JP2005020386A - Image pickup device and image pickup method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To adjust the code amount of image data to be compression-encoded without putting processing burdens on a microcomputer by a small memory capacity even for different output image sizes. <P>SOLUTION: A code amount change ratio calculation part 14a changes an SCF value to be multiplied with a basic quantization table and calculates the change ratio of the code amount of the image data corresponding to the change of the SCF value. In the case that the code amount of the image data is out of a prescribed range, an SCF value deciding part 14b refers to the calculated change ratio and decides the SCF value to turn the code amount to be within the prescribed range. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

【００４２】
図１０は、符号減少割合テーブルの例である。
ここで、“ＴＢＬ＿ＳＣＦ”はテーブル上でのＳＣＦ値、“ＴＢＬ＿ＲＥＤ＿ＲＡＴＥ”はテーブル上での符号減少割合を示す。なお、図１０のように、符号減少割合テーブル作成には、ＳＣＦ値＝１．０を真中にし、式（３）で計算したようなステップで、ＳＣＦ値に対応する符号減少割合のテーブルが作成される。
【００４３】
次に、ＳＣＦ値決定部１４ｂでは、圧縮符号化処理後の画像データの符号量を検出し、符号量が制限符号量で規定された範囲外である場合に、符号量の変化割合（例えば、図１０のような符号減少割合テーブル）を参照し、制限符号量の範囲内に符号量を調整するようなＳＣＦ値を決定する。具体的な調整方法については後述する。
【００４４】
量子化部１４では、決定されたＳＣＦ値を式（１）のように基本量子化テーブルに掛け合わせ量子化テーブルを作成し、その量子化テーブルを用いて量子化を行う。
【００４５】
図１１は、Ｙ（輝度）成分用の基本量子化テーブルの例である。
また、図１２は、Ｃ（色差）成分用の基本量子化テーブルの例である。
図１１、１２のように高い周波数成 分（図の右下部分）は目立ちにくい特徴があり、低い周波数成分（図の左上部分）は目立ちやすい特徴があるので、低い周波数成分は省略を少なくするために、量子化ステップを小さく、高い周波数成分は省略を多くするために、量子化ステップを大きくしている。
【００４６】
なお、人間の視覚は輝度成分に比べて色差成分の感度が低い特徴があるので、図１２のＣ成分用の基本量子化テーブルのように色差成分の省略を多くして符号量を輝度成分に比べて少なくしている。
【００４７】
量子化された画像データは、エントロピー符号化部１５により、ハフマン符号化や算術符号化され、ＪＰＥＧデータとして出力される。
以上のように、異なる画像サイズ、大幅に異なる符号量の画像についても同一の符号量の変化割合（例えば、図１０の符号減少割合テーブル）を用いて、符号量の調整が可能であるので、少ないメモリ容量で符号量予測及び符号量の調整が可能になる。
【００４８】
次に、本発明の実施の形態の撮像装置１０を具体的に説明する。
図１３は、撮像装置システムのハードウェア構成図である。
図のように、撮像装置システムは、図１で示した撮像装置１０と対応しているカメラモジュール２０と、表示制御ブロック３０とからなる。
【００４９】
カメラモジュール２０は、レンズ２１、センサ２２、ＣＤＳ（相関二重サンプリング）／ＡＧＣ（自動利得制御）２３、カメラ信号処理ブロック２４、ＪＰＥＧエンコーダ２５、マイコン２６、ＥＥＰＲＯＭ（Ｅｌｅｃｔｒｉｃａｌｌｙ Ｅｒａｓａｂｌｅ Ｐｒｏｇｒａｍｍａｂｌｅ Ｒｅａｄ−Ｏｎｌｙ Ｍｅｍｏｒｙ）２７から構成される。
【００５０】
ここで、レンズ２１は、被写体に対し、集光と結像を行う。
センサ２２は、光を電気信号に変換する。また、電子シャッターにより電荷の蓄積時間が可変する。センサ２２は、例えばＣＣＤ（Ｃｈａｒｇｅ Ｃｏｕｐｌｅｄ Ｄｅｖｉｃｅ）センサである。
【００５１】
ＣＤＳ／ＡＧＣ２３は、相関二重サンプリング回路によるノイズ除去と自動利得制御を行う。
カメラ信号処理ブロック２４は、Ａ／Ｄ（アナログ／デジタル）コンバータを有し、センサ２２から入力されたアナログ信号をデジタル信号に変換する。さらに、入力信号を映像信号（図ではＹＵＶ信号）に変換し出力する。なお、“ＳＹＮＣ信号”は同期信号であり、カメラ信号処理ブロック２４からＪＰＥＧエンコーダ２５及び表示制御ブロック３０に入力され、各ブロックに処理での同期をとる。ＹＵＶ信号は、Ｙ成分と、Ｙ成分と赤色成分（Ｃｒ）の差（Ｕ）、Ｙ成分と青色成分（Ｃｂ）の差（Ｖ）の３つの情報で色を表した形式による映像信号である。
【００５２】
ＪＰＥＧエンコーダ２５は、カメラ信号処理ブロック２４からの出力をＪＰＥＧ形式にエンコード（圧縮符号化）する。
マイコン２６は、図示を省略しているが、図１３の撮像装置システムを制御するためのプログラムを格納したＲＯＭ及びプログラムの実行時に必要なデータを一時格納するＲＡＭを有し、撮像装置システムの各部を制御する。
【００５３】
ＥＥＰＲＯＭ２７は、前述した基本量子化テーブル及び、符号量の変化割合を格納する。
表示制御ブロック３０は、ＬＣＤ（Ｌｉｑｕｉｄ Ｃｒｙｓｔａｌ Ｄｉｓｐｌａｙ）出力、画像伸張などを行う。また、図示しない上位の制御ブロック（マイコン）からのコマンドをＩＩＣ（Ｉｎｔｅｒ Ｉｎｔｅｇｒａｔｅｄ Ｃｉｒｃｕｉｔ）通信などで、マイコン２６に通知する。
【００５４】
図１３において、レンズ２１及びセンサ２２は、図１で示した撮像部１１に対応している。また、カメラ信号処理ブロック２４（Ａ／Ｄコンバータ）は、デジタル化部１２に対応している。また、ＪＰＥＧエンコーダ２５と、マイコン２６の機能により、ＤＣＴ変換部１３、量子化部１４、符号量変化割合算出部１４ａ、ＳＣＦ値決定部１４ｂ、エントロピー符号化部１５の機能が実現できる。
【００５５】
このような撮像装置システムにおいて、被写体がレンズ２１により集光され、センサ２２に結像されると、センサ２２は光を電気信号に変換する。さらに電気信号になった画像データに対して、ＣＤＳ／ＡＧＣ２３でノイズ除去や自動利得制御が行われたのちカメラ信号処理ブロック２４に入力され、ＹＵＶ信号に変換されて表示制御ブロック３０に出力される。例えば、デジタルカメラなどにおいて、シャッターが押されていない場合、表示モードによっては、映像信号は表示制御ブロック３０で処理され、液晶モニタ（図示せず）上で被写体を常に表示させるようにしてもよい。このとき例えば、被写体が、３０フレーム／秒で表示される。
【００５６】
シャッターが押された場合、そのときのフレームの映像信号が８×８の６４画素を１ブロックとして、ブロックごとにＪＰＥＧエンコーダ２５で処理されＪＰＥＧ形式に圧縮符号化される。
【００５７】
ここで、本発明の実施の形態におけるＪＰＥＧエンコーダ２５の詳細を説明する。
