JP4415446B2 - Colorimetric image conversion method, colorimetric image conversion apparatus, and computer-readable information recording medium recording a colorimetric image conversion program - Google Patents

Description

【００５０】
（ここで ｉ：チャンネル（Ｒ，Ｇ，Ｂ），ｊ：色票番号）
へ変換する。
【００５１】
（フローチャートのｓ１４：露光条件の標準化）
標準板を撮影した際にデジタルカメラで設定した露光条件であるシャッタースピード（ｔ）とその絞り値（ｆ）とを、画像データが記録されたファイルから読み込み、ｎ個の線形階調値データをこれらの露光条件によりシャッタースピード１あたりかつ絞り値１あたりの階調値に標準化し、応答値データ
【００５２】
【数２】

【００５３】
とし、これらをｎ次元の１列の応答値ベクトル
【００５４】
【数３】

【００５５】
（フローチャートのｓ１５：説明変数の行列作成）
あらかじめ記録してある、デジタルカメラ２の入射光の分光エネルギー（例えば）波長域３８０ｎｍ〜７８０ｎｍの範囲で５ｎｍ間隔に対する応答値の関係を記述した３チャンネルそれぞれの分光感度特性データ；
ｃ_i （λ）
と、
標準板１上の全色票の分光反射率をやはり３８０ｎｍ〜７８０ｎｍの範囲で５ｎｍ間隔で記述した２４個の分光反射率データ
ｓ_j （λ）
と、
照明光の分光分布をｄ次元で表現するためのｄ個の基底関数；
ｂ_k （λ）
とから、
ｎ行×ｄ列の説明変数行列；
【００５６】
【数４】

【００５７】
を次のように作成する。
【００５８】
【数５】

【００５９】
【数６】

【００６０】
ここで、
λ： 波長（λ１，λ２，λ３， ・・・，λ８１）
Ｂ： 各列が基底関数をｂ_k （λ）を示す基底関数ベクトル（８１行×ｄ列）
【００６１】
（フローチャートのｓ１６：重回帰分析）
照明の分光分布をｄ個の基底関数；
ｂ_k （λ）
の線形結合で表すためのｄ個の重み係数から成る重みベクトルを
【００６２】
【数７】

【００６３】
とすると、応答値ベクトル；
【００６４】
【数８】

【００６５】
は、
【００６６】
【数９】

【００６７】
と表すことができることから、説明変数；
【００６８】
【数１０】

【００６９】
に対して、応答値ベクトル；
【００７０】
【数１１】

【００７１】
を目的変数とした線形重回帰分析を施し、得られた偏回帰係数を重みベクトル
【００７２】
【数１２】

【００７３】
とする。
但し、照明の分光分布を基底関数の線形結合で表すための次数ｄは、適正応答範囲内の応答値の数ｎより小さくなければならない。
【００７４】
（フローチャートのｓ１７：線形変換）
ｄ個の基底関数を収めた基底関数行列；
【００７５】
【数１３】

【００７６】
とｄ次元の重みベクトル
【００７７】
【数１４】

【００７８】
とにより、
【００７９】
【数１５】

【００８０】
を計算し、未知であった照明光の分光分布データを、３８０ｎｍ〜７８０ｎｍの範囲の５ｎｍ刻みで合計８１個の分光分布の推定値を収めたベクトル
【００８１】
【数１６】

【００８２】
として得る。
【００８３】
２．３ 光源分析処理についての詳細
以下、本発明に関わる推定方法あるいは一例における、光源分析処理について、図４のフローチャートに基づいて説明を行う。
【００８４】
あらかじめ、撮影に用いたデジタルカメラの色再現特性を様々な照明光各々に対して把握しておき、その把握した色再現特性に関する情報を記述した色再現情報ファイルを作成する。この色再現情報ファイルとしては、例えばＩＣＣプロファイルを適用する。
この色再現情報ファイルに、どのような照明光かを識別できる文字列をファイル名として付けておき、これを指標とする。また、対応する照明光の情報を色再現情報ファイルに記述しておく。作成した複数の色再現情報ファイルは、ファイル群として計算機内に搭載しておく。
【００８５】
（フローチャートのｓ１８：分光特性推定値入力）
解析処理で出力された照明光の分光特性推定値を計算機３に入力する。
【００８６】
（フローチャートのｓ１９：照明光の種類の特定）
入力された照明光の分光特性推定値と、作成しておいた色再現情報ファイルに記述してある照明光の情報を用いて、照明光の種類を特定する。
【００８７】
ここで、照明光の分光特性推定値と複数の照明光の情報から照明光の種類を特定する方法として、以下の方法を挙げる。
【００８８】
（１）解析処理で出力された分光分布データより得られる分光分布波形を用いて照明光の種類を割り出す方法
解析処理により３８０ｎｍ〜７８０ｎｍの間において５ｎｍ刻みで出力される分光分布データを、

として、両者の比較を行う。
各波長における差の絶対値を全波長域で平均した値が最も小さくなる分光分布データを求め、そのデータに対応する照明光を求める照明光として決定する。
【００８９】
（２）解析処理で提示した、ｄ次元の重みベクトルから照明光の種類を割り出 す方
分光分布の推定値は、解析処理で提示したｄ次元の重みベクトル
【００９０】
【数１７】

