JP3737919B2 - Photogrammetry target reference point calculation device, photogrammetry target reference point calculation method, and recording medium storing photogrammetry target reference point calculation program - Google Patents

Description

【００４６】
（１）式において、αはＸｓ軸のＸｃ軸回りの回転角、βはＹｓ軸のＹｃ軸回りの回転角、γはＺｓ軸のＺｃ軸回りの回転角であり、Ｒはα、β、γの回転行列、Δはシーン座標からカメラ座標へのベクトルである。
【００４７】
また、カメラ座標系における基準点の座標値Ｐｃｉ（Ｐｃｘｉ，Ｐｃｙｉ，Ｐｃｚｉ）（ｉ＝１、２、３）と、写真座標系における基準点の座標値Ｄｉ（Ｄｘｉ、Ｄｙｉ）（ｉ＝１、２、３）との関係は式（２）で表わされる。尚、（２）式における「ｆ」はカメラ１００の撮影レンズの焦点距離である。
【００４８】
【数２】

【００４９】
以上のように、上述の（１）式および（２）式により、基準点のシーン座標系における座標値から、写真座標系における座標値が算出される。一方、写真座標系における基準点の座標値は、後述するように撮影画像より自動抽出される。従って、（１）式および（２）式から算出した基準点の座標値と、撮影画像から自動抽出した座標値とを比較することにより、カメラ位置を定義するΔＸ、ΔＹ、ΔＺ、α、β、γの値を算出することができる。
【００５０】
図１７は、ＩＣカード２４に格納された画像データに基づいて写真測量が行なわれる写真測量装置の座標算出システムのブロック図である。図１７に示すように、座標算出システムはコンピュータ１８４を備えるコンピュータシステムで構成される。コンピュータ１８４は、中央演算処理装置（ＣＰＵ）１８４Ａ、オペレーティングシステム、常数等が格納された読出し専用メモリ（ＲＯＭ）１８４Ｂ、データ等を一時的に格納する書込み／読出し自在なメモリ（ＲＡＭ）１８４Ｃ、ＣＰＵ１８４Ａで実行された演算結果を格納するためのハードディスク１８４Ｄ、入出力インターフェース（Ｉ／Ｏ）１８４Ｅを有する。
【００５１】
コンピュータシステムはまた、Ｉ／Ｏ１８４Ｅを介してコンピュータ１８４に接続されるＩＣカードリーダ１８６を備える。ＩＣカードリーダ１８６はＩＣメモリカード２４が装着されるスロットを備えており、１フレーム分のデジタル画像や他のデータを読み込むためのＩＣカードドライバ１８８を有する。
【００５２】
また、コンピュータシステムは、Ｉ／Ｏ１８４Ｅを介してコンピュータ１８４に接続されるＣＤ−ＲＯＭリーダ１９７を備える。ＣＤ−ＲＯＭリーダ１９７は、ＣＤ−ＲＯＭ１９８を装着するためのトレーを備えており、ＣＤ−ＲＯＭドライブユニット１９９を有する。後述するターゲットの基準点やカメラ位置の算出プログラムは、ＣＤ−ＲＯＭ１９８に格納されている。この算出プログラムはＣＤ−ＲＯＭ１９８から読み出され、ハードディスク１８４Ｄに記憶される。
【００５３】
さらに、コンピュータシステムは、ＩＣメモリカード２４から読み出された１フレーム分のデジタル画像データに基づく撮影画像の再生や、コンピュータ１８４により作成される測量図の再生を行なうためのモニタ１９０、種々のコマンド信号やデータをコンピュータ１８４に入力するためのキーボード１９２、モニタ１９０に表示されるカーソルを操作するためのマウス１９４、必要に応じて測量図を印刷するためのプリンタ１９６を備える。
【００５４】
図１１は、上述のＣＰＵ１８４Ａにより実行される、本実施形態のカメラ位置算出の手順を示すフローチャートである。ステップＳ１０００において、上述のターゲット１０を設置して撮影した測量現場のデジタル画像が本実施形態の写真測量装置にロードされ、次いでステップＳ１００２でグレースケール処理が行なわれる。グレースケール処理とは、カラー画像から色成分を除去し、輝度情報のみを抽出する処理である。デジタル画像を構成する全ての画素毎に２５６階調の輝度情報が抽出され、配列形式の変数Ｉｍｇに格納される。すなわち、各画素の輝度情報はＩｍｇ（ｊ，ｋ）で表わされる。ｊおよびｋは画素のデジタル画像における２次元座標値（写真座標系の座標値）である。
【００５５】
ステップＳ１００４で各画素毎に抽出された輝度情報の２値化処理が行なわれる。変数Ｉｍｇに格納された輝度情報を所定の閾値を比較し、その輝度情報が高く閾値を超えていれば「１」、輝度情報が低く閾値以下であれば「０」をＩｍｇと同様の配列形式の変数Ｂｉｎに格納する。すなわち、各画素の輝度情報を２値化したデータはＢｉｎ（ｊ，ｋ）で表わされる。上述のように、本実施形態で用いられるターゲット１０の基準点部材３１、３４、３６、補助点部材３２、３３、３５には反射シートが貼付され、その周りには、黒色の無反射部材４１、４２、４３、４４、４５、４６が設けられている。従って、基準点部材３１、３４、３６および補助点部材３２、３３、３５に対応する画素の２値化データは「１」となり、周囲の無反射部材４１、４２、４３、４４、４５、４６に対応する画素の２値化データは「０」となる。
【００５６】