図１４は、ＪＰＥＧエンコーダ及びそのエンコード処理に用いられる構成を示す図である。
【００５８】
図のように、ＪＰＥＧエンコーダ２５は、ＪＰＥＧエンコードブロック２５−１、符号量レジスタ２５−２を有する。
また、ＪＰＥＧエンコードブロック２５−１は、Ｙ成分用基本量子化テーブルレジスタ２５−１ａ、Ｃ成分用基本量子化テーブルレジスタ２５−１ｂ、Ｙ成分用ＳＣＦレジスタ２５−１ｃ、Ｃ成分用ＳＣＦレジスタ２５−１ｄを有する。
【００５９】
Ｙ成分用基本量子化テーブルレジスタ２５−１ａは、量子化の際、ＥＥＰＲＯＭ２７から取り出されたＹ成分用の基本量子化テーブルを一時格納する。
Ｃ成分用基本量子化テーブルレジスタ２５−１ｂは、量子化の際、ＥＥＰＲＯＭ２７から取り出されたＣ成分用の基本量子化テーブルを一時格納する。
【００６０】
Ｙ成分用ＳＣＦレジスタ２５−１ｃは、量子化の際、ＥＥＰＲＯＭ２７から取り出されたＹ成分用のＳＣＦ値を一時格納する。
Ｃ成分用ＳＣＦレジスタ２５−１ｄは、量子化の際、ＥＥＰＲＯＭ２７から取り出されたＣ成分用のＳＣＦ値を一時格納する。
【００６１】
符号量レジスタ２５−２は、ＪＰＥＧエンコードブロック２５−１から出力されたＪＰＥＧ形式の画像データの符号量を計数する。例えばカウンタのような構成であればよい。
【００６２】
このようなＪＰＥＧエンコーダ２５において、８×８画素のブロックごとに、マイコン２６の制御のもと、ＤＣＴ変換処理がおこなわれ、その後、量子化処理が行われる。
【００６３】
量子化処理時には、Ｙ成分用基本量子化テーブルレジスタ２５−１ａにはＹ成分用の基本量子化テーブルが、Ｃ成分用基本量子化テーブルレジスタ２５−１ｂにはＣ成分用の基本量子化テーブルが、Ｙ成分用ＳＣＦレジスタ２５−１ｃにはＹ成分用のＳＣＦ値が、Ｃ成分用ＳＣＦレジスタ２５−１ｄにはＣ成分用のＳＣＦ値が、ＥＥＰＲＯＭ２７から取り出され格納される。
【００６４】
画像データは、取り出されたこれらのものを用いて量子化される。
すなわち、ＳＣＦ値を基本量子化テーブルの各要素値に掛け合わせて量子化テーブルを作成し、その量子化テーブルの各要素値で、ＤＣＴ変換処理の結果得られた８×８のＤＣＴ係数を割る。このようにして量子化し、エントロピー符号化処理を行った後、得られたＪＰＥＧ形式の画像データの符号量を符号量レジスタ２５−２で計数する。その結果、出力符号量が上限符号量から下限符号量の制限符号量以内である場合は、そのままＪＰＥＧエンコーダ２５から出力する。
【００６５】
なお、上限符号量と下限符号量はＪＰＥＧエンコードブロック２５−１に設定し、制限符号量エラーをマイコン２６に通知するようにしてもよいし、外部から指定された値をマイコン２６内部のＲＡＭに持ち、出力される符号量と比較して、制限符号量エラーをマイコン２６で検知するようにしてもよい。
【００６６】
以下、出力符号量が制限符号量以内ではない場合の量子化処理を説明する。
本発明の実施の形態では、例えば、工場出荷前にある標準画像を写し、マイコン２６の制御のもと、量子化の際に用いる基本量子化テーブルに乗算するＳＣＦ値を所定のステップで変化させ、ＳＣＦ値の変化に応じた画像データの符号量の変化割合を算出し、例えば、図１０に示したような符号減少割合テーブルとして管理する。このような符号減少割合テーブルは、ＥＥＰＲＯＭ２７に格納される。
【００６７】
画像データの符号量が、制限符号量の範囲外である場合に、マイコン２６は、初期化処理時にＲＡＭに展開しておいたＥＥＰＲＯＭ２７に格納された符号量の変化割合を参照して以下の処理を行う。
【００６８】
まず目標とする符号量（以下目標符号量と呼ぶ）を定める。
目標符号量としては、例えば以下のようなものがある。
１．上限符号量と下限符号量の中間の符号量
２．上限符号量エラー時には、上限符号量よりも小さな符号量で、下限符号量エラー時には、下限符号量よりも大きな符号量。
３．どちらのエラー時にも上限符号量より少し小さな値。
【００６９】
その後、符号量エラー時の符号量より、目標符号量にするための割合（以下絶対目標符号量割合と呼ぶ）を求める。絶対目標符号量割合は、エラー時の符号量を目標符号量で割ることによって算出される。
【００７０】
さらに、マイコン２６は、エラー時に使用したＳＣＦ値の符号量の変化割合に絶対目標符号量割合を掛け、相対的な目標符号量割合（以下相対目標符号量割合と呼ぶ）を算出する。算出した相対目標符号量割合を用いて、例えば図１０のような、符号減少割合テーブルを参照して、次のフレームの量子化に用いるＳＣＦ値を決定する。
【００７１】
その後、ＪＰＥＧエンコードブロック２５−１において、Ｙ成分用ＳＣＦレジスタ２５−１ｃにはＹ成分用のＳＣＦ値が、Ｃ成分用ＳＣＦレジスタ２５−１ｄにはＣ成分用のＳＣＦ値が格納され、式（１）のように基本量子化テーブルに掛け合わせ量子化テーブルを作成し、その量子化テーブルを用いて量子化が行われる。
【００７２】
上記のマイコン２６による、前述した目標符号量の２番目の例についてのＳＣＦ値の自動調整処理について、フローチャートでまとめる。
図１５は、ＳＣＦ値の自動調整処理の流れを示すフローチャートである。
【００７３】
なお、図では、主に上限符号量エラーの場合の処理内容を表記している。
また、ここでは、符号量の変化割合として符号減少割合をテーブル化したものを利用する場合について示しており、“ＴＢＬ＿ＲＥＤ＿ＲＡＴＥ［ＣＮＴ］”は、カウンタ値ＣＮＴに対応する符号減少割合テーブル上での符号減少割合を示す。また、“ＴＢＬ＿ＳＣＦ［ＣＮＴ］”はカウンタ値ＣＮＴに対応する符号減少割合テーブル上のＳＣＦ値を示している。なお、カウンタ値ＣＮＴは、符号減少割合テーブル上での位置を示し、例えば、図１０で示したような符号減少割合テーブルにおいて、ＳＣＦ値が最小（図１０の例では、ＴＢＬ＿ＳＣＦ＝０．００３９の値）の場合は、カウンタ値ＣＮＴ＝０となり、ＳＣＦ値が最大（図１０の例では、ＴＢＬ＿ＳＣＦ＝２５６）の場合は、カウンタ値ＣＮＴも最大となる。
【００７４】
ＳＣＦ値の自動調整処理は、まず、カウンタ更新処理を行う。ここではカウンタ値ＣＮＴに、符号量エラー時のフレームに使用していたＳＣＦ値のカウンタ値ＰＡＳＴ＿ＣＮＴをセットする（ステップＳ１０）。
【００７５】
次に相対目標符号量割合を計算する。相対目標符号量割合は、絶対目標符号量（出力符号量／制限符号量）に、現在のカウンタ値ＣＮＴにおける符号減少割合ＴＢＬ＿ＲＥＤ＿ＲＡＴＥ［ＣＮＴ］を掛けることによって算出する。ここで制限符号量は、上限符号量エラー時は上限符号量であり、下限符号量エラー時は下限符号量である（ステップ１１）。
【００７６】
次にカウンタ値ＣＮＴを上限符号量エラー時にはプラス１し、下限符号量エラー時にはマイナス１する（ステップＳ１２）。
次に、上限符号量エラー時には、符号減少割合ＴＢＬ＿ＲＥＤ＿ＲＡＴＥ［ＣＮＴ］が、ステップＳ１１で算出した相対目標符号量割合を上回るか否かを判断する。上回る場合はステップＳ１５の処理に進み、そうでない場合は、ステップＳ１４に進む。一方、下限符号量エラー時には、符号減少割合ＴＢＬ＿ＲＥＤ＿ＲＡＴＥ［ＣＮＴ］が、ステップＳ１１で算出した相対目標符号量割合を下回るか否かを判断する。下回る場合はステップＳ１５の処理に進み、そうでない場合は、ステップＳ１４に進む（ステップＳ１３）。
【００７７】
ステップＳ１４の処理では、上限符号量エラー時には、カウンタ値ＣＮＴが最大値か否かの判定を行う。ここで、カウンタ値ＣＮＴが最大値ならばステップＳ１５へ、そうでない場合は、ステップＳ１２からの処理を繰り返す。一方、下限符号量エラー時には、カウンタ値ＣＮＴが最小値か否かの判定を行う。ここで、カウンタ値ＣＮＴが最小値ならばステップＳ１５へ、そうでない場合は、ステップＳ１２からの処理を繰り返す。
【００７８】