【００９１】
と、
基底関数行列
【００９２】
【数１８】

【００９３】
から導出される値であるので、
ｄ次元の重みベクトル
【００９４】
【数１９】

【００９５】
から、どのような照明光であるかを特定することもできる。
そして、解析処理で提示したｄ次元の重みベクトルを
【００９６】
【数２０】

【００９７】
、また同様に、選択対象となる複数の照明光におけるｄ次元の重みベクトルを
【００９８】
【数２１】

【００９９】
とする。尚、
【０１００】
【数２２】

【０１０１】
については、それぞれに対応する照明光の分光分布を、あらかじめ
【０１０２】
【数２３】

【０１０３】
の形式で近似を行っておく。
【０１０４】
【数２４】

【０１０５】
と、
【０１０６】
【数２５】

【０１０７】
それぞれとの比較を行い、最も近しいベクトルを特定し、それに対応する照明光を求める明光として決定する。
【０１０８】
（フローチャートのｓ２０：照合処理）
照明光の種類が特定されたことで、ファイル群内の各ファイルのファイル名と計算機３上で照合することにより、ファイル群の中から撮影時の照明光に対応した色再現情報ファイルを自動的に選択する。
【０１０９】
（フローチャートのｓ２１：対応ファイルの出力）
選択された色再現情報ファイルを、撮影時の照明光に対応した色再現情報ファイルとして出力する。
【０１１０】
＜３ 変形例＞
（ａ）撮像手段として用いる媒体は、デジタルスチルカメラでなくとも、デジタルビデオカメラやスキャナなど、画像データを取得できる技術であるならばどのような媒体でも構わない。
ただしスキャナの場合は単体でなく、フィルム撮影用カメラと組にして扱うことにする。このとき、撮像手段の色再現情報ファイルは、スキャナとカメラ各々に対して用意する必要がある。
（ｂ）設定手段は、撮像手段と一体である必要性はなく、撮像手段の部分、例えば計算機に装備されていることも好ましい。
【０１１１】
（ｃ）撮像手段の色再現情報ファイルの形態としては、ＩＣＣプロファイルを適用するしているが、それに限らず、それ以外の自由な形態で作成されたデータを適用できることも好ましい。
（ｄ）光源分析手段において、計算機を用いずとも、独立した専用の装置を設けたり、また撮像手段として用いた媒体中に選択を行う処理系を組み込んでおくことも好ましい。
【０１１２】
（ｅ）光源分析手段において、照明光の種類を特定する際、色再現情報ファイルそのものに対応する照明光についての情報を記述しておくに限らず、対応する照明光についての情報のみを集めた対応テーブルを別途に用意し、それを用いて照明光の種類を特定を行うことも好ましい。
（ｆ）光源分析処理により選択された色再現情報ファイルを用いてデジタルカメラで撮影した画像をＣＩＥＸＹＺ画像データに変換するが、このときＣＩＥＸＹＺ画像データに限らず、ＣＩＥＬＡＢ画像データに変換するのも好ましい。
【０１１３】
（ｇ）光源分析処理により選択された色再現情報ファイルを用いてデジタルカメラで撮影した画像をＣＩＥＸＹＺ画像データに変換する。このＣＩＥＸＹＺ画像データを、表示系または出力系のプロファイルを用いて表示・出力用の画像データへ変換するのだが、この表示用または出力用の画像データ変換だけでなく、画像解析や画像の分類など、ＣＩＥＸＹＺ画像データを応用して様々な作業を行うことも好ましい。
（ｈ）照明光の分光特性推定値から、光源分析処理を行うにあたり、指標として用いるのは、分光分布データに限らず、例えば色温度といった分光分布データから算出される値を用いるのも好ましい。
【０１１４】
【発明の効果】
以上のように、本発明によれば、まず撮影に用いた撮像系の色再現特性を、様々な照明光各々に対して把握しておき、その把握した色再現特性に関する情報を記述したファイル（色再現情報ファイル）をあらかじめ作成しておく。この複数の色再現情報ファイルをファイル群として計算機に搭載しておき、撮像系で撮影した際の照明光の分光特性推定値とファイル群から撮影時の照明光の種類を特定し、照合を行うことで、撮影時の照明光に対応した色再現情報ファイルを自動的に選択することが可能となる。この選択された色再現情報ファイルを用いて、撮影時と同一の照明光下で同一の入力媒体を用いて撮影した画像をＣＩＥＸＹＺ画像データに変換する。このＣＩＥＸＹＺ画像データをモニタで表示したり印刷物を出力したりする際、表示デバイスまたは出力デバイスの色再現特性に関する情報を考慮して、表示用または出力用の画像データへの変換処理ができ、照明光の影響を考慮した画像の色変換が容易に行うことができる。
【図面の簡単な説明】
【図１】本発明に関わる測色画像変換方法、測色画像変換装置、並びに測色画像変換プログラムの一例について、その構成の一部を模式的に示す説明図。（標準板を撮像装置で撮影する様子）
【図２】本発明に関わる測色画像変換方法、測色画像変換装置、並びに測色画像変換プログラムの一例について、その全体の動作フローを説明するフローチャート。
【図３】本発明に関わる測色画像変換方法、測色画像変換装置、並びに測色画像変換プログラムにおいて、解析手段が行う処理例について説明するフローチャート。
【図４】本発明に関わる測色画像変換方法、測色画像変換装置、並びに測色画像変換プログラムにおいて、光源分析手段が行う処理例について説明するフローチャート。
【符号の説明】
１・・・ 標準板
２・・・ 撮像手段（代表例はデジタルカメラ）
３・・・ 計算機[0001]
BACKGROUND OF THE INVENTION
Information that grasps the color reproduction characteristics of the imaging system corresponding to the illumination light used at the time of shooting can be automatically specified, and this information is used to convert the image data captured by the imaging system. Thus, it is possible to obtain basic image data that is useful when it is necessary to obtain display or output image data itself, or to obtain various display or output image data. In particular, by obtaining CIEXYZ image data by the conversion process, it is possible to perform a process of converting CIEXYZ image data into image data for display or output using information on the color reproduction characteristics of the display device or output device. . Also, various operations can be performed by applying CIEXYZ image data such as converted image analysis and image classification into image data for display or output.
The CIEXYZ image data referred to in this specification means image data expressed by a CIEXYZ display system.
[0002]
[Prior art]
If you try to reproduce the appearance of a color that is recognized by human vision when an image is taken of a scene with a digital camera, the effect of the illumination light at the time of shooting will be taken into account. Color conversion processing must be performed. In order to perform this color conversion process, information indicating the illumination light at the time of photographing is required. The following technologies are involved in obtaining information about the illumination light, but have problems.
[0003]
Measurement equipment and sensors that can easily obtain illumination characteristics are provided, and some products are integrated with the image processing equipment of the imaging system or display system, or the imaging system itself has the function. There are also things. However, low-dimensional measurement values such as luminance and illuminance were obtained from this simple measuring device and sensor. Therefore, it is insufficient as information on illumination light.
[0004]
Thus, consider obtaining spectral characteristics as illumination light information. Below are some of the technologies involved in obtaining spectral characteristics, but they also have problems.
[0005]
There is a measuring device called a “spectral radiometer” as a device used when measuring the spectral characteristics of illumination light, but there are various problems such as being expensive and heavy, and taking a long time for measurement.
[0006]
There is also an apparatus that corrects the color of an image obtained by photographing a real scene with a digital camera in consideration of the influence of a light source. However, in order to correct such a color, as described above, information indicating what kind of illumination light the scene was photographed is necessary. This information needs to be instructed by a person who performs correction to the apparatus, and there is no means for automatically determining the type of illumination light from an image with high accuracy.
[0007]
[Problems to be solved by the invention]
As described above, there is no conventional technique for automatically obtaining information indicating what kind of illumination light was taken when imaged with an imaging system. Therefore, the spectral characteristics of the illumination light are estimated, and the obtained spectral characteristic estimation value is used to grasp the information of the color reproduction characteristics for the illumination light at the time of shooting. By using this information, the same illumination is obtained. Below, it would be very convenient if image data of a landscape or subject photographed with the same digital camera can be converted to another image data while maintaining calorimetric information. The present invention uses such an estimated value of spectral characteristics of illumination light to grasp information on color reproduction characteristics with respect to illumination light at the time of photographing, and landscapes and subjects photographed with the same digital camera under the same illumination It is an object of the present invention to make it possible to convert the image data into another image data while maintaining the colorimetric information. In particular, the other image data resulting from the conversion is, for example, CIEXYZ. If it is image data, it can be converted into image data for display or output in consideration of information on the color reproduction characteristics of various display devices or output devices as conversion source data. Since it is very convenient in practice, a colorimetric image conversion method, a colorimetric image conversion device, and a colorimetric image conversion program are described that make it possible to execute these easily and at low cost. And to provide a computer-readable information recording medium.
[0008]
[Means for Solving the Problems]
In order to solve the above problems, the means provided by the present invention is as follows.
A colorimetric image that automatically identifies information that captures the color reproduction characteristics of the imaging system corresponding to the illumination light used at the time of shooting and uses this to convert the captured image data into another image data Conversion method,