次いで、ステップＳ１００６でラベリングが行なわれる。ラベリングとは、２値化データが「１」である画素が連続している部分を１つのグループとして抽出する処理であり、デジタル画像において輝度情報の高い領域が抽出される。各グループ毎に、そのグループを構成する画素の数およびそれぞれの２次元座標値が所定の変数に格納される。
【００５７】
次いで、ステップＳ１００８で各グループの面積、すなわち画素数が算出され、ステップＳ１０１０で、各グループの画素数に基づいて、そのグループが基準点部材若しくは補助点部材に相当するか否かを判別する。具体的には、画素数が下限値Ｔｈ１と上限値Ｔｈ２の範囲内にあるか否かをチェックする。下限値Ｔｈ１および上限値Ｔｈ２は、それにより規定される画素数の範囲が、デジタル画像内の基準点部材、補助点部材に相当するグループの画素数を十分に含み、かつ最小の範囲となるよう設定されている。従ってステップＳ１０１０の処理により、デジタル画像のノイズに相当する画素数が１、２個程度のグループや、測量現場の道路上においてペンキが塗布されたような、広範囲にわたって高い輝度情報を有する領域に相当するグループが除去される。
【００５８】
次いでステップ１０１２において、各グループの輝度情報を考慮した重心の算出が行われる。図１２は、重心算出の処理手順を示すフローチャートである。ステップＳ１０２２で変数ＳｕｍＸ、ＳｕｍＹ、ＳｕｍＩｍｇが初期化される。変数ＳｕｍＸには、グループを構成する各画素の輝度情報とｘ座標値との積が累積して格納され、変数ＳｕｍＹには、グループを構成する各画素の輝度情報とｙ座標値との積が累積して格納される。また、変数ＳｕｍＩｍｇは、グループを構成する各画素の輝度情報を加算した総和が格納される。
【００５９】
ステップＳ１０２６で、画素の輝度情報（Ｉｍｇ（ｊ，ｋ））と画素のｘ座標値（ｊ）の積が算出され、その時点で変数ＳｕｍＸに格納されている値に加算した上で変数ＳｕｍＸに格納される。次いでステップＳ１０２８で、画素の輝度情報と画素のｙ座標値（ｋ）の積が算出され、その時点で変数ＳｕｍＹに格納されている値に加算した上で変数ＳｕｍＹに格納される。さらに、ステップＳ１０３０で、画素の輝度情報がその時点で変数ＳｕｍＩｍｇに格納されている値に加算され、変数ＳｕｍＩｍｇに格納される。以上のステップＳ１０２６〜Ｓ１０３０の処理が、グループを構成する全ての画素に対して繰り返される。
【００６０】
すなわち、変数ＳｕｍＸには、各画素の輝度情報をｘ方向の座標値に基づいて重み付けをした値の総和が格納され、変数ＳｕｍＹには、各画素の輝度情報をｙ方向の座標値に基づいて重み付けをした値の総和が格納される。また、変数ＳｕｍＩｍｇには、各画素の輝度情報の総和が格納される。
【００６１】
ステップＳ１０３２でグループを構成する全ての画素に対して、上述の輝度情報の累積演算処理が行われたか否かがチェックされる。グループを構成する全ての画素についてステップＳ１０２６〜Ｓ１０３０の処理が行われたら、ステップＳ１０３４へ進む。ステップＳ１０３４では、各画素の輝度情報をｘ方向の座標値に基づいて重み付けをした値の総和を、各画素の輝度情報の総和により除算する処理が行なわれ、グループの重心のｘ座標値が算出される。ステップＳ１０３６では、各画素の輝度情報をｙ方向の座標値に基づいて重み付けをした値の総和を、各画素の輝度情報の総和により除算する処理が行なわれ、グループの重心のｙ座標値が算出される。
【００６２】
次いで、ステップＳ１０３８において、撮影画像内の全てのグループについて輝度情報を考慮した重心の座標の算出が行われたか否かがチェックされる。全てのグループについての重心の座標の算出処理が終了していない場合は、ステップＳ１０２０〜Ｓ１０３６までの処理が繰り返される。全てのグループについての重心の座標の算出処理が終了したら、図１１のステップＳ１０１４へ進む。
【００６３】
図１３は、各グループの重心座標の算出が行なわれた後の撮影画像の各画素の状態を模式的に示す図であり、輝度情報の低い画素には斜線を付してある。図１３に示すように、グループＧ１〜Ｇ１３が抽出されており、それぞれの重心はＧＣ１〜ＧＣ１３である。
【００６４】
ステップＳ１０１４では、ターゲット１０の基準点部材３１、３４、３６、補助点部材３２、３３、３５の相対関係を満たす、グループの組み合わせを抽出する。図１４はグループの組み合わせ抽出処理の前半を示すフローチャート、図１５は後半を示すフローチャートである。以降、グループの組合せ抽出処理の手順を図１３を参照しながら説明する。
【００６５】
ステップＳ１０４０で画像内のグループから任意の２グループを選択され、ステップＳ１０４２で、選択した２グループの重心を含む直線が算出される。次いでステップＳ１０４４で直線上において、選択した２グループの重心の間に重心を有するグループが検出される。
【００６６】
ステップＳ１０４６において、ステップＳ１０４４で検出されたグループの数がチェックされる。直線上に重心を有するグループがステップＳ１０４０で選択された２グループを含めて４つの場合はステップＳ１０４８へ進み、３つの場合はステップＳ１０５０へ進み、その他の場合はステップＳ１０５２へ進む。
【００６７】
例えば、グループＧ１とグループＧ２が選択された場合、それぞれの重心ＧＣ１、ＧＣ２を含む直線上には他のグループの