ステップＳ１５の処理では、ＳＣＦレジスタ（図１４のＹ成分用ＳＣＦレジスタ２５−１ｃ及びＣ成分用ＳＣＦレジスタ２５−１ｄ）と、量子化に用いるＳＣＦ値のカウンタ値ＰＡＳＴ＿ＣＮＴを更新する。すなわち、ステップＳ１４までの処理で決まったカウンタ値ＣＮＴにより、ＳＣＦレジスタには、符号量の変化割合テーブル上のＳＣＦ値ＴＢＬ＿ＳＣＦ［ＣＮＴ］をセットし、量子化に用いるＳＣＦ値のカウンタ値ＰＡＳＴ＿ＣＮＴには、カウンタ値ＣＮＴをセットする。以上により、ＳＣＦ値の自動調整処理を終了する。
【００７９】
このように、異なる画像サイズ、符号量の画像についても同一の符号量の変化割合のデータを用いてＳＣＦ値を調整することで、少ないメモリ容量で符号量予測及び符号量の調整が可能になる。
【００８０】
さらに、従来カメラモジュール２０の上位の制御ブロックで行っているファイルサイズ制御をカメラモジュール２０で行ったので、余分な処理を上位の制御ブロックに入れる必要がなくなる。よって、上位の制御ブロックにおけるＲＯＭ、ＲＡＭ容量の削減、また、処理負担を軽減することができる。
【００８１】
なお、上記では、符号量の変化割合をマイコン２６の制御のもと算出するとして説明したが、符号量の変化割合を、例えば、図１０で示したようなテーブルデータとして外部からＥＥＰＲＯＭ２７に書き込むようにしてもよい。
【００８２】
また、符号量の変化割合は、前述の考察のように、基本量子化テーブルによって異なるものにするべきであるため、工場出荷前に基本量子化テーブルを変更するような場合は、変更した基本量子化テーブルに対応した符号量変化割合を前述の方法により算出して調整しなおすか、外部からＥＥＰＲＯＭ２７に与えることによって対応することができる。
【００８３】
また、上記では、制御可能なＳＣＦ値を１／２５６〜２５６としたが、量子化テーブルの要素の値は１〜２５５までであることから、基本量子化テーブルの最大要素と最小要素からＳＣＦ値の必要な制御範囲を求めるようにしてもよい。
【００８４】
例えば、ＳＣＦ値の制御最大値＝２５５／基本量子化テーブルの最小値、ＳＣＦ値の制御最小値＝１／基本量子化テーブルの最大値、という式で制御範囲を決定する。このようにＳＣＦ値の制御範囲を限定することで、高速に制御することが可能になる。
【００８５】
【発明の効果】
以上説明したように本発明では、基本量子化テーブルに乗算するＳＣＦ値を変化させ、ＳＣＦ値の変化に応じた画像データの符号量の変化割合を算出して、画像データの符号量が、所定の範囲外である場合、その変化割合を参照して、符号量を所定の範囲内とするようなＳＣＦ値を決定するので、出力画像サイズに係わらず、同一の符号量の変化割合を用いて符号量の調整が可能となる。
【００８６】
これによって、少ないメモリ容量で符号量の調整が可能になる。
【図面の簡単な説明】
【図１】本発明の実施の形態の撮像装置の概略の機能ブロック図である。
【図２】ＳＣＦ値とＪＰＥＧ形式の画像データの符号量との関係を示す図である。
【図３】ＳＣＦ値と符号減少割合の関係を示す図である。
【図４】同じ画を異なる３つの画像サイズでとった場合の画像データの符号量とＳＣＦ値の関係を示す図である。
【図５】同じ画を異なる３つの画像サイズでとった場合の画像データの符号減少割合とＳＣＦ値の関係を示す図である。
【図６】同じ画像サイズであるが、極端に符号量の異なる画のＳＣＦ値に対する符号量の変化を示す図である。
【図７】符号量の異なる画のＳＣＦ値に対する符号減少割合の変化を示す図である。
【図８】３種類の基本量子化テーブルについてＳＣＦ値と符号量の関係を示す図である。
【図９】３種類の基本量子化テーブルについてＳＣＦ値と符号減少割合の関係を示す図である。
【図１０】符号減少割合テーブルの例である。
【図１１】Ｙ（輝度）成分用の基本量子化テーブルの例である。
【図１２】Ｃ（色差）成分用の基本量子化テーブルの例である。
【図１３】撮像装置システムのハードウェア構成図である。
【図１４】ＪＰＥＧエンコーダ及びそのエンコード処理に用いられる構成を示す図である。
【図１５】ＳＣＦ値の自動調整処理の流れを示すフローチャートである。
【符号の説明】
１０……撮像装置、１１……撮像部、１２……デジタル化部、１３……ＤＣＴ変換部、１４……量子化部、１４ａ……符号量変化割合算出部、１４ｂ……ＳＣＦ値決定部、１５……エントロピー符号化部[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image pickup apparatus and an image pickup method, and more particularly to an image pickup apparatus and an image pickup method for taking in and compressing and encoding image data of a subject.
[0002]
[Prior art]
JPEG (Joint Photographic Expert Group) is known as a method for capturing and encoding image data of a subject captured as a still image.
[0003]
In the JPEG method, image data captured by an algorithm such as Discrete Cosine Transform (DCT), quantization, or entropy coding is compression-coded.
[0004]
Hereinafter, conventional quantization processing will be described.
In the quantization, image data is quantized (divided) for each block of 8 × 8 64 pixels after DCT processing using a quantization table. The quantization table is a value obtained by multiplying the basic quantization table and a scaling factor (hereinafter referred to as SCF or SCF value), and is represented by the following expression.
[0005]
[Expression 1]
QTable [i, j] = SCF × BaseQTable [i, j] (1)
Here, “QTable [i, j]” indicates the [i, j] -th value of the quantization table, and “BaseQTable [i, j]” indicates the [i, j] -th value of the basic quantization table. ing. The SCF value changes the quantization step width and changes the code amount of the image data. The code amount decreases as the SCF value increases, and the code amount increases as the SCF value decreases. Also, there are quantization tables for Y (luminance) components and C (color difference) components, and SCF values for Y component and C component are prepared separately.
[0006]
