JP2004046329A

JP2004046329A - Image contour enhancement device

Info

Publication number: JP2004046329A
Application number: JP2002199803A
Authority: JP
Inventors: Tsutomu Takizawa; 滝沢　努
Original assignee: Pentax Corp
Current assignee: Pentax Corp
Priority date: 2002-07-09
Filing date: 2002-07-09
Publication date: 2004-02-12
Anticipated expiration: 2022-07-09
Also published as: JP4190221B2

Abstract

<P>PROBLEM TO BE SOLVED: To prevent the occurrence of a pseudo signal in an endoscope and to provide a clear display picture. <P>SOLUTION: A luminance difference with adjacent pixels in respective pixels of image data is quantized within a prescribed range. A reference luminance difference is set to an almost intermediate value being the luminance difference which needs contour enhancement the most. The luminance difference whose difference with a reference luminance difference is the minimum is selected among the luminance differences with the adjacent pixels. In a region where the selected luminance difference becomes the reference luminance difference, a contour is enhanced the most. In a region where the difference between the selected luminance difference and the reference luminance difference becomes relatively large, a contour enhancement amount is relatively reduced. <P>COPYRIGHT: (C)2004,JPO

Description

【０００１】
【発明の属する技術分野】
本発明は、例えば内視鏡によって得られる画像に対して輪郭強調を施す輪郭強調装置に関するものである。
【０００２】
【従来の技術】
従来、画像処理において、輪郭強調は画面領域全体に対して均等に施されている。すなわち、隣接画素との輝度差が小さい領域では、輪郭強調量が小さく掛けられ、得られる画像データの輝度値も元データに比べあまり増大させられない。一方、隣接画素との輝度差が大きい領域では、輪郭強調量が大きく掛けられ、得られる画像データの輝度値も増大させられる。これにより、領域ごとの輝度差が明確にされ、シャープな画像が得られる。
【０００３】
しかし、輝度差が非常に大きくなると、輪郭強調量もそれに併せて大きく掛けられるので、画像データの輝度値が大きくなりすぎて、その領域にはハレーションなどの擬似信号が発生することがある。
【０００４】
特に内視鏡によって得られる画像データは、その光学特性から暗い領域と明るい領域が混在しており、輝度差が大きい領域が多い。また、内視鏡の被観察体は体液等の液体が付着したものが多く、これらに光が照射された場合、光が過度に反射され、その近傍部分との輝度差が非常に大きくなることがある。したがって、内視鏡によって得られる画像は擬似信号が多く発生するため、不自然な画像になることが多い。また、輝度差が少ない暗い領域においては、ノイズが多く、この領域に輪郭強調が均等に施されると、ノイズが強調され、画質が劣化することがある。
【０００５】
【発明が解決しようとする課題】
本発明は、輪郭強調による擬似信号の発生を防止して、鮮明で観察しやすい画像を得ることを目的としている。
【０００６】
【課題を解決するための手段】
本発明に係る画像輪郭強調装置は、画像データを構成する各画素に対して隣接画素との輝度差に応じた輪郭強調を行う輪郭強調手段と、輝度差が予め設定されている基準輝度差となる領域において、輪郭強調における輪郭強調量を最大にする輪郭強調調整手段とを備える。輪郭強調調整手段は輝度差と基準輝度差との差が相対的に大きくなる領域において輪郭強調量を相対的に小さくすることが好ましい。これにより、輝度差が大きくなりすぎる領域において、輪郭強調量が抑制され、擬似信号の発生が防止される。
【０００７】
画像データの輝度値は一定の範囲内で量子化され、基準輝度差は一定の範囲内の略中間値であることが好ましい。好ましくは、輪郭強調手段は、隣接画素との輝度差のうち基準輝度差との差が最も小さい輝度差に応じて輪郭強調を行う。画像データを得るための撮像装置は内視鏡であることが好ましい。
【０００８】
輪郭強調調整手段は、輝度差が基準輝度差よりも小さいと検出される領域においてのみ輪郭強調を行うことが好ましい。また、輪郭強調調整手段は、輝度差が基準輝度差よりも大きいと検出される領域においてのみ輪郭強調を行ってもよい。これにより、必要に応じて、輝度差が大きい領域若しくは、小さい領域にて、輪郭強調を掛けないことが可能である。
【０００９】
本発明に係る画像輪郭強調装置画像は、データに対して輪郭強調を行う輪郭強調手段と、隣接画素との輝度差が予め設定されている基準輝度差よりも小さいと検出される領域においてのみ輪郭強調を行う輪郭強調調整手段を備える。また、本発明に係る画像輪郭強調装置は、画像データに対して輪郭強調を行う輪郭強調手段と、隣接画素との輝度差が予め設定されている基準輝度差よりも大きいと検出される領域においてのみ輪郭強調を行う輪郭強調調整手段とを備える。
【００１０】
また、本発明に係る画像輪郭強調方法は画像データを構成する各画素に対して隣接画素との輝度差に応じた輪郭強調を行う輪郭強調ステップと、輝度差が予め設定されている基準輝度差となる領域において、輪郭強調における輪郭強調量を最大にする輪郭強調調整ステップとを備える。
【００１１】
【発明の実施の形態】
以下、本発明の実施形態について添付図面を参照して説明する。
図１は本発明の第１の実施形態である画像輪郭強調装置を備えた内視鏡システムを示し、内視鏡システムは、内視鏡１０および画像処理装置から成る。内視鏡１０は被写体から画像データを得るための撮像装置であり、画像処理装置は内視鏡１０によって得られた画像データを画像として出力できるように処理するための装置である。
【００１２】
ＣＰＵ１２では、制御情報が生成され、生成された制御情報はマスク生成回路１３、輪郭強調回路１４に送られる。マスク生成回路１３と輪郭強調回路１４は生成された制御情報に基づき制御される。
【００１３】
内視鏡１０によって得られた画像データは映像信号に変換され、映像信号は、加算器１１に送られる。加算器１１では、送られてきた映像信号に、マスク生成回路１３で生成されたマスク情報が加算される。マスク情報が加算された映像信号は輪郭強調回路１４に送られ、輪郭強調が掛けられる。
【００１４】
輪郭強調が掛けられた映像信号は、映像信号処理回路１５に送られ、映像信号処理回路１５でＡ／Ｄ変換されるとともに、ＲＧＢに分離され、それぞれＲメモリ１６ａ、Ｇメモリ１６ｂ、Ｂメモリ１６ｃに格納される。格納されたそれぞれの映像信号はビデオ画像処理回路１７に入力され、Ｄ／Ａ変換された後、出力モニタに画像として出力される。同期信号生成部１８では同期信号が生成される。同期信号は、タイミング回路１９に送られ、映像処理回路１５、ビデオ処理回路１７の動作タイミングの制御と、メモリ１６ａ、１６ｂ、１６ｃに対する書込みおよび読み出し制御に用いられる。
【００１５】
図２は輪郭強調回路１４を示す。輪郭強調回路１４では、画像データを構成する各画素に対して、隣接画素との輝度差に応じた輪郭強調が行なわれる。なお、ここでは、輝度差は８ビット（２５６段階）で量子化され演算が行なわれているが、目的に応じて、分解能を変更してもよい。
【００１６】
入力信号ａは、遅延器２０に送られると共に減算器２１および加算器２２に送られる。遅延器２０では、入力信号ａが１画素遅延させられ、１画素遅延信号ｂが得られる。１画素遅延信号ｂは、遅延器２３に送られると共に、減算器２１、２４、２６および加算器２７に送られる。遅延器２３では、１画素遅延信号ｂが１画素遅延させられ２画素遅延信号ｃが得られ、２画素遅延信号ｃは、加算器２２および減算器２４に送られる。
【００１７】
加算器２２では、入力信号ａと２画素遅延信号ｃが加算された後、係数（１／２）が乗じられ、加算信号ｄ（＝（ａ＋ｃ）／２）が得られる。減算器２６では、１画素遅延信号ｂから加算器２２より送られてきた加算信号ｄが減ぜられ、輪郭強調信号ｅ（＝ｂ−（ａ＋ｃ）／２）が得られる。得られた輪郭強調信号ｅは、乗算器２８に送られる。
【００１８】
減算器２１では、１画素遅延信号ｂから入力信号ａが減ぜられ、その絶対値（｜ｂ−ａ｜）が求められる。減算器２４でも同様に１画素遅延信号ｂから２画素遅延信号ｃが減ぜられ、その絶対値（｜ｂ−ｃ｜）が求められる。減算器２１、２４で求められた絶対値は隣接画素との輝度差としてフィルタ２５に送られる。フィルタ２５では、各画素における隣接画素との輝度差（｜ｂ−ａ｜若しくは｜ｂ−ｃ｜）のうち基準輝度差との差が小さい輝度差が選択される。なお、｜ｂ−ａ｜と｜ｂ−ｃ｜の値が等しくなった場合は、輝度差の値の大きい方が選択される。その選択された値は，フィルタ２９に送られ、フィルタ２９で基準輝度差より大きいか否かが読み取られ、基準輝度差以下であるならば、変換器３０へ、基準輝度差より大きければ変換器３１へと送られる。なお、基準輝度差は輪郭強調が最も必要とされる輝度差に設定される。本実施形態では、基準輝度差は、輝度差が量子化された範囲内の中間値である１２７と予め設定されている。