(A) Spectral distribution data output step of obtaining and outputting spectral distribution data of illumination light at the time of photographing by analyzing image data obtained by photographing a standard plate on which a plurality of color charts are arranged with an imaging system;
(B) Comparison of spectral distribution data of output illumination light, color reproduction characteristics for each imaging system for each illumination light, and a file describing illumination light information for each wavelength A light source analysis step that identifies the most similar object as the illumination light at the time of shooting;
(C) a selected file output step of selecting and outputting a file corresponding to the illumination light information at the time of photographing by performing collation from (a);
(D) an image data conversion step of converting image data obtained by photographing with the same imaging system under the same illumination using a file corresponding to the selected illumination light into another image data;
A colorimetric image conversion method characterized by performing all the steps (a) to (d) above.
[0009]
The standard plate in the present invention is a plate in which a plurality of color charts having different spectral reflectances are arranged on the same surface. The reflection characteristics of the surface of each arranged color chart are highly diffusive, and the diffusivity and reflectance are made substantially uniform within the same color chart. The standard plate is preferably a flat surface. However, the plane does not necessarily have to be a plane as long as it does not cause a problem in terms of accuracy related to estimation of spectral characteristics.
[0010]
In the invention of claim 2,
2. The colorimetric image conversion method according to claim 1, wherein the other image data of (D) is CIEXYZ image data.
[0011]
In the invention of claim 3,
A colorimetric image that automatically identifies information that captures the color reproduction characteristics of the imaging system corresponding to the illumination light used at the time of shooting and uses this to convert the captured image data into another image data Conversion method,
(E) By analyzing image data obtained by photographing a standard plate with multiple color charts with an imaging system, a weight vector for representing the spectral distribution of illumination light at the time of photographing with a linear combination of basis functions is obtained. Obtaining and outputting a weight vector output step;
(F) A file that describes the weight vector for representing the spectral distribution of illumination light by a linear combination of basis functions, information that captures the color reproduction characteristics of each imaging system for each type of illumination light, and information about the illumination light A light source analysis step for comparing each weight vector using and identifying the closest one as illumination light at the time of photographing;
(G) A selection file output step of selecting and outputting a file corresponding to the illumination light information at the time of photographing by performing collation from (F);
(H) An image data conversion step of converting image data obtained by photographing with the same imaging system under the same illumination to another image data using a file corresponding to the selected illumination light;
A colorimetric image conversion method characterized by performing all of the above steps (e) to (c).
[0012]
In the invention of claim 4,
4. The colorimetric image conversion method according to claim 3, wherein the other image data of (K) is CIEXYZ image data.
[0013]
In the invention of claim 5,
A colorimetric image that automatically identifies information that captures the color reproduction characteristics of the imaging system corresponding to the illumination light used at the time of shooting and uses this to convert the captured image data into another image data A conversion device,
(G) Storage means for storing a file group in which information describing the color reproduction characteristics of each imaging system for each of various illumination lights is stored in a computer;
(G) Spectral distribution data output means for obtaining and outputting spectral distribution data of illumination light at the time of photographing by analyzing image data obtained by photographing a standard plate having a plurality of color charts with an imaging system;
(Sa) Information on the spectral distribution data of the output illumination light and information on the color reproduction characteristics of each imaging system stored in the storage means of (g) for each of the various illumination lights and information on the illumination light A light source analysis means for comparing each wavelength by using a file describing, and specifying the most similar one as illumination light at the time of photographing;
(F) A selected file output means for selecting and outputting a file corresponding to the illumination light information at the time of photographing by performing collation from (a);
(S) Image data conversion means for converting image data obtained by photographing with the same imaging system under the same illumination using a file corresponding to the selected illumination light into another image data;
A colorimetric image conversion apparatus comprising all of the above (K) to (S).
[0014]
In the invention of claim 6,
6. The colorimetric image conversion apparatus according to claim 5, wherein the other image data (C) is CIEXYZ image data.
[0015]
In the invention of claim 7,
A colorimetric image that automatically identifies information that captures the color reproduction characteristics of the imaging system corresponding to the illumination light used at the time of shooting and uses this to convert the captured image data into another image data A conversion device,
(C) Storage means for storing, in a computer, a file group describing information obtained by grasping color reproduction characteristics for each imaging system for each of various illumination lights;
(E) By analyzing the image data obtained by photographing a standard plate with multiple color charts with an imaging system, a weight vector for representing the spectral distribution of illumination light at the time of photographing with a linear combination of basis functions is obtained. A weight vector output means for obtaining and outputting;
(T) a weight vector for representing the spectral distribution of illumination light by a linear combination of basis functions;
(C) The color reproduction characteristics of each imaging system stored in the storage means are compared for each weight vector using a file describing the information of the illumination light and the information regarding the illumination light. A light source analysis means for identifying the closest object as illumination light at the time of photographing;
(H) a selection file output means for selecting and outputting a file corresponding to the illumination light information at the time of photographing by performing collation from the above (ta);
(Iv) Image data conversion means for converting image data obtained by photographing with the same imaging system under the same illumination using a file corresponding to the selected illumination light into another image data;
A colorimetric image conversion apparatus comprising all of the above (C) to (V).
[0016]
In the invention of claim 8,