In general, the larger the code amount, the better the image quality, but the code amount allowed in the system (hereinafter referred to as the limited code amount) is different.
Conventionally, when the SCF value at the time of quantization has a plurality of fixed values and is quantized with the first SCF value and not within the range defined by the limited code amount, the next SCF value is used to quantize The process of performing the conversion is repeated until it falls within the range of the limited code amount. In such a case, there is a problem that it is necessary to mount a frame memory in order to obtain an optimum SCF value, which is costly and increases the chip size.
[0007]
For this reason, it is necessary to predict the code amount for the SCF value, and several techniques are disclosed.
For example, in code amount prediction, there is a code amount for each SCF value as a table, and a code amount corresponding to an input SCF value is uniquely predicted (see Patent Document 1).
[0008]
[Patent Document 1]
JP 2002-369198 (paragraph numbers [0083], [0084], FIG. 7)
[0009]
[Problems to be solved by the invention]
However, the above prior art has the following problems.
In recent years, for example, image data captured and compressed and encoded by a camera module is required to correspond to a plurality of output image sizes. As output image sizes, 640 (pixel) × 480 (line) pixel VGA (Video Graphics Array), 320 × 240 pixel QVGA (Quarter Video Graphics Array), 352 × 288 pixel CIF (Common Intermediate Format), 176 There are QCIF (Quarter Common Intermediate Format) of 144 pixels. Conventionally, since this difference in output image size is not dealt with, there is a problem that it is necessary to have code amount table data for SCF values or parameters for each output image size. In such a case, there is a problem that the required memory capacity such as ROM (Read Only Memory) and RAM (Random Access Memory) increases as the types of output image sizes increase.
[0010]
Further, in the conventional method, since the SCF value is controlled by a microcomputer on the upper side (hereinafter abbreviated as a microcomputer) of the camera module, there is a problem that a processing load is imposed on the upper microcomputer.
[0011]
In the conventional method, since the code amount is uniquely determined for the SCF value, it is impossible to adjust the code amount according to the code amount error.
The present invention has been made in view of these points, and an object of the present invention is to provide an imaging apparatus and an imaging method capable of adjusting the code amount with a small memory capacity and without placing a processing burden on a microcomputer.
[0012]
[Means for Solving the Problems]
In the present invention, in order to solve the above-described problem, in an imaging apparatus that captures and compresses and encodes image data of a subject, an SCF value multiplied by a basic quantization table is changed for the predetermined image data, and the SCF value is changed. A code amount change rate calculating unit that calculates a change rate of the image data according to the change of the image data, and when the code amount of the other image data is out of a predetermined range, the change rate is referred to, and the code And an SCF value determining unit that determines the SCF value so that the amount is within the predetermined range.