【００１９】
変換器３０、３１では、フィルタ２５で選択された輝度差が基準輝度差となる領域において輪郭強調量が最大になり、さらに選択された輝度差と基準輝度差との差が相対的に大きくなる領域において輪郭強調量が相対的に小さくなるように輪郭強調係数が設定される。すなわち、変換器３０では、選択された輝度差が大きくなる領域の輪郭強調係数は大きくなるように、変換器３１では、選択された輝度差の値が大きくなる領域の輪郭強調係数は小さくなるように設定されている。また、本実施形態では、選択された輝度差と輪郭強調係数の関係が比例関係となるように設定され、かつ輪郭強調係数の最小値は０で、最大値は１となるように設定されている。
【００２０】
例えば、変換器３０では、選択された輝度差が０となる画素の輪郭強調係数は０に、輝度差が１となる画素の輪郭強調係数は１／１２７に、輝度差が１２６となる画素の輪郭強調係数は１２６／１２７に、輝度差が１２７となる画素の輪郭強調係数は１となるように設定される。一方、変換器３１では、輝度差が１２８となる画素の輪郭強調係数は１に、輝度差が１２９となる画素の輪郭強調係数は１２６／１２７に、輝度差が２５４となる画素の輪郭強調係数は１／１２７に、輝度差が２５５となる画素の輪郭強調係数は０となるように設定される。
【００２１】
変換器３０，３１で求められた輪郭強調係数は、乗算器２８に送られ、乗算器２８では、輪郭強調信号ｅに輪郭強調係数が乗じられ、強調係数乗算信号ｆ（＝ｅ×輪郭強調係数）が求められる。強調係数乗算信号ｆは加算器２７に送られ、加算器２７では、１画素遅延信号ｂに強調係数乗算信号ｆが加算され、出力信号ｇが得られる。
【００２２】
図３から図５に輪郭強調回路１４で得られる信号例を示す。
図３は入力信号ａの各画素の輝度値が０、２５５、２５５、２５５、２５５、０、０であるとき、輪郭強調回路１４で得られる信号例を示す。例えば、各画素の輝度差（｜ｂ−ａ｜、｜ｂ−ｃ｜）は、左から１画素目の輝度差が０、０となり、２画素目が２５５、０となる。この場合において、フィルタ２５（図２参照）では、例えば左から１画素目の２つの輝度差がそれぞれ０であるので、この画素における輝度差は０と選択される。２画素目の輝度差は２５５、０であるので、１２７との差が小さい０が選択される。以下の全ての画素についても同様に輝度差が選択され、本実施例においては、全ての画素における輝度差が０として選択される。
【００２３】
選択された輝度差に応じて変換器３０、３１（図２参照）において各画素の輪郭強調係数が求められる。図３に示す信号では選択された輝度差がすべて０であるので、すべての画素における輪郭強調係数は０となる。
【００２４】
加算器２２（図２参照）では、加算信号ｄが求められ、減算器２６で輪郭強調信号ｅが求められる。輪郭強調信号ｅは乗算器２８に送られ、輪郭強調係数が乗ぜられる。図３の信号例の全ての画素において、輪郭強調係数は０となる。したがって、強調係数乗算信号ｆは、すべての画素の輝度値が０となり、出力信号ｇは１画素遅延信号ｂと同一の信号となる。このようにして、輝度差が最大（２５５）の画素には、輪郭強調が全く掛けられず、擬似信号の発生が防止される。
【００２５】
図４は入力信号の各画素の輝度値が０、１２７、１２７、１２７、１２７、０、０のときの輪郭強調回路１４で得られる信号例を示す。例えば、各画素の輝度差｜ｂ−ａ｜、｜ｂ−ｃ｜は、左から１画素目の輝度差が０、０となり、２画素目が１２７，０となる。フィルタ２５（図２参照）では、例えば、２画素目の輝度差は１２７、０であるので、１２７との差が小さい１２７が選択される。以下の全ての画素についても同様に輝度差が選択され、本実施例においては、左から０、１２７、１２７、０、０、１２７、１２７、０が各画素の輝度差として選択される。
【００２６】
選択された輝度差に応じて変換器３０、３１（図２参照）において各画素の輪郭強調係数が求められる。図４に示す信号では選択された輝度差が０の画素における輪郭強調係数は０となる。一方、選択された輝度差が１２７の画素における輪郭強調係数は１となり、左から輪郭強調係数は０、１、１、０、０、１、１、０となる。減算器２６で得られた輪郭強調信号ｅのそれぞれの画素に、これらの輪郭強調係数が乗算器２８で乗ぜられ、強調係数乗算信号ｆが得られる。すなわち、選択された輝度差が基準輝度差と一致する画素では、輪郭強調信号ｅに輪郭強調係数の最大値である１が乗ぜられ強調係数乗算信号ｆが得られる。輪郭強調係数信号ｆには、一画素遅延信号ｂが加えられ出力信号ｇが得られる。このようにして、輝度差が基準輝度差となる画素においては、輪郭強調が最大に掛けられた出力信号ｇが得られる。
【００２７】
図５は入力信号の各画素の輝度値が０、６３、２５５、２５５、２５５、０、０のときの輪郭強調回路１４で得られる信号例を示す。各画素の輝度差｜ｂ−ａ｜、｜ｂ−ｃ｜は、例えば、左から１画素目の輝度差が０、０となり、２画素目が６３、０となり、３画素目が１９２、６３、４画素目が０、１９２となる。この場合において、各画素の輝度差は、１２７との差が小さい輝度差が選択されるので、それぞれ左から０、６３、６３、１９２、０、２５５、２５５が選択される。変換器３０，３１（図２参照）では、その輝度差に応じて輪郭強調係数が求められる。輪郭強調係数は、左から０、６３／１２７、６３／１２７、６３／１２７、０、０、０となり、その輪郭強調係数に応じた輪郭強調係数信号ｆが得られる。輪郭強調係数信号ｆには、一画素遅延信号ｂが加えられ出力信号ｇが得られる。このようにして、信号が複雑な波形を示した場合でも、基準輝度差との差に応じて輪郭強調量が調整される。
【００２８】
なお、本実施形態において、出力信号ｇは８ビットを超えてしまうこともあるが、その場合、出力信号における分解能が９ビット若しくは１０ビットに変更される。
【００２９】
本発明の第１の実施形態では、輝度差が大きい領域若しくは小さい領域では輪郭強調が抑制されるため、高輝度部や暗部の不自然な輪郭調教が低減できる。また、輪郭強調が必要な領域では輪郭強調が充分に掛けられるため、シャープな画像が得られる。
【００３０】
なお、フィルタ２９は図６に示すような構成にしても良い。すなわち、８ビットに量子化された輝度差がビット列に配列され、基準輝度値と比較される構成である。本実施形態においては、輝度差のデータが０であるならばすべてのビットが０にセットされる。そして、輝度差に応じて各ビットが１若しくは０にセットされていき、輝度差が２５５であるとき、すべてのビットが１にセットされる。輝度差のデータが１２７ならば、Ｄ０からＤ６ビットに１がセットされ、Ｄ７ビットに０がセットされる。輝度差のデータが１２８ならば、Ｄ７ビットに１が、Ｄ０からＤ６ビットに０がセットされる。
【００３１】
フィルタ２９では、この構成を利用して、Ｄ７ビットに１がセットされているか否かにより、フィルタ２５で選択された輝度差が基準輝度差（１２７）より大きいか否かが判定される。Ｄ７ビットに１がセットされていなければ、輝度差が基準輝度差（１２７）以下なので、輪郭強調信号ｅは変換器３０に送られる。Ｄ７ビットに１がセットされていれば、輝度差が基準輝度差（１２７）より大きいので、輪郭強調信号ｅは変換器３１に送られる。本変形例では、輝度値の判別においてＤ７ビットが０か１であるかのみを判別するのみで良く、データが基準輝度差より大きいか否かが素早く判定される。
【００３２】
図７に本発明の第２の実施形態における輪郭強調回路１４を示す。第２の実施形態が第１の実施形態と相違する点は、変換器３０、３１と乗算器２８の間にスイッチ３２が取り付けられている点である。以下相違点のみを説明する。
【００３３】
入力信号ａは、第１の実施形態と同様の処理がなされ、減算器２６で輪郭強調信号ｅ（＝ｂ−（ａ＋ｃ）／２）が得られる。また、変換器３０若しくは３１でも、輝度差に応じて輪郭強調係数が得られる。スイッチ３２は、目的に応じて、手動もしくは自動で変換器３０、３１のどちらか一方のみが、乗算回路２８に接続されるように構成されている。
【００３４】
スイッチ３２によって、変換器３０が乗算回路２８に接続されると、輝度差が基準輝度差（１２７）以下と検出される領域においてのみ輪郭強調が掛けられ、輝度差が基準輝度差（１２７）より大きい領域において、輪郭強調が掛けられない。これは輝度値が大きい領域が多い場合、出力画像全体が明るくなりすぎることを防止するのに有効な手段である。
【００３５】
一方、スイッチ３２によって、変換器３１が乗算器２８に接続されると、輝度差が基準輝度差（１２７）よりも大きいと検出される領域においてのみ輪郭強調が掛けられ、輝度差が基準輝度差（１２７）以下の領域において、輪郭強調が掛けられない。輝度差の小さいノイズが多い場合、そのノイズの強調を防止するのに有効な手段である。
【００３６】
以上のように第２の実施形態においては、目的に応じて、輝度差が大きい領域若しくは、小さい領域において選択的に、輪郭強調が掛けられない。これによりノイズ強調や擬似信号の抑制が効果的に抑えられる。
【００３７】
なお、本発明の第１、第２の実施形態では、１画素遅延信号および２画素遅延により、出力信号が生成されているが、１画素遅延信号のみによって出力信号ｇが生成される構成にしてもよい。
【００３８】
【発明の効果】
本発明においては、基準輝度差を観察に最適と考えられる輝度差に設定することにより、それぞれの輝度差における輪郭強調量が必要に応じて抑制される。これにより、擬似信号の発生やノイズの強調が防止され、鮮明な画像が得られる。
【図面の簡単な説明】
【図１】本発明の第一の実施形態である輪郭強調装置を備えた内視鏡システムを示すブロック図である。
【図２】本発明における輪郭強調回路の第１の実施形態を示す処理ブロック図である。
【図３】輝度差最大のときに輪郭強調回路で得られる信号を示す図である。
【図４】輝度差が中間値のときに輪郭強調回路で得られる信号を示す図である。
【図５】輪郭強調回路で得られる信号を示す図である。
【図６】フィルタで行なわれている処理の変形例を示す模式図である。
【図７】本発明における輪郭強調回路の第２の実施形態を示す処理ブロック図である。
【符号の説明】
１０　内視鏡
１４　輪郭強調回路
２０、２３　遅延器
２１、２４、２６　減算器
２２、２７　加算器
２５、２９　フィルタ
３０、３１　変換器
３２　スイッチ[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a contour emphasizing device that performs contour emphasis on an image obtained by an endoscope, for example.