8. The colorimetric image conversion apparatus according to claim 7, wherein the other image data is CIEXYZ image data.
[0017]
In the invention of claim 9,
Automatically identifies information that captures the color reproduction characteristics of the imaging system corresponding to the illumination light used at the time of shooting, and uses this to record a program that converts the captured image data into another image data Computer-readable information recording medium
(T) Storage processing that creates a group of files that describe the information obtained by grasping the color reproduction characteristics of each imaging system for each of various illumination lights, and stores them in a computer;
(G) Accepts input of image data obtained by photographing a standard plate with multiple color charts with an imaging system and analyzes it to obtain and output spectral distribution data of illumination light at the time of photographing. Distribution data output processing;
(Na) Information on the spectral distribution data of the output illumination light and the information on the color reproduction characteristics of each imaging system stored in the storage means of (te) for each of various illumination lights and information on the illumination light A light source analysis process that compares each wavelength using a file that describes and identifies the most similar as the illumination light at the time of shooting;
(D) A selected file output process for selecting and outputting a file corresponding to illumination light information at the time of photographing by performing collation from (n);
(Nu) An image data conversion process that specifies the imaging system used and converts image data obtained by photographing using a file corresponding to the selected illumination light into another image data;
A computer-readable information recording medium storing a program characterized by performing all of the above (te) to (nu).
[0018]
In the invention of claim 10,
The computer-readable information recording medium recording the program according to claim 9, wherein the other image data is CIEXYZ image data.
[0019]
In the invention of claim 11,
Automatically identifies information that captures the color reproduction characteristics of the imaging system corresponding to the illumination light used at the time of shooting, and uses this to record a program that converts the captured image data into another image data Computer-readable information recording medium
(E) A storage process for creating a file group describing information obtained by grasping the color reproduction characteristics of each imaging system for each of various illumination lights, and storing the file group in a computer;
(B) Accepts input of image data obtained by photographing a standard plate with multiple color charts with an imaging system, and analyzes this to represent the spectral distribution of illumination light at the time of photographing as a linear combination of basis functions A weight vector output process for obtaining and outputting a weight vector for
(C) A weight vector for representing the spectral distribution of illumination light by a linear combination of basis functions and the color reproduction characteristics for each imaging system stored in the storage means in (ne) for each of various illumination lights. A light source analysis process that compares each weight vector using a file that describes the grasped information and illumination light information, and identifies the closest one as the illumination light at the time of photographing;
(V) a selection file output process for selecting and outputting a file corresponding to the illumination light information at the time of photographing by performing collation from (c);
(F) An image data conversion process for specifying the imaging system used and converting image data obtained by photographing using a file corresponding to the selected illumination light into another image data;
A computer-readable information recording medium storing a program characterized by performing all of the above (ne) to (f).
[0020]
In the invention of claim 12,
12. The computer-readable information recording medium on which the program according to claim 11 is recorded, wherein the other image data of (F) is CIEXYZ image data.
[0021]
In practice, it is preferable that the color reproduction characteristics of the imaging system used for shooting be grasped for each of various illumination lights, and a file describing information regarding the grasped color reproduction characteristics (hereinafter referred to as a color reproduction information file). ).
As this color reproduction information file, a so-called ICC profile can be cited as a suitable example (if the color reproduction information is described, the ICC profile is not necessarily required). The plurality of color reproduction information files are mounted as a file group in the computer.
[0022]
If the type of illumination light is specified from the estimated spectral characteristics of the illumination light, a color reproduction information file corresponding to the illumination light at the time of shooting is automatically selected from the file group by matching the file group with a computer. Is done. Using the selected color reproduction information file, the image data of a landscape or a subject photographed with the same imaging system under the same illumination is converted into CIEXYZ image data. (Claims 1 to 12 are applicable, and in particular, corresponding to Claims 2, 4, 6, 8, 10, and 12)
The CIEXYZ image data is converted into image data for display or output in consideration of information on the color reproduction characteristics of the display device or output device.
As for the imaging system used, it is also possible to prepare several types of imaging systems in advance and create a color reproduction information file for each.
[0023]
Instead of converting image data into CIEXYZ image data, it is also possible to obtain image data necessary for display and output (without passing through CIEXYZ image data). (Corresponding to the claims 1, 3, 5, 7, 9, 11)
[0024]
The information recording medium on which the image conversion program according to the present invention is recorded is, for example, a medium using magnetism, light, magnetomagnetism, or other technology, and can read the program on which the computer or the processing apparatus is recorded. Anything is acceptable.
[0025]
DETAILED DESCRIPTION OF THE INVENTION
<1 configuration>
Hereinafter, an apparatus configuration related to the present invention will be exemplified and described in detail.
[0026]
FIG. 1 is an explanatory diagram showing a configuration of an automatic selection apparatus according to an embodiment of the present invention.
[0027]
In FIG. 1, reference numeral 1 denotes a standard plate, in which a plurality of color charts having different spectral reflectances are arranged on the same surface, and the surface on which the color chart is arranged in the direction of the imaging means is directed to the measurement target position in the space. Install.
The reflection characteristics of the surface of each color chart are highly diffusive, and the diffusivity and reflectivity are made substantially uniform within the same color chart.
[0028]
2 uses a digital camera as an imaging system. It processes the potential signal detected by the optical sensor and controls the shutter speed and aperture value, which are the exposure conditions of the imaging means, and the imaging means that records the image file using the gradation value of each pixel as color image data. The digital camera also has setting means for recording each exposure condition in an image file together with image data.
[0029]
A computer 3 functions as an analysis unit, a light source analysis unit, an image data conversion unit (in particular, a CIEXYZ image data conversion unit), and an image conversion unit.
[0030]
As the analysis means, first, after specifying the model of the digital camera 2 used by the user, the basis function for expressing the spectral distribution of the illumination light in a specific dimension and the imaging characteristic data of the digital camera 2 (gradation characteristics / (Spectral sensitivity characteristics / appropriate response range) and the image file of the standard plate obtained by the imaging means are analyzed, and for example, spectral characteristic data of illumination light is output in increments of 5 nm between 380 nm and 780 nm.
[0031]
As the light source analysis means, the color reproduction characteristics of the digital camera 2 are previously grasped for each of several illumination lights, a file describing the grasped information is created, and the plurality of color reproduction information files are created. These are installed in the computer 3 as a file group. An ICC profile is applied as this color reproduction information file. The computer 3 collates the analyzed spectral characteristic estimated value with the mounted file group, and determines a color reproduction information file corresponding to the illumination light at the time of photographing. In addition, a file group is created for each digital camera so as to support a plurality of digital cameras.
[0032]
The CIEXYZ image data conversion means converts image data obtained by photographing with the same digital camera under the same illumination to CIEXYZ image data.
As the image conversion means, when the obtained CIEXYZ image data is displayed on a monitor or printed matter is output, the image data is converted into image data for display or output in consideration of information on the color reproduction characteristics of the display device or output device. Perform the conversion process.
[0033]
<2 operation>
2.1 Overall operation
Hereinafter, the overall operation in an example of the selection method and the selection apparatus according to the present invention will be described.
[0034]
FIG. 2 is a flowchart for explaining the overall operation using the automatic selection apparatus according to the embodiment of the present invention.
[0035]
(S1: Flowchart installation)
The user installs the standard plate 1 on the measurement target, installs the digital camera 2 at a measurement position where the standard plate 1 is imaged in an appropriate size within the imaging range of the digital camera 2, and uses the standard plate 1 method. The directions of the standard plate 1 and the digital camera 2 are adjusted so that the line direction and the optical axis of the imaging system of the digital camera 2 coincide.
[0036]
(S2 in flowchart: setting of imaging means)
The user sets the shutter speed and the aperture value, which are the exposure conditions of the digital camera 2, so that the light reflected from the surface of the standard plate 1 falls within the appropriate exposure range of the digital camera 2.
[0037]
(S3 in the flowchart: imaging)
The user releases the shutter of the digital camera 2, and the digital camera 2 captures an image and records the image data and exposure conditions in an image file.
[0038]
(S4 in flowchart: image data transfer)
The user transfers the image file recorded in the digital camera 2 to the computer 3.
[0039]
(Flowchart s5: Standard plate position designation)
The user displays the image data in the image file on the monitor of the computer 3, observes it, and designates the position occupied by the standard plate 1 in the image.
[0040]
(S6 in flowchart: analysis processing)
First, the user designates the model of the digital camera 2 used for photographing on the computer 3.
After the designation, the computer 3 generates a basis function for expressing the spectral distribution of the illumination light in a specific dimension, imaging characteristic data (gradation characteristics / spectral sensitivity characteristics / appropriate response range) of the digital camera 2, and the standard plate 1. Based on the spectral characteristic data of each color chart surface, analysis is performed corresponding to the image data of the standard plate recorded in the image file and the exposure conditions of the digital camera 2 recorded in the image file at the time of photographing the standard plate Processing is performed, and spectral characteristic data of illumination light irradiated on the standard plate 1 is output.
[0041]
(S7 in the flowchart: light source analysis process)
On the computer 3, the spectral characteristic data of the illumination light obtained by the analysis process and the color reproduction information file group of the digital camera 2 for each illumination light are collated, and a color reproduction information file corresponding to the illumination light at the time of photographing is automatically obtained. To choose.
[0042]
(Flowchart s8: CIEXYZ image data conversion)
Using the color reproduction information profile selected in the light source analysis process, an image of a landscape or a subject photographed with the same digital camera under the same illumination is converted into CIEXYZ image data.
[0043]
(S9 in the flowchart: image conversion process)
When displaying image data obtained by CIEXYZ image data conversion on a monitor or outputting a printed matter, image data for display or output in consideration of information on color reproduction characteristics of the display device or output device Process to convert to.
[0044]
2.2 Details about the analysis process
Hereinafter, an analysis process for estimation in the estimation method or an example according to the present invention will be described based on the flowchart of FIG.
[0045]
(S10 in the flowchart: image data input)
A file describing image data obtained by photographing a scene including the standard plate 1 with a digital camera is transferred from the digital camera connected via a communication line to the computer 3.
[0046]
(S11 in the flowchart: color chart sampling)
The image data of the standard plate is displayed on the monitor, and the position of the standard plate in the photographed image is designated by the user. Pixels in a certain area near the center in each color chart specified from the designated position (for example, ) The average value of each of the three channels (R, G, B) is calculated, and (for example) gradation values corresponding to 24 color charts:
gRj, gGj, gBj (j: color chart number)
For example, a total of 72 pieces of gradation value data are obtained.
The three channels (R, G, B) are the most common.
[0047]
(S12 in flowchart: selection of appropriate gradation value)
From (for example) 3 × 24 gradation value data corresponding to all color charts, n not within the appropriate response range of the digital camera 2 recorded in advance₀N in total except for [= (72−n₀)] Gradation value data.
[0048]
(S13 in flowchart: gradation linearization)
N gradation value data are obtained by gradation characteristic data recorded in advance as a function F (x) as a nonlinear relationship between the energy amount of incident light to the digital camera 2 and the gradation value output by the digital camera 2. Linear tone value data that is linear in energy and standardized from 0 to 1;
[0049]
[Expression 1]

[0050]
(Where i: channel (R, G, B), j: color chart number)
Convert to
[0051]
(S14 in flowchart: standardization of exposure conditions)
The shutter speed (t) and the aperture value (f), which are exposure conditions set by the digital camera when the standard plate is photographed, are read from the file in which the image data is recorded, and n linear gradation value data are read. These exposure conditions are standardized to gradation values per shutter speed and aperture value 1, and response value data
[0052]
[Expression 2]

[0053]
And these are n-dimensional single-column response value vectors
[0054]
[Equation 3]

[0055]
(S15 in flowchart: creation of matrix of explanatory variables)
Spectral sensitivity characteristic data of each of the three channels describing the relationship of response values with respect to 5 nm intervals in the spectral energy (for example) wavelength range of 380 nm to 780 nm of the incident light of the digital camera 2 recorded in advance;
c_i(Λ)
When,
24 spectral reflectance data describing spectral reflectances of all color charts on the standard plate 1 in the range of 380 nm to 780 nm at intervals of 5 nm.
s_j(Λ)
When,
D basis functions for expressing the spectral distribution of illumination light in d dimensions;
b_k(Λ)
And
explanatory variable matrix of n rows × d columns;
[0056]
[Expression 4]

[0057]
Is created as follows.
[0058]
[Equation 5]

[0059]
[Formula 6]

[0060]
here,
λ: Wavelength (λ1, λ2, λ3, ..., λ81)
B: Each column represents a basis function b_kBasis function vector indicating (λ) (81 rows × d columns)
[0061]
(S16 in the flowchart: multiple regression analysis)
D basis functions for the spectral distribution of illumination;
b_k(Λ)
A weight vector consisting of d weighting coefficients to represent the linear combination of
[0062]
[Expression 7]

[0063]
Then the response value vector;
[0064]
[Equation 8]

[0065]
Is
[0066]
[Equation 9]

[0067]
Because it can be expressed as:
[0068]
[Expression 10]

[0069]
Response value vector;
[0070]
## EQU11 ##

[0071]
Is used as the objective variable, and the partial regression coefficient obtained is used as the weight vector.
[0072]
[Expression 12]

[0073]
And
However, the order d for expressing the spectral distribution of illumination by a linear combination of basis functions must be smaller than the number n of response values within the appropriate response range.
[0074]
(S17 in the flowchart: linear conversion)
a basis function matrix containing d basis functions;
[0075]
[Formula 13]

[0076]
And d-dimensional weight vector
[0077]
[Expression 14]

[0078]
And by
[0079]
[Expression 15]

[0080]
Is a vector that contains the spectral distribution data of unknown illumination light, with estimated values of a total of 81 spectral distributions in 5 nm increments ranging from 380 nm to 780 nm.
[0081]
[Expression 16]

[0082]
Get as.
[0083]
2.3 Details of light source analysis processing
Hereinafter, the light source analysis process in the estimation method or an example according to the present invention will be described based on the flowchart of FIG.
[0084]
In advance, the color reproduction characteristics of the digital camera used for photographing are grasped for each of various illumination lights, and a color reproduction information file describing information on the grasped color reproduction characteristics is created. For example, an ICC profile is applied as the color reproduction information file.
A character string that can identify what kind of illumination light is attached to the color reproduction information file as a file name, and this is used as an index. The corresponding illumination light information is described in the color reproduction information file. The created plurality of color reproduction information files are mounted in the computer as a file group.
[0085]
(S18 in the flowchart: spectral characteristic estimation value input)
The estimated spectral characteristic value of the illumination light output in the analysis process is input to the computer 3.
[0086]
(S19 in the flowchart: identification of the type of illumination light)
The type of illumination light is specified by using the spectral characteristic estimation value of the input illumination light and the illumination light information described in the created color reproduction information file.
[0087]
Here, as a method for specifying the type of illumination light from the estimated value of the spectral characteristics of illumination light and the information of the plurality of illumination lights, the following method is given.
[0088]
(1) A method of determining the type of illumination light using a spectral distribution waveform obtained from spectral distribution data output by analysis processing
Spectral distribution data output in increments of 5 nm between 380 nm and 780 nm by analysis processing,

As a comparison between the two.
Spectral distribution data in which the absolute value of the difference at each wavelength is averaged over the entire wavelength range is determined to be the smallest, and the illumination light corresponding to the data is determined as the illumination light to be determined.
[0089]
(2) Determine the type of illumination light from the d-dimensional weight vector presented in the analysis process Who
The estimated value of the spectral distribution is the d-dimensional weight vector presented in the analysis process.
[0090]
[Expression 17]

[0091]
When,
Basis function matrix
[0092]
[Expression 18]

[0093]
Is a value derived from
d-dimensional weight vector
[0094]
[Equation 19]

[0095]
Therefore, it is possible to specify what kind of illumination light is used.
Then, the d-dimensional weight vector presented in the analysis process is
[0096]
[Expression 20]

[0097]
Similarly, a d-dimensional weight vector for a plurality of illumination lights to be selected is
[0098]
[Expression 21]

[0099]
And still,
[0100]
[Expression 22]

[0101]
For the spectral distribution of illumination light corresponding to each
[0102]
[Expression 23]

[0103]
Approximation is performed in the form of
[0104]
[Expression 24]

[0105]
When,
[0106]
[Expression 25]

[0107]
Comparison with each is performed, the closest vector is specified, and it determines as the bright light which calculates | requires the illumination light corresponding to it.
[0108]
(S20 in the flowchart: collation process)
By specifying the type of illumination light, the file name of each file in the file group is checked on the computer 3 to automatically generate a color reproduction information file corresponding to the illumination light at the time of shooting from the file group. Select
[0109]
(S21 in the flowchart: output of the corresponding file)
The selected color reproduction information file is output as a color reproduction information file corresponding to the illumination light at the time of shooting.
[0110]
<3 Modification>
(A) The medium used as the imaging means is not limited to a digital still camera, and any medium such as a digital video camera or a scanner can be used as long as it can acquire image data.
However, in the case of a scanner, not a single unit but a pair with a film camera. At this time, it is necessary to prepare the color reproduction information file of the imaging means for each of the scanner and the camera.
(B) It is not necessary for the setting means to be integral with the imaging means, and it is also preferable that the setting means is provided in a part of the imaging means, for example, a computer.
[0111]
(C) Although the ICC profile is applied as the format of the color reproduction information file of the imaging means, it is not limited to this, and it is also preferable that data created in other free formats can be applied.
(D) In the light source analysis means, it is also preferable to provide an independent dedicated device without using a computer, or to incorporate a processing system for selection in the medium used as the imaging means.
[0112]
(E) When the type of illumination light is specified in the light source analysis means, not only information about illumination light corresponding to the color reproduction information file itself is described, but only information about the corresponding illumination light is collected. It is also preferable to prepare a correspondence table separately and specify the type of illumination light using it.
(F) An image captured by a digital camera is converted into CIEXYZ image data using the color reproduction information file selected by the light source analysis process. At this time, it is preferable to convert the image into CIELAB image data as well as CIEXYZ image data. .
[0113]
(G) An image photographed with a digital camera is converted into CIEXYZ image data using the color reproduction information file selected by the light source analysis process. This CIEXYZ image data is converted to display / output image data using a display system or output system profile. In addition to this display or output image data conversion, image analysis, image classification, etc. It is also preferable to perform various operations by applying CIEXYZ image data.
(H) When performing the light source analysis process from the estimated spectral characteristic value of the illumination light, the index is not limited to the spectral distribution data, but it is also preferable to use a value calculated from the spectral distribution data such as color temperature.
[0114]
【The invention's effect】
As described above, according to the present invention, first, a color reproduction characteristic of an imaging system used for photographing is grasped for each of various illumination lights, and a file (in which information related to the grasped color reproduction characteristic is described) Create a color reproduction information file) in advance. The multiple color reproduction information files are installed in the computer as a file group, and the type of illumination light at the time of shooting is identified and collated from the estimated spectral characteristics of the illumination light when shooting with the imaging system and the file group. This makes it possible to automatically select a color reproduction information file corresponding to the illumination light at the time of shooting. Using the selected color reproduction information file, an image photographed using the same input medium under the same illumination light as that at the time of photographing is converted into CIEXYZ image data. When this CIEXYZ image data is displayed on a monitor or printed matter is output, it can be converted into image data for display or output in consideration of information on the color reproduction characteristics of the display device or output device. Color conversion of an image in consideration of the influence of light can be easily performed.
[Brief description of the drawings]
FIG. 1 is an explanatory diagram schematically showing a part of the configuration of an example of a colorimetric image conversion method, a colorimetric image conversion apparatus, and a colorimetric image conversion program according to the present invention. (Shooting a standard plate with an imaging device)
FIG. 2 is a flowchart for explaining an overall operation flow of an example of a colorimetric image conversion method, a colorimetric image conversion apparatus, and a colorimetric image conversion program according to the present invention.
FIG. 3 is a flowchart for explaining a processing example performed by an analysis unit in the colorimetric image conversion method, colorimetric image conversion apparatus, and colorimetric image conversion program according to the present invention.
FIG. 4 is a flowchart for explaining a processing example performed by a light source analysis unit in the colorimetric image conversion method, colorimetric image conversion apparatus, and colorimetric image conversion program according to the present invention.
[Explanation of symbols]
1 ... Standard plate
2 ... Imaging means (typical example is digital camera)
3 ... Calculator

Claims (12)

A colorimetric image that automatically identifies information that captures the color reproduction characteristics of the imaging system corresponding to the illumination light used at the time of shooting and uses this to convert the captured image data into another image data Conversion method,
(A) Spectral distribution data output step of obtaining and outputting spectral distribution data of illumination light at the time of photographing by analyzing image data obtained by photographing a standard plate on which a plurality of color charts are arranged with an imaging system;
(B) Comparison of spectral distribution data of output illumination light, color reproduction characteristics for each imaging system for each illumination light, and a file describing illumination light information for each wavelength A light source analysis step that identifies the most similar object as the illumination light at the time of shooting;
(C) a selected file output step of selecting and outputting a file corresponding to the illumination light information at the time of photographing by performing collation from (a);
(D) an image data conversion step of converting image data obtained by photographing with the same imaging system under the same illumination using a file corresponding to the selected illumination light into another image data;
A colorimetric image conversion method characterized by performing all of the above steps (a) to (d). 2. The colorimetric image conversion method according to claim 1, wherein the other image data of (d) is CIEXYZ image data. A colorimetric image that automatically identifies information that captures the color reproduction characteristics of the imaging system corresponding to the illumination light used at the time of shooting and uses this to convert the captured image data into another image data Conversion method,
(E) By analyzing image data obtained by photographing a standard plate with multiple color charts with an imaging system, a weight vector for representing the spectral distribution of illumination light at the time of photographing with a linear combination of basis functions is obtained. Obtaining and outputting a weight vector output step;
(F) A file that describes the weight vector for representing the spectral distribution of illumination light by a linear combination of basis functions, information that captures the color reproduction characteristics of each imaging system for each type of illumination light, and information about the illumination light A light source analysis step for comparing each weight vector using and identifying the closest one as illumination light at the time of photographing;
(G) A selection file output step of selecting and outputting a file corresponding to the illumination light information at the time of photographing by performing collation from (F);
(H) An image data conversion step of converting image data obtained by photographing with the same imaging system under the same illumination to another image data using a file corresponding to the selected illumination light;
A colorimetric image conversion method characterized by performing all of the above steps (e) to (c). 4. The colorimetric image conversion method according to claim 3, wherein the other image data of (K) is CIEXYZ image data. A colorimetric image that automatically identifies information that captures the color reproduction characteristics of the imaging system corresponding to the illumination light used at the time of shooting and uses this to convert the captured image data into another image data A conversion device,
(G) Storage means for storing a file group in which information describing the color reproduction characteristics of each imaging system for each of various illumination lights is stored in a computer;
(G) Spectral distribution data output means for obtaining and outputting spectral distribution data of illumination light at the time of photographing by analyzing image data obtained by photographing a standard plate having a plurality of color charts with an imaging system;
(Sa) Information on the spectral distribution data of the output illumination light and information on the color reproduction characteristics of each imaging system stored in the storage means of (g) for each of the various illumination lights and information on the illumination light A light source analysis means for comparing each wavelength by using a file describing, and specifying the most similar one as illumination light at the time of photographing;
(F) A selected file output means for selecting and outputting a file corresponding to the illumination light information at the time of photographing by performing collation from (a);
(S) Image data conversion means for converting image data obtained by photographing with the same imaging system under the same illumination using a file corresponding to the selected illumination light into another image data;
A colorimetric image conversion apparatus comprising all of the above (K) to (S). 6. The colorimetric image conversion apparatus according to claim 5, wherein the other image data is CIEXYZ image data. A colorimetric image that automatically identifies information that captures the color reproduction characteristics of the imaging system corresponding to the illumination light used at the time of shooting and uses this to convert the captured image data into another image data A conversion device,
(C) Storage means for storing, in a computer, a file group describing information obtained by grasping color reproduction characteristics for each imaging system for each of various illumination lights;
(E) By analyzing the image data obtained by photographing a standard plate with multiple color charts with an imaging system, a weight vector for representing the spectral distribution of illumination light at the time of photographing with a linear combination of basis functions is obtained. A weight vector output means for obtaining and outputting;
(T) A weight vector for representing the spectral distribution of illumination light by a linear combination of basis functions and the color reproduction characteristics for each imaging system stored in the storage means in (C) for each of various illumination lights. Light source analysis means for comparing each weight vector using a file describing the grasped information and illumination light information, and specifying the closest one as the illumination light at the time of photographing;
(H) a selection file output means for selecting and outputting a file corresponding to the illumination light information at the time of photographing by performing collation from the above (ta);
(Iv) Image data conversion means for converting image data obtained by photographing with the same imaging system under the same illumination using a file corresponding to the selected illumination light into another image data;
A colorimetric image conversion apparatus comprising all of the above (se) to (tu). 8. The colorimetric image conversion apparatus according to claim 7, wherein the other image data is CIEXYZ image data. Automatically identifies information that captures the color reproduction characteristics of the imaging system corresponding to the illumination light used at the time of shooting, and uses this to record a program that converts the captured image data into another image data Computer-readable information recording medium
(T) Storage processing that creates a group of files that describe the information obtained by grasping the color reproduction characteristics of each imaging system for each of various illumination lights, and stores them in a computer;
(G) Accepts input of image data obtained by photographing a standard plate with multiple color charts with an imaging system and analyzes it to obtain and output spectral distribution data of illumination light at the time of photographing. Distribution data output processing;
(Na) Information on the spectral distribution data of the output illumination light and the information on the color reproduction characteristics of each imaging system stored in the storage means of (te) for each of various illumination lights and information on the illumination light A light source analysis process that compares each wavelength using a file that describes and identifies the most similar as the illumination light at the time of shooting;
(D) A selected file output process for selecting and outputting a file corresponding to illumination light information at the time of photographing by performing collation from (n);
(Nu) An image data conversion process that specifies the imaging system used and converts image data obtained by photographing using a file corresponding to the selected illumination light into another image data;
A computer-readable information recording medium recording a program characterized by performing all of the above (te) to (nu). The computer-readable information recording medium recording the program according to claim 9, wherein the other image data is CIEXYZ image data. Automatically identifies information that captures the color reproduction characteristics of the imaging system corresponding to the illumination light used at the time of shooting, and uses this to record a program that converts the captured image data into another image data Computer-readable information recording medium
(E) A storage process for creating a file group describing information obtained by grasping the color reproduction characteristics of each imaging system for each of various illumination lights, and storing the file group in a computer;
(B) Accepts input of image data obtained by photographing a standard plate with multiple color charts with an imaging system, and analyzes this to represent the spectral distribution of illumination light at the time of photographing as a linear combination of basis functions A weight vector output process for obtaining and outputting a weight vector for
(C) A weight vector for representing the spectral distribution of illumination light by a linear combination of basis functions and the color reproduction characteristics for each imaging system stored in the storage means in (ne) for each of various illumination lights. A light source analysis process that compares each weight vector using a file that describes the grasped information and illumination light information, and identifies the closest one as the illumination light at the time of photographing;
(V) a selection file output process for selecting and outputting a file corresponding to the illumination light information at the time of photographing by performing collation from (c);
(F) An image data conversion process for specifying the imaging system used and converting image data obtained by photographing using a file corresponding to the selected illumination light into another image data;
A computer-readable information recording medium storing a program characterized by performing all of the above (ne) to (f). The other image data of (F) is CIEXYZ image data,
A computer-readable information recording medium on which the program according to claim 11 is recorded.

Images