[0013]
According to the above configuration, when the code amount of the image data is outside the predetermined range, the SCF value determining unit refers to the change rate calculated by the code amount change rate calculating unit, and determines the code amount to be predetermined. The SCF value is determined so as to be within the range.
[0014]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a schematic functional block diagram of an imaging apparatus according to an embodiment of the present invention.
[0015]
Here, an imaging apparatus that performs compression encoding in the JPEG format is shown.
The imaging device 10 includes an imaging unit 11, a digitizing unit 12, a DCT conversion unit 13, a quantization unit 14, and an entropy encoding unit 15, and further, when the quantization unit 14 performs quantization. It comprises a functioning code amount change rate calculating unit 14a and an SCF value determining unit 14b.
[0016]
The imaging unit 11 captures subject image data.
The digitizing unit 12 converts the captured image data into digital data.
The DCT conversion unit 13 performs discrete cosine conversion to convert image data into frequency components.
[0017]
The quantization unit 14 quantizes the DCT coefficient using the quantization table.
The entropy encoding unit 15 replaces a sequence of frequently occurring data with a short code and replaces a sequence of less frequently occurring data with a long code by Huffman coding or arithmetic coding.
[0018]
The code amount change rate calculation unit 14a changes the SCF value multiplied by the basic quantization table, and calculates the change rate of the code amount of the image data according to the change of the SCF value. The change ratio is calculated based on the code amount (reference code amount) when the SCF value is a reference. The code amount change rate includes a code amount decrease rate (hereinafter referred to as a code decrease rate) or an increase rate (hereinafter referred to as a code increase rate). The reference code amount will be described later.
[0019]
The SCF value determining unit 14b refers to the change ratio when the code amount of the image data is outside the predetermined range defined by the limited code amount, and adjusts the code amount within the predetermined range. Is determined in advance.
[0020]
Here, the reason for calculating and using the change rate of the code amount in the imaging apparatus 10 according to the embodiment of the present invention will be described.
FIG. 2 is a diagram illustrating the relationship between the SCF value and the code amount of image data in JPEG format.
[0021]
The horizontal axis is the SCF value, the vertical axis is the code amount, and both are logarithmic axes. Here, it is assumed that the SCF values of the Y (luminance) component and the C (color difference) component are the same. 3 to 9 described below are similarly logarithmic axes, and the SCF values of the Y (luminance) component and the C (color difference) component are the same.
[0022]
As shown in this figure, it is understood that the code amount increases when the SCF value is decreased, and the code amount decreases when the SCF value is increased.
The sign reduction rate is expressed by the following equation.
[0023]
[Expression 2]
RED_RATE (X) = TOTAL (BaseSCF) / TOTAL (X) (2)
In the above equation, the code reduction rate “RED_RATE (X)” when the SCF value is “X” is the code amount “TOTAL (BaseSCF)” of the reference SCF value “BaseSCF” (this is the reference code amount) Is divided by the code amount “TOTAL (X)” when the SCF value is “X”.
[0024]
In equation (2), if “TOTAL (X) / TOTAL (BaseSCF)” is used, the sign increase rate can be obtained.
When the code amount shown in FIG. 2 is converted into the code reduction ratio according to the equation (2), the following figure is obtained.
[0025]
FIG. 3 is a diagram illustrating the relationship between the SCF value and the code reduction rate.
The horizontal axis is the SCF value, and the vertical axis is the sign reduction rate.
However, here, the reference SCF value “BaseSCF = 1.0” is executed and Expression (2) is executed. As shown in the figure, the code reduction ratio is a relative code amount when the code amount when the SCF value is 1.0 is set as the reference code amount.
[0026]
FIG. 4 is a diagram showing the relationship between the code amount of image data and the SCF value when the same image is taken with three different image sizes.
It is assumed that the basic quantization table used for quantization is the same.
[0027]
Here, the change by the SCF value of the code amount when the same image is taken with three different image sizes of VGA, QVGA, and QCIF is shown. The VGA with the largest image size has the largest code amount, and the code amount of QCIF with the smallest image size is the smallest. As shown in the figure, it can be seen that even if the SCF values are the same, the code amount varies greatly depending on the output image size.
[0028]
Next, when this code amount is converted into a code reduction rate according to the equation (2), the following figure is obtained.
FIG. 5 is a diagram showing the relationship between the code reduction ratio of image data and the SCF value when the same image is taken with three different image sizes.
[0029]
As shown in the figure, it can be seen that the code reduction rate is almost the same regardless of the output image size. This is because the method of changing the code amount with respect to the change of the SCF value is similar regardless of the output image size.
[0030]
Next, consider the case where the output image size is the same and the code amount is extremely different due to the redundancy of the image.
FIG. 6 is a diagram illustrating a change in the code amount with respect to the SCF value of an image having the same image size but extremely different code amount.
[0031]
As shown in the figure, it can be seen that even if the SCF values are the same, the code amounts are different.
Next, when this code amount is converted into a code reduction rate according to the equation (2), the following figure is obtained.
[0032]
FIG. 7 is a diagram illustrating a change in the code reduction ratio with respect to the SCF values of images having different code amounts.
As shown in the figure, it can be seen that similar characteristics can be obtained by converting the images having different code amounts into the code reduction ratio. This is because, even in images with different code amounts, the method of changing the code amount with respect to the change of the SCF value is similar.
[0033]
Next, a case where the basic quantization table is different will be described.
FIG. 8 is a diagram illustrating the relationship between the SCF value and the code amount for three types of basic quantization tables.
[0034]
As shown in the figure, when the basic quantization table is different, the method of changing the code amount is also different according to the change of the SCF value. When this is converted into a code reduction ratio in accordance with equation (2), the following diagram is obtained.
[0035]
FIG. 9 is a diagram illustrating the relationship between the SCF value and the code reduction rate for three types of basic quantization tables.
As shown in the figure, it can be seen that when the basic quantization table is different, the code reduction rate is different.
[0036]
From the above considerations, it was found that if the basic quantization table is the same, the code reduction rate is similar regardless of the output image size and the difference in code amount. It goes without saying that the same applies to the sign increase rate.
[0037]
For this reason, the imaging apparatus 10 according to the embodiment of the present invention calculates the change rate of the code amount with respect to the SCF value, and has, for example, table data, which is used to determine the SCF value at the time of quantization. .
[0038]
Hereinafter, the operation of the imaging apparatus 10 will be described.
When the subject is captured by the imaging unit 11, the image data captured by the digitizing unit 12 is converted into digital data. The image data is subjected to discrete cosine transform in the DCT transform unit 13 and transformed into frequency components.
[0039]
The image data converted into the frequency component is quantized by the quantization unit 14. In the code amount change ratio calculation unit 14a, for example, when a certain standard image is copied before factory shipment, the SCF value to be multiplied by the basic quantization table is changed in a predetermined step, and the frame of the image data at that time is changed. From the code amount change, the code amount change ratio with respect to the SCF value is calculated. The calculated change rate of the code amount is managed as a table (a code decrease rate table or a code increase rate table), for example.
[0040]
The predetermined step “STEP” is represented by a ratio “TBL_SCF [x + 1] / TBL_SCF [x]” of the SCF values on the adjacent table, and in the range of 1/256 to 256 that can take SCF values, for example, When the number of elements “NUMPOINT” is taken at equal intervals on the logarithmic axis, it is expressed by the following equation.
[0041]
[Equation 3]

[0042]
FIG. 10 is an example of a code reduction ratio table.
Here, “TBL_SCF” indicates an SCF value on the table, and “TBL_RED_RATE” indicates a code reduction rate on the table. As shown in FIG. 10, the code reduction rate table is created by creating a code reduction rate table corresponding to the SCF value in the step calculated by Equation (3) with SCF value = 1.0 in the middle. Is done.
[0043]
Next, the SCF value determination unit 14b detects the code amount of the image data after compression encoding processing, and when the code amount is outside the range defined by the limit code amount, the code amount change rate (for example, The SCF value that adjusts the code amount within the range of the limited code amount is determined with reference to the code reduction ratio table as shown in FIG. A specific adjustment method will be described later.
[0044]
The quantization unit 14 creates a quantization table by multiplying the determined SCF value by the basic quantization table as shown in Expression (1), and performs quantization using the quantization table.
[0045]
FIG. 11 is an example of a basic quantization table for the Y (luminance) component.
FIG. 12 is an example of a basic quantization table for C (color difference) components.
As shown in FIGS. 11 and 12, the high frequency component (lower right part of the figure) has a feature that is inconspicuous, and the low frequency component (upper left part of the figure) has a feature that is noticeable. Therefore, the quantization step is increased in order to reduce the quantization step and increase the omission of high frequency components.
[0046]
Since human vision has a characteristic that the sensitivity of the color difference component is lower than that of the luminance component, the omission of the color difference component is increased as in the basic quantization table for the C component in FIG. Compared to less.
[0047]
The quantized image data is subjected to Huffman coding or arithmetic coding by the entropy coding unit 15 and output as JPEG data.
As described above, the code amount can be adjusted using the same code amount change ratio (for example, the code decrease ratio table in FIG. 10) for images having different image sizes and significantly different code amounts. Code amount prediction and code amount adjustment are possible with a small memory capacity.
[0048]
Next, the imaging device 10 according to the embodiment of the present invention will be specifically described.
FIG. 13 is a hardware configuration diagram of the imaging apparatus system.
As illustrated, the imaging apparatus system includes a camera module 20 corresponding to the imaging apparatus 10 illustrated in FIG. 1 and a display control block 30.
[0049]
The camera module 20 includes a lens 21, a sensor 22, a CDS (correlated double sampling) / AGC (automatic gain control) 23, a camera signal processing block 24, a JPEG encoder 25, a microcomputer 26, an EEPROM (Electrically Erasable Programmable Read-Only Memory). 27.
[0050]
Here, the lens 21 performs focusing and image formation on the subject.
The sensor 22 converts light into an electrical signal. In addition, the charge accumulation time varies depending on the electronic shutter. The sensor 22 is, for example, a CCD (Charge Coupled Device) sensor.
[0051]
The CDS / AGC 23 performs noise removal and automatic gain control by a correlated double sampling circuit.
The camera signal processing block 24 includes an A / D (analog / digital) converter, and converts an analog signal input from the sensor 22 into a digital signal. Further, the input signal is converted into a video signal (YUV signal in the figure) and output. The “SYNC signal” is a synchronization signal, which is input from the camera signal processing block 24 to the JPEG encoder 25 and the display control block 30 to synchronize each block in processing. The YUV signal is a video signal in a format in which colors are represented by three pieces of information: a Y component, a difference (U) between the Y component and the red component (Cr), and a difference (V) between the Y component and the blue component (Cb). .
[0052]
The JPEG encoder 25 encodes (compresses) the output from the camera signal processing block 24 into the JPEG format.
Although not shown, the microcomputer 26 includes a ROM that stores a program for controlling the image pickup apparatus system of FIG. 13 and a RAM that temporarily stores data necessary for executing the program. To control.
[0053]
The EEPROM 27 stores the basic quantization table and the change rate of the code amount.
The display control block 30 performs LCD (Liquid Crystal Display) output, image expansion, and the like. In addition, a command from a higher-level control block (microcomputer) (not shown) is notified to the microcomputer 26 by IIC (Inter Integrated Circuit) communication or the like.
[0054]
In FIG. 13, the lens 21 and the sensor 22 correspond to the imaging unit 11 shown in FIG. The camera signal processing block 24 (A / D converter) corresponds to the digitizing unit 12. Moreover, the functions of the DCT conversion unit 13, the quantization unit 14, the code amount change rate calculation unit 14a, the SCF value determination unit 14b, and the entropy encoding unit 15 can be realized by the functions of the JPEG encoder 25 and the microcomputer 26.
[0055]
In such an imaging apparatus system, when a subject is condensed by the lens 21 and imaged on the sensor 22, the sensor 22 converts light into an electrical signal. Further, the image data that has become an electrical signal is subjected to noise removal and automatic gain control by the CDS / AGC 23, and then input to the camera signal processing block 24, converted into a YUV signal, and output to the display control block 30. . For example, in a digital camera or the like, when the shutter is not pressed, depending on the display mode, the video signal may be processed by the display control block 30 so that the subject is always displayed on a liquid crystal monitor (not shown). . At this time, for example, the subject is displayed at 30 frames / second.
