecting each combining means in the combining means selection unit 458 and the specific processing content in the selected combining means is not limited to the above description. Further, it is conceivable to select each combining means according to the designated resolution, the type of document, or the like, or to automatically select each combining means and switch the processing contents.
[0072]
As another embodiment of the image reading apparatus of the present invention, for example, an optical imaging system in which a photoelectric conversion element is configured by a scan type of a one-dimensional CCD element, a mirror is incorporated in the optical imaging system, and the optical axis direction is variable. A device with a control mechanism, a device including a dedicated controller that performs connection processing of divided images, zooming and focusing of the optical imaging system, and reading position control of each divided image, intermediate for converting from a photoelectric conversion element to a digital image There may be variations such as those using signals other than NTSC signals and those using digital image output without using analog signals, and combinations thereof are also possible.
[0073]
【The invention's effect】
According to the present invention, when adjacent divided images are connected, the pixel value of the overlapping region is not replaced entirely with the pixel value of one of the divided images, but instead of one of the divided images for each pixel. Even if there is a minute error in the connection position of both divided images, the overlapping area is synthesized by replacing it with the pixel value or by replacing it with the calculated value using the pixel value of the pixel of both divided images. In addition, it is possible to obtain a connection image in which the discontinuity of pixel values is well dispersed and the joint is not conspicuous. Therefore, by connecting a plurality of divided images in the same manner, it is possible to obtain a high-quality read image that is faithful to the image original with an arbitrary resolution.
In addition, by automatically selecting the synthesis method, it is possible to effectively disperse the discontinuity of pixel values and to perform a connection process that can withstand practical use while reducing the processing load.
[Brief description of the drawings]
FIG. 1 is an external explanatory diagram of an image reading apparatus according to the present invention.
FIG. 2 is a control block diagram showing a main part of the image reading apparatus according to the present invention.
FIG. 3 is a flowchart showing a part of the image reading method according to the present invention.
FIG. 4 is an explanatory diagram showing a procedure of a connection position calculation method in the image reading method according to the present invention.
FIG. 5 is an explanatory diagram showing a procedure of a connection position calculation method in the image reading method according to the present invention.
FIG. 6 is an explanatory diagram showing a procedure of a connection position calculation method in the image reading method according to the present invention.
FIG. 7 is an explanatory diagram showing a procedure of a connection position calculation method in the image reading method according to the present invention.
FIG. 8 is an explanatory diagram showing a procedure of a connection position calculation method in the image reading method according to the present invention.
FIG. 9 is an explanatory diagram showing a procedure of a connection position calculation method in the image reading method according to the present invention.
FIG. 10 is an explanatory diagram showing the position of each pixel in the overlapping region in the image reading method according to the present invention.
FIG. 11 is an explanatory diagram illustrating a connection result of adjacent divided images according to the first embodiment.
FIG. 12 is an explanatory diagram illustrating a connection result of adjacent divided images according to the third embodiment.
FIG. 13 is a control block diagram showing a main part of the image reading apparatus according to the present invention.
FIG. 14 is a flowchart showing a part of the image reading method according to the present invention.
FIG. 15 is a control block diagram showing a main part of the image reading apparatus according to the present invention.
FIG. 16 is a flowchart showing a part of the image reading method according to the present invention.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 100 ... Reading mechanism, 100a ... Imaging part, 100b ... Imaging control part, 100c ... Support stand, 106 ... Document, 400 ... Control part, 402 ... Image signal conversion part (video capture card), 403 ... RS232C case 404 ... NTSC video signal line 410 ... Processing control unit 430 ... Primary image memory 440 ... Perspective conversion / resolution adjustment unit 450, 450 ', 450 "... Image composition unit 451 ... Connected image memory 452 ... connection position calculation unit, 453 ... overlapping pixel selection determination unit, 454 ... secondary image memory, 455 ... overlapping region memory, 456 ... overlapping pixel calculation unit, 457 ... weighting factor generation unit, 458 ... synthesis means selection unit, 460, 460 ', 460 "... Synthesis control unit, 470 ... Output I / F, 500 ... Input unit