[0002]
[Prior art]
Conventionally, in image processing, contour emphasis has been uniformly applied to the entire screen area. That is, in a region where the luminance difference between adjacent pixels is small, the amount of edge enhancement is multiplied by a small amount, and the luminance value of the obtained image data is not much increased compared to the original data. On the other hand, in a region where the luminance difference between adjacent pixels is large, the contour enhancement amount is multiplied by a large amount, and the luminance value of the obtained image data is also increased. As a result, the brightness difference between the regions is clarified, and a sharp image is obtained.
[0003]
However, when the luminance difference becomes extremely large, the contour enhancement amount is also multiplied accordingly, so that the luminance value of the image data becomes too large, and a pseudo signal such as halation may be generated in the area.
[0004]
In particular, image data obtained by an endoscope has a mixture of dark and bright regions due to its optical characteristics, and many regions have a large luminance difference. In addition, many objects to be observed with an endoscope have liquids such as bodily fluids attached thereto, and when these are irradiated with light, the light is excessively reflected, and the luminance difference from the vicinity thereof becomes extremely large. There is. Therefore, many pseudo signals are generated in an image obtained by the endoscope, so that the image is often unnatural. Further, in a dark area where the luminance difference is small, there is a lot of noise. If contour enhancement is equally applied to this area, the noise is enhanced and the image quality may be degraded.
[0005]
[Problems to be solved by the invention]
SUMMARY OF THE INVENTION It is an object of the present invention to obtain a clear and easy-to-view image by preventing generation of a pseudo signal due to contour enhancement.
[0006]
[Means for Solving the Problems]
An image edge enhancement device according to the present invention includes an edge enhancement unit that performs edge enhancement in accordance with a brightness difference between adjacent pixels for each pixel constituting image data, and a reference brightness difference in which the brightness difference is set in advance. A region emphasis adjusting means for maximizing the amount of outline emphasis in the outline emphasis. It is preferable that the contour emphasis adjusting means makes the contour emphasis amount relatively small in a region where the difference between the luminance difference and the reference luminance difference is relatively large. Thus, in a region where the luminance difference becomes too large, the amount of contour enhancement is suppressed, and generation of a pseudo signal is prevented.
[0007]
Preferably, the luminance value of the image data is quantized within a certain range, and the reference luminance difference is a substantially intermediate value within the certain range. Preferably, the contour emphasizing means performs the contour emphasis in accordance with the luminance difference having the smallest difference from the reference luminance difference among the luminance differences with the adjacent pixels. The imaging device for obtaining image data is preferably an endoscope.
[0008]
It is preferable that the contour emphasis adjusting means performs the contour emphasis only in a region where the luminance difference is detected to be smaller than the reference luminance difference. Further, the contour emphasis adjusting means may perform the contour emphasis only on an area where the luminance difference is detected to be larger than the reference luminance difference. This makes it possible to prevent contour enhancement in a region where the luminance difference is large or small as necessary.
[0009]
The image edge enhancement device image according to the present invention includes an edge enhancement unit that performs edge enhancement on data, and an edge enhancement only in an area where a brightness difference between adjacent pixels is detected to be smaller than a preset reference brightness difference. A contour emphasis adjusting means for emphasizing is provided. In addition, the image edge enhancement device according to the present invention includes an edge enhancement unit that performs edge enhancement on image data, and an image enhancement unit that performs an edge enhancement on an area where a brightness difference between adjacent pixels is detected to be larger than a preset reference brightness difference. And a contour emphasis adjusting unit for performing contour emphasis only.
[0010]
The image edge enhancement method according to the present invention further includes an edge enhancement step of performing edge enhancement for each pixel constituting the image data according to a brightness difference between adjacent pixels, and a reference brightness difference in which the brightness difference is set in advance. And a contour emphasis adjusting step of maximizing the amount of contour emphasis in the contour emphasis in the following region.
[0011]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.
FIG. 1 shows an endoscope system provided with an image contour enhancement device according to a first embodiment of the present invention. The endoscope system includes an endoscope 10 and an image processing device. The endoscope 10 is an imaging device for obtaining image data from a subject, and the image processing device is a device for processing image data obtained by the endoscope 10 so that the image data can be output as an image.
[0012]
In the CPU 12, control information is generated, and the generated control information is sent to the mask generation circuit 13 and the outline emphasis circuit 14. The mask generation circuit 13 and the outline emphasis circuit 14 are controlled based on the generated control information.
[0013]
The image data obtained by the endoscope 10 is converted into a video signal, and the video signal is sent to the adder 11. The adder 11 adds the mask information generated by the mask generation circuit 13 to the sent video signal. The video signal to which the mask information has been added is sent to the contour emphasizing circuit 14 and subjected to contour emphasis.
[0014]
The video signal on which the contour enhancement has been applied is sent to the video signal processing circuit 15, where the video signal is A / D-converted and separated into RGB by the video signal processing circuit 15, and is separated into R memory 16a, G memory 16b, and B memory 16c, respectively. Is stored in Each of the stored video signals is input to the video image processing circuit 17, subjected to D / A conversion, and then output as an image to an output monitor. The synchronization signal generator 18 generates a synchronization signal. The synchronization signal is sent to the timing circuit 19 and is used for controlling the operation timing of the video processing circuit 15 and the video processing circuit 17 and for controlling the writing and reading of the memories 16a, 16b and 16c.
[0015]
FIG. 2 shows the outline emphasis circuit 14. The contour emphasis circuit 14 performs contour emphasis on each pixel constituting the image data in accordance with a luminance difference between adjacent pixels. Here, the luminance difference is quantized by 8 bits (256 steps) and the calculation is performed, but the resolution may be changed according to the purpose.
[0016]
The input signal a is sent to the delay unit 20 and sent to the subtractor 21 and the adder 22. In the delay unit 20, the input signal a is delayed by one pixel, and a one-pixel delayed signal b is obtained. The one-pixel delay signal b is sent to the delay unit 23 and sent to the subtracters 21, 24, 26 and the adder 27. In the delay unit 23, the one-pixel delay signal b is delayed by one pixel to obtain a two-pixel delay signal c. The two-pixel delay signal c is sent to the adder 22 and the subtractor 24.
[0017]
In the adder 22, after the input signal a and the two-pixel delay signal c are added, the result is multiplied by a coefficient (1/2) to obtain an addition signal d (= (a + c) / 2). In the subtracter 26, the addition signal d sent from the adder 22 is subtracted from the one-pixel delay signal b, and an edge enhancement signal e (= b− (a + c) / 2) is obtained. The obtained contour emphasis signal e is sent to the multiplier 28.
[0018]
In the subtracter 21, the input signal a is subtracted from the one-pixel delay signal b, and its absolute value (| ba |) is obtained. Similarly, the subtracter 24 also subtracts the two-pixel delay signal c from the one-pixel delay signal b, and obtains the absolute value (| bc-). The absolute values obtained by the subtracters 21 and 24 are sent to the filter 25 as a luminance difference between adjacent pixels. The filter 25 selects a luminance difference having a small difference from the reference luminance difference among luminance differences (| ba | or | bc |) between each pixel and an adjacent pixel. If the values of | ba | and | bc− become equal, the one with the larger value of the luminance difference is selected. The selected value is sent to the filter 29, and the filter 29 reads whether or not the difference is larger than the reference luminance difference. It is sent to 31. Note that the reference luminance difference is set to a luminance difference at which contour enhancement is most required. In the present embodiment, the reference luminance difference is set in advance to 127 which is an intermediate value within a range where the luminance difference is quantized.
[0019]
In the converters 30 and 31, the contour enhancement amount is maximized in a region where the luminance difference selected by the filter 25 is the reference luminance difference, and the difference between the selected luminance difference and the reference luminance difference is relatively large. The contour emphasis coefficient is set so that the contour emphasis amount becomes relatively small in the region. That is, converter 30 increases the contour emphasis coefficient in the region where the selected luminance difference increases, and converter 31 decreases the contour emphasis coefficient in the region where the selected luminance difference value increases. Is set to In the present embodiment, the relationship between the selected luminance difference and the contour emphasis coefficient is set to be proportional, and the minimum value of the contour emphasis coefficient is set to 0 and the maximum value is set to 1 I have.
[0020]
For example, in the converter 30, the contour enhancement coefficient of the selected pixel whose luminance difference is 0 is 0, the contour enhancement coefficient of the pixel whose luminance difference is 1 is 1/127, and the pixel whose luminance difference is 126 is 126. The contour emphasis coefficient is set to 126/127, and the contour emphasis coefficient of a pixel having a luminance difference of 127 is set to 1. On the other hand, in the converter 31, the contour emphasis coefficient of the pixel whose luminance difference is 128 is 1, the contour emphasis coefficient of the pixel whose luminance difference is 129 is 126/127, and the contour emphasis coefficient of the pixel whose luminance difference is 254. Is set to 1/127, and the contour emphasis coefficient of a pixel having a luminance difference of 255 is set to 0.
[0021]
The contour emphasis coefficients obtained by the converters 30 and 31 are sent to a multiplier 28, where the contour emphasis signal e is multiplied by the contour emphasis coefficient, and an emphasis coefficient multiplication signal f (= e × contour emphasis coefficient) ) Is required. The enhancement coefficient multiplication signal f is sent to the adder 27, where the enhancement coefficient multiplication signal f is added to the one-pixel delay signal b to obtain an output signal g.
[0022]
3 to 5 show signal examples obtained by the contour emphasizing circuit 14.
FIG. 3 shows an example of a signal obtained by the contour emphasizing circuit 14 when the luminance value of each pixel of the input signal a is 0, 255, 255, 255, 255, 0, 0. For example, the brightness difference (| ba |, | bc |) of each pixel is such that the brightness difference of the first pixel from the left is 0, 0, and that of the second pixel is 255, 0. In this case, in the filter 25 (see FIG. 2), for example, since the two luminance differences of the first pixel from the left are each 0, the luminance difference in this pixel is selected as 0. Since the luminance difference of the second pixel is 255, 0, 0 having a small difference from 127 is selected. The luminance difference is similarly selected for all the following pixels, and in this embodiment, the luminance difference for all the pixels is selected as 0.
[0023]
The converters 30 and 31 (see FIG. 2) determine the outline emphasis coefficient of each pixel according to the selected luminance difference. In the signal shown in FIG. 3, the selected luminance differences are all 0, and thus the edge enhancement coefficients for all the pixels are 0.
[0024]
The adder 22 (see FIG. 2) calculates the addition signal d, and the subtracter 26 calculates the contour emphasis signal e. The contour emphasis signal e is sent to the multiplier 28, where it is multiplied by a contour emphasis coefficient. The contour enhancement coefficient is 0 for all the pixels in the signal example of FIG. Therefore, in the enhancement coefficient multiplication signal f, the luminance values of all pixels are 0, and the output signal g is the same signal as the one-pixel delay signal b. In this way, no contour emphasis is applied to the pixel having the maximum luminance difference (255), thereby preventing generation of a pseudo signal.
[0025]
FIG. 4 shows a signal example obtained by the contour emphasizing circuit 14 when the luminance value of each pixel of the input signal is 0, 127, 127, 127, 127, 0, 0. For example, as for the luminance differences | ba− || bc | of the respective pixels, the luminance difference of the first pixel from the left is 0, 0, and the luminance difference of the second pixel is 127, 0. In the filter 25 (see FIG. 2), for example, since the luminance difference of the second pixel is 127, 0, 127 having a small difference from 127 is selected. The luminance difference is similarly selected for all the pixels below, and in the present embodiment, 0, 127, 127, 0, 0, 127, 127, 0 from the left are selected as the luminance difference of each pixel.
[0026]
The converters 30 and 31 (see FIG. 2) determine the outline emphasis coefficient of each pixel according to the selected luminance difference. In the signal shown in FIG. 4, the contour emphasis coefficient of the selected pixel having the luminance difference of 0 is 0. On the other hand, the contour enhancement coefficient of the selected pixel having a luminance difference of 127 is 1, and the contour enhancement coefficients are 0, 1, 1, 0, 0, 1, 1, 0 from the left. Each pixel of the contour emphasis signal e obtained by the subtracter 26 is multiplied by these contour emphasis coefficients by the multiplier 28 to obtain an emphasis coefficient multiplication signal f. That is, at the pixel where the selected luminance difference matches the reference luminance difference, the outline emphasis signal e is multiplied by 1 which is the maximum value of the outline emphasis coefficient, and the emphasis coefficient multiplication signal f is obtained. The one-pixel delay signal b is added to the contour emphasis coefficient signal f, and an output signal g is obtained. In this manner, an output signal g with the maximum edge enhancement is obtained for a pixel whose luminance difference is equal to the reference luminance difference.
[0027]
FIG. 5 shows an example of a signal obtained by the contour emphasizing circuit 14 when the luminance value of each pixel of the input signal is 0, 63, 255, 255, 255, 0, 0. The luminance difference | ba− || bc | of each pixel is, for example, that the luminance difference of the first pixel from the left is 0, 0, the second pixel is 63, 0, and the third pixel is 192, 63. And the fourth pixel is 0,192. In this case, since a luminance difference between each pixel and a difference from 127 is small, 0, 63, 63, 192, 0, 255, 255 are selected from the left. In the converters 30 and 31 (see FIG. 2), an outline emphasis coefficient is obtained according to the luminance difference. The contour enhancement coefficients are 0, 63/127, 63/127, 63/127, 0, 0, 0 from the left, and a contour enhancement coefficient signal f corresponding to the contour enhancement coefficient is obtained. The one-pixel delay signal b is added to the contour emphasis coefficient signal f, and an output signal g is obtained. In this way, even when the signal has a complicated waveform, the contour enhancement amount is adjusted according to the difference from the reference luminance difference.
[0028]
In this embodiment, the output signal g may exceed 8 bits, but in this case, the resolution of the output signal is changed to 9 bits or 10 bits.
[0029]
In the first embodiment of the present invention, contour enhancement is suppressed in a region where the luminance difference is large or small, so that unnatural contour training in a high luminance portion or a dark portion can be reduced. In a region where contour enhancement is required, a sharp image is obtained because the contour enhancement is sufficiently applied.
[0030]
Note that the filter 29 may be configured as shown in FIG. That is, the luminance difference quantized to 8 bits is arranged in a bit string and compared with the reference luminance value. In the present embodiment, if the luminance difference data is 0, all bits are set to 0. Then, each bit is set to 1 or 0 according to the luminance difference, and when the luminance difference is 255, all bits are set to 1. If the data of the luminance difference is 127, 1 is set in the D0 to D6 bits, and 0 is set in the D7 bit. If the luminance difference data is 128, 1 is set in the D7 bit and 0 is set in the D0 to D6 bits.
[0031]
Using this configuration, the filter 29 determines whether or not the luminance difference selected by the filter 25 is larger than the reference luminance difference (127) based on whether or not the D7 bit is set to 1. If the D7 bit is not set to 1, the luminance difference is equal to or smaller than the reference luminance difference (127), and the contour emphasis signal e is sent to the converter 30. If the D7 bit is set to 1, the luminance difference is larger than the reference luminance difference (127), and the contour emphasis signal e is sent to the converter 31. In this modification, it is only necessary to determine whether the D7 bit is 0 or 1 in the determination of the luminance value, and it is quickly determined whether or not the data is larger than the reference luminance difference.
[0032]
FIG. 7 shows an outline emphasis circuit 14 according to the second embodiment of the present invention. The second embodiment is different from the first embodiment in that a switch 32 is provided between the converters 30 and 31 and the multiplier 28. Hereinafter, only the differences will be described.
[0033]
The input signal a is subjected to the same processing as in the first embodiment, and the contour emphasis signal e (= b− (a + c) / 2) is obtained by the subtractor 26. Also, the converter 30 or 31 can also obtain an outline emphasis coefficient according to the luminance difference. The switch 32 is configured such that only one of the converters 30 and 31 is connected to the multiplication circuit 28 manually or automatically according to the purpose.
[0034]
When the converter 30 is connected to the multiplying circuit 28 by the switch 32, contour emphasis is applied only to a region where the luminance difference is detected to be equal to or smaller than the reference luminance difference (127), and the luminance difference is compared with the reference luminance difference (127). In a large area, contour enhancement is not applied. This is an effective means for preventing the entire output image from becoming too bright when there are many areas having a large luminance value.
[0035]
On the other hand, when the converter 32 is connected to the multiplier 28 by the switch 32, contour enhancement is applied only to an area where the luminance difference is detected to be larger than the reference luminance difference (127), and the luminance difference is changed to the reference luminance difference. (127) In the following areas, contour enhancement is not applied. When there is a lot of noise with a small luminance difference, this is an effective means for preventing the noise from being emphasized.
[0036]
As described above, in the second embodiment, contour enhancement is not selectively applied to a region having a large luminance difference or a region having a small luminance difference depending on the purpose. This effectively suppresses noise enhancement and suppression of pseudo signals.
[0037]
In the first and second embodiments of the present invention, the output signal is generated by the one-pixel delay signal and the two-pixel delay, but the output signal g is generated only by the one-pixel delay signal. Is also good.
[0038]
【The invention's effect】
In the present invention, by setting the reference luminance difference to a luminance difference considered to be optimal for observation, the amount of contour enhancement at each luminance difference is suppressed as necessary. Thereby, generation of a pseudo signal and enhancement of noise are prevented, and a clear image is obtained.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating an endoscope system including a contour emphasis device according to a first embodiment of the present invention.
FIG. 2 is a processing block diagram showing a first embodiment of the contour emphasizing circuit according to the present invention.
FIG. 3 is a diagram showing a signal obtained by an outline emphasizing circuit when the luminance difference is maximum.
FIG. 4 is a diagram illustrating a signal obtained by an outline emphasizing circuit when a luminance difference is an intermediate value.
FIG. 5 is a diagram illustrating signals obtained by an outline emphasis circuit.
FIG. 6 is a schematic diagram showing a modified example of the processing performed by the filter.
FIG. 7 is a processing block diagram showing a second embodiment of the contour emphasizing circuit according to the present invention.
[Explanation of symbols]
Reference Signs List 10 endoscope 14 contour emphasis circuit 20, 23 delay unit 21, 24, 26 subtractor 22, 27 adder 25, 29 filter 30, 31 converter 32 switch

Claims

Contour enhancing means for performing contour enhancement in accordance with the luminance difference between adjacent pixels for each pixel constituting the image data; and a contour in the contour emphasis in a region where the luminance difference is a preset reference luminance difference. An image edge enhancement device comprising: an edge enhancement adjustment unit that maximizes an enhancement amount.

2. The image edge enhancement according to claim 1, wherein the edge enhancement adjustment unit relatively reduces the edge enhancement amount in the region where the difference between the luminance difference and the reference luminance difference is relatively large. apparatus.

2. The image contour emphasizing device according to claim 1, wherein the luminance value of the image data is quantized within a certain range, and the reference luminance difference is set to a substantially intermediate value of the range.

2. The image edge enhancement device according to claim 1, wherein the edge enhancement unit performs edge enhancement in accordance with a brightness difference having a smallest difference from the reference brightness difference among brightness differences from adjacent pixels. 3.

2. The image edge enhancement device according to claim 1, wherein the edge enhancement adjustment unit performs edge enhancement only in the area where the luminance difference is detected to be smaller than the reference luminance difference.

2. The image edge enhancement device according to claim 1, wherein the edge enhancement adjustment unit performs edge enhancement only in an area where the luminance difference is detected to be larger than a reference luminance difference.

2. The image edge enhancement device according to claim 1, wherein the imaging device for obtaining the image data is an endoscope.

An image comprising: an outline emphasis unit that performs outline emphasis on image data; and an outline emphasis adjustment unit that performs outline emphasis only in an area where a luminance difference between adjacent pixels is detected to be smaller than a preset reference luminance difference. Contour enhancement device.

Contour enhancement means for performing contour enhancement on image data, and contour enhancement adjustment means for performing the contour enhancement only in an area where a luminance difference between adjacent pixels is detected to be larger than a preset reference luminance difference. Image edge enhancement device provided.

A contour emphasizing step of performing contour emphasis on each pixel constituting the image data in accordance with a luminance difference between adjacent pixels, and a contour in the contour emphasis in a region where the luminance difference is a preset reference luminance difference. An edge enhancement adjusting step of maximizing an enhancement amount.