[0056]
When the shutter is pressed, the video signal of the frame at that time is processed by the JPEG encoder 25 for each block with 64 pixels of 8 × 8 as one block, and is compressed and encoded into the JPEG format.
[0057]
Here, details of the JPEG encoder 25 in the embodiment of the present invention will be described.
FIG. 14 is a diagram showing a configuration used for the JPEG encoder and its encoding process.
[0058]
As illustrated, the JPEG encoder 25 includes a JPEG encode block 25-1 and a code amount register 25-2.
The JPEG encoding block 25-1 includes a Y component basic quantization table register 25-1a, a C component basic quantization table register 25-1b, a Y component SCF register 25-1c, and a C component SCF register 25-. 1d.
[0059]
The Y component basic quantization table register 25-1a temporarily stores the Y component basic quantization table extracted from the EEPROM 27 at the time of quantization.
The C component basic quantization table register 25-1b temporarily stores the C component basic quantization table extracted from the EEPROM 27 at the time of quantization.
[0060]
The Y component SCF register 25-1c temporarily stores the S component value for the Y component extracted from the EEPROM 27 at the time of quantization.
The C component SCF register 25-1d temporarily stores the C component SCF value extracted from the EEPROM 27 during quantization.
[0061]
The code amount register 25-2 counts the code amount of the JPEG format image data output from the JPEG encode block 25-1. For example, a configuration like a counter may be used.
[0062]
In such a JPEG encoder 25, a DCT conversion process is performed for each 8 × 8 pixel block under the control of the microcomputer 26, and then a quantization process is performed.
[0063]
At the time of quantization processing, the basic quantization table for the Y component is stored in the basic quantization table register 25-1a for the Y component, and the basic quantization table for the C component is stored in the basic quantization table register 25-1b for the C component. The Y component SCF register 25-1c is extracted from the EEPROM 27 and stored in the C component SCF register 25-1d.
[0064]
The image data is quantized using these extracted data.
That is, the SCF value is multiplied by each element value of the basic quantization table to create a quantization table, and the 8 × 8 DCT coefficient obtained as a result of the DCT conversion process is divided by each element value of the quantization table. . After quantization and entropy encoding processing in this way, the code amount of the obtained image data in JPEG format is counted by the code amount register 25-2. As a result, if the output code amount is within the limit code amount from the upper limit code amount to the lower limit code amount, the output is output from the JPEG encoder 25 as it is.
[0065]
The upper limit code amount and the lower limit code amount may be set in the JPEG encode block 25-1, and a limit code amount error may be notified to the microcomputer 26, or an externally designated value may be stored in the RAM inside the microcomputer 26. The microcomputer 26 may detect the limited code amount error as compared with the output code amount.
[0066]
Hereinafter, a quantization process when the output code amount is not within the limit code amount will be described.
In the embodiment of the present invention, for example, a standard image before factory shipment is copied, and under the control of the microcomputer 26, the SCF value to be multiplied by the basic quantization table used for quantization is changed in a predetermined step. The change rate of the code amount of the image data corresponding to the change of the SCF value is calculated and managed as a code reduction rate table as shown in FIG. 10, for example. Such a code reduction ratio table is stored in the EEPROM 27.
[0067]
When the code amount of the image data is outside the range of the limit code amount, the microcomputer 26 refers to the change rate of the code amount stored in the EEPROM 27 developed in the RAM at the time of initialization processing, and performs the following processing I do.
[0068]
First, a target code amount (hereinafter referred to as a target code amount) is determined.
Examples of the target code amount include the following.
1. Code amount intermediate between upper limit code amount and lower limit code amount
2. When the upper limit code amount error occurs, the code amount is smaller than the upper limit code amount, and when the lower limit code amount error occurs, the code amount is larger than the lower limit code amount.
3. A value slightly smaller than the upper limit code amount in both errors.
[0069]
Thereafter, a ratio (hereinafter referred to as an absolute target code amount ratio) for obtaining the target code amount is obtained from the code amount at the time of the code amount error. The absolute target code amount ratio is calculated by dividing the error code amount by the target code amount.
[0070]
Further, the microcomputer 26 multiplies the change rate of the code amount of the SCF value used at the time of the error by the absolute target code amount rate to calculate a relative target code amount rate (hereinafter referred to as a relative target code amount rate). Using the calculated relative target code amount ratio, an SCF value used for quantization of the next frame is determined with reference to a code reduction ratio table as shown in FIG. 10, for example.
[0071]
Thereafter, in the JPEG encoding block 25-1, the SCF value for the Y component is stored in the SCF register for the Y component 25-1c, and the SCF value for the C component is stored in the SCF register for the C component 25-1d. As in 1), a basic quantization table is multiplied to create a quantization table, and quantization is performed using the quantization table.
[0072]
The automatic adjustment process of the SCF value for the second example of the target code amount described above by the microcomputer 26 is summarized in a flowchart.
FIG. 15 is a flowchart showing the flow of the automatic adjustment process of the SCF value.
[0073]
In the figure, the processing contents in the case of the upper limit code amount error are mainly shown.
In addition, here, a case where a table in which the code decrease rate is tabulated as the code amount change rate is used, and “TBL_RED_RATE [CNT]” indicates the code on the code decrease rate table corresponding to the counter value CNT. Indicates the rate of decrease. “TBL_SCF [CNT]” indicates the SCF value on the code reduction rate table corresponding to the counter value CNT. The counter value CNT indicates the position on the code reduction rate table. For example, in the code reduction rate table as shown in FIG. 10, the SCF value is minimum (in the example of FIG. 10, TBL_SCF = 0.039). Value), the counter value CNT = 0, and when the SCF value is maximum (in the example of FIG. 10, TBL_SCF = 256), the counter value CNT is also maximum.
[0074]
In the automatic adjustment process of the SCF value, first, a counter update process is performed. Here, the counter value PAST_CNT of the SCF value used for the frame at the time of the code amount error is set to the counter value CNT (step S10).
[0075]
Next, the relative target code amount ratio is calculated. The relative target code amount ratio is calculated by multiplying the absolute target code amount (output code amount / restricted code amount) by the code decrease rate TBL_RED_RATE [CNT] in the current counter value CNT. Here, the limited code amount is the upper limit code amount at the time of the upper limit code amount error, and is the lower limit code amount at the time of the lower limit code amount error (step 11).
[0076]
Next, the counter value CNT is incremented by 1 when the upper limit code amount error occurs, and is decremented by 1 when the lower limit code amount error occurs (step S12).
Next, when the upper limit code amount error occurs, it is determined whether or not the code decrease rate TBL_RED_RATE [CNT] exceeds the relative target code amount rate calculated in step S11. When it exceeds, it progresses to the process of step S15, and when that is not right, it progresses to step S14. On the other hand, when the lower limit code amount error occurs, it is determined whether or not the code decrease rate TBL_RED_RATE [CNT] is below the relative target code amount rate calculated in step S11. If it is lower, the process proceeds to step S15, and if not, the process proceeds to step S14 (step S13).
[0077]
In the process of step S14, it is determined whether or not the counter value CNT is the maximum value when the upper limit code amount error occurs. If the counter value CNT is the maximum value, the process proceeds to step S15. If not, the process from step S12 is repeated. On the other hand, when the lower limit code amount error occurs, it is determined whether or not the counter value CNT is the minimum value. If the counter value CNT is the minimum value, the process proceeds to step S15. If not, the process from step S12 is repeated.
[0078]
In the process of step S15, the SCF register (Y component SCF register 25-1c and C component SCF register 25-1d in FIG. 14) and the SCF counter value PAST_CNT used for quantization are updated. That is, the SCF value TBL_SCF [CNT] on the code amount change rate table is set in the SCF register by the counter value CNT determined by the processing up to step S14, and the SCF value counter value PAST_CNT used for quantization is set in the SCF value. The counter value CNT is set. Thus, the SCF value automatic adjustment processing is completed.
[0079]
In this way, by adjusting the SCF value using the same code amount change ratio data for images of different image sizes and code amounts, it is possible to predict the code amount and adjust the code amount with a small memory capacity. .
[0080]
Furthermore, since the file size control performed in the upper control block of the conventional camera module 20 is performed in the camera module 20, it is not necessary to put extra processing in the upper control block. Therefore, the ROM and RAM capacities in the upper control block can be reduced and the processing load can be reduced.
[0081]
In the above description, the change rate of the code amount is calculated under the control of the microcomputer 26. However, the change rate of the code amount is written to the EEPROM 27 from the outside as table data as shown in FIG. It may be.
[0082]
In addition, since the change rate of the code amount should be different depending on the basic quantization table as described above, if the basic quantization table is changed before shipment from the factory, the changed basic quantum table is used. The code amount change rate corresponding to the conversion table can be calculated and adjusted again by the above-described method, or can be dealt with by giving it to the EEPROM 27 from the outside.
[0083]
In the above description, the controllable SCF value is set to 1/256 to 256. However, since the value of the quantization table element is 1 to 255, the SCF value is determined from the maximum element and the minimum element of the basic quantization table. The necessary control range may be obtained.
[0084]
For example, the control range is determined by the following equation: SCF value maximum control = 255 / basic quantization table minimum value, SCF value control minimum value = 1 / basic quantization table maximum value. By limiting the control range of the SCF value in this way, it becomes possible to control at high speed.
[0085]
【The invention's effect】
As described above, in the present invention, the SCF value multiplied by the basic quantization table is changed, the change rate of the code amount of the image data corresponding to the change of the SCF value is calculated, and the code amount of the image data is set to a predetermined value. If it is out of the range, the SCF value that determines the code amount within the predetermined range is determined with reference to the change rate, so the change rate of the same code amount is used regardless of the output image size. The code amount can be adjusted.
[0086]
As a result, the code amount can be adjusted with a small memory capacity.
[Brief description of the drawings]
FIG. 1 is a schematic functional block diagram of an imaging apparatus according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating a relationship between an SCF value and a code amount of image data in JPEG format.
FIG. 3 is a diagram illustrating a relationship between an SCF value and a code reduction rate.
FIG. 4 is a diagram illustrating a relationship between a code amount of image data and an SCF value when the same image is taken with three different image sizes.
FIG. 5 is a diagram illustrating a relationship between a code reduction ratio of image data and an SCF value when the same image is taken with three different image sizes.
FIG. 6 is a diagram illustrating changes in code amount with respect to SCF values of images having the same image size but extremely different code amounts.
FIG. 7 is a diagram illustrating a change in a code reduction rate with respect to SCF values of images having different code amounts.
FIG. 8 is a diagram illustrating the relationship between SCF values and code amounts for three types of basic quantization tables.
FIG. 9 is a diagram illustrating a relationship between an SCF value and a code reduction rate for three types of basic quantization tables.
FIG. 10 is an example of a code reduction rate table.
FIG. 11 is an example of a basic quantization table for a Y (luminance) component.
FIG. 12 is an example of a basic quantization table for a C (color difference) component.
FIG. 13 is a hardware configuration diagram of the imaging apparatus system.
FIG. 14 is a diagram showing a configuration used for a JPEG encoder and its encoding process.
FIG. 15 is a flowchart showing a flow of automatic adjustment processing of an SCF value.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 10 ... Imaging device, 11 ... Imaging part, 12 ... Digitization part, 13 ... DCT conversion part, 14 ... Quantization part, 14a ... Code amount change rate calculation part, 14b ... SCF value determination part , 15 ... Entropy encoding unit

Claims (5)

In an imaging device that captures and encodes image data of a subject,
A code amount change rate calculating unit that calculates a change rate of a code amount of the image data according to a change in the scaling factor by changing a scaling factor to be multiplied to a basic quantization table for the predetermined image data;
A scaling factor determining unit that determines the scaling factor so that the code amount is within the predetermined range with reference to the change ratio when the code amount of the other image data is outside the predetermined range. When,
An imaging device comprising: The imaging apparatus according to claim 1, wherein the code amount change rate calculation unit adjusts the change rate according to a type of the basic quantization table used. The imaging apparatus according to claim 1, wherein the code amount change rate calculation unit limits a range of change of the scaling factor according to a maximum element value or a minimum element value of the basic quantization table used. . A storage unit that stores the change rate calculated by the code amount change rate calculation unit;
The imaging apparatus according to claim 1, wherein the scaling factor determination unit determines the scaling factor with reference to the change ratio stored in the storage unit. In an imaging method that captures and encodes image data of a subject,
For the predetermined image data, change the scaling factor to be multiplied to the basic quantization table, calculate the change rate of the code amount of the image data according to the change of the scaling factor,
When the code amount of the other image data is outside a predetermined range, the scaling factor is determined with reference to the change rate and the code amount within the predetermined range.
An imaging method characterized by the above.

Images