Claims

An image reading method for dividing an original image into a plurality of divided images, sequentially reading each divided image with an adjacent divided image, and obtaining the data of the original image by connecting the divided images.
When connecting two adjacent divided images, the reference divided image is the reference image, the other divided image is the connection image, and any one pixel of the connection image is the search reference pixel, and the reference image The connected image is relatively moved within a certain range, and the average value of the density difference between the corresponding pixels of both images in a certain area of the overlapping portion of both images, In determining the connection position, first, the search reference pixel is moved at intervals of several pixels within a rough search region having an area corresponding to the position accuracy of the reading mechanism with the position of the search reference pixel as the center, and the density difference The position of the search reference pixel when the average value of is the minimum is the provisional connection position between the reference image and the connection image,
Next, the search reference pixel is moved one pixel at a time within the range of the dense search region having an area smaller than the coarse search region with the provisional connection position as the center, and the average value of the density difference is minimized. The position of the search reference pixel is the connection position between the reference image and the connection image,
The image reading is characterized in that the two divided images are connected by setting the pixel value of the overlapping area of the two divided images at the connection position as the pixel value of one of the divided images for each pixel. Method.

An imaging optical unit capable of changing the magnification, a reading mechanism unit that holds a solid-state imaging device that captures an optical image of a document image formed using the imaging optical unit, and the reading mechanism unit The position is controlled so that the entire original image is imaged, and the original image is divided into a plurality of divided images and sequentially captured at a necessary magnification. An image reading apparatus having control means for obtaining data of an original image of an arbitrary resolution,
The control means includes
Divided image storage means for storing the pixel value data of each divided image in association with the reading position on the document image;
When connecting two adjacent divided images, the reference divided image is the reference image, the other divided image is the connection image, and any one pixel of the connection image is the search reference pixel, and the reference image The connected image is relatively moved within a certain range, and the average value of the density difference between the corresponding pixels of both images in a certain area of the overlapping portion of both images, In determining the connection position, first, the search reference pixel is moved at intervals of several pixels within a rough search region having an area corresponding to the position accuracy of the reading mechanism with the position of the search reference pixel as the center, and the density difference The position of the search reference pixel when the average value of the reference value is the minimum is the provisional connection position between the reference image and the connection image, and then the range of the dense search region that is centered on the provisional connection position and smaller in area than the coarse search region In front The search reference pixel moves by one pixel, and a connection position calculating means the average value of the density difference is to determine the connection position between the and position the reference image in the search reference pixel connected image when a minimum,
Overlapping pixel selection means for setting the pixel value of the overlapping area of both divided images at the connection position to the pixel value of the pixel of either one of the divided images for each pixel;
An image reading apparatus comprising:

3. The image reading apparatus according to claim 2, wherein the overlapping pixel selection unit determines each pixel value of the overlapping area of the two divided images based on the position of the pixel in the overlapping area. An image reading apparatus characterized by determining whether or not the pixel value is to be determined.

The image reading apparatus according to claim 3, wherein the overlapping pixel selection unit is configured to diagonally divide each pixel value of the overlapping area of the two divided images into a position where the pixel in the overlapping area bisects the overlapping area. An image reading apparatus that determines which of the divided images the pixel value of the pixel is based on which of the two areas divided by the above.

An image reading method for dividing an original image into a plurality of divided images, sequentially reading each divided image with an adjacent divided image, and obtaining the data of the original image by connecting the divided images.
When connecting two adjacent divided images, the reference divided image is the reference image, the other divided image is the connection image, and any one pixel of the connection image is the search reference pixel, and the reference image The connected image is relatively moved within a certain range, and the average value of the density difference between the corresponding pixels of both images in a certain area of the overlapping portion of both images, In determining the connection position, first, the search reference pixel is moved at intervals of several pixels within a rough search region having an area corresponding to the position accuracy of the reading mechanism with the position of the search reference pixel as the center, and the density difference The position of the search reference pixel when the average value of is the minimum is the provisional connection position between the reference image and the connection image,
Next, the search reference pixel is moved one pixel at a time within the range of the dense search region having an area smaller than the coarse search region with the provisional connection position as the center, and the average value of the density difference is minimized. The position of the search reference pixel is the connection position between the reference image and the connection image,
The two divided images are connected by setting the pixel value of the overlapping region of the two divided images at the connection position to a calculated value using the pixel value of the pixel of the two divided images for each pixel. Image reading method.

An imaging optical unit capable of changing the magnification, a reading mechanism unit that holds a solid-state imaging device that captures an optical image of a document image formed using the imaging optical unit, and the reading mechanism unit The position is controlled so that the entire original image is imaged, and the original image is divided into a plurality of divided images and sequentially captured at a necessary magnification. An image reading apparatus having control means for obtaining data of an original image of an arbitrary resolution,
The control means includes
Divided image storage means for storing the pixel value data of each divided image in association with the reading position on the document image;
When connecting two adjacent divided images, the reference divided image is the reference image, the other divided image is the connection image, and any one pixel of the connection image is the search reference pixel, and the reference image The connected image is relatively moved within a certain range, and the average value of the density difference between the corresponding pixels of both images in a certain area of the overlapping portion of both images, In determining the connection position, first, the search reference pixel is moved at intervals of several pixels within a rough search region having an area corresponding to the position accuracy of the reading mechanism with the position of the search reference pixel as the center, and the density difference The position of the search reference pixel when the average value of the reference value is the minimum is the provisional connection position between the reference image and the connection image, and then the range of the dense search region that is centered on the provisional connection position and smaller in area than the coarse search region In front The search reference pixel moves by one pixel, and a connection position calculating means the average value of the density difference is to determine the connection position between the and position the reference image in the search reference pixel connected image when a minimum,
Overlapping pixel calculation means for setting the pixel value of the overlapping area of both divided images at the connection position to a calculation value using the pixel value of the pixel of both divided images for each pixel;
An image reading apparatus comprising:

The image reading apparatus according to claim 6 , wherein the overlapping pixel calculation unit sets each pixel value of the overlapping area of the two divided images as a calculation value based on a weighted average of the pixel values of the pixels of the two divided images. An image reading apparatus.

An imaging optical unit capable of changing the magnification, a reading mechanism unit that holds a solid-state imaging device that captures an optical image of a document image formed using the imaging optical unit, and the reading mechanism unit The position is controlled so that the entire original image is imaged, and the original image is divided into a plurality of divided images and sequentially captured at a necessary magnification. An image reading apparatus having control means for obtaining data of an original image of an arbitrary resolution,
The control means includes divided image storage means for storing pixel value data of each of the divided images in association with a reading position on the document image, and
When connecting two adjacent divided images, the reference divided image is the reference image, the other divided image is the connection image, and any one pixel of the connection image is the search reference pixel, and the reference image The connected image is relatively moved within a certain range, and the average value of the density difference between the corresponding pixels of both images in a certain area of the overlapping portion of both images, In determining the connection position, first, the search reference pixel is moved at intervals of several pixels within a rough search region having an area corresponding to the position accuracy of the reading mechanism with the position of the search reference pixel as the center, and the density difference The position of the search reference pixel when the average value of the reference value is the minimum is the provisional connection position between the reference image and the connection image, and then the range of the dense search region that is centered on the provisional connection position and smaller in area than the coarse search region In front The search reference pixel moves by one pixel, and a connection position calculating means the average value of the density difference is to determine the connection position between the and position the reference image in the search reference pixel connected image when a minimum,
Overlapping pixel selection means for setting the pixel value of the overlapping area of both divided images at the connection position to the pixel value of the pixel of either one of the divided images for each pixel;
Overlapping pixel calculation means for setting the pixel value of the overlapping area of both divided images at the connection position to a calculation value using the pixel value of the pixel of both divided images for each pixel;
Combining means selecting means for selecting the overlapping pixel selecting means or the overlapping pixel calculating means according to a difference in pixel value between corresponding pixels in the overlapping area of both divided images at the connection position;
An image reading apparatus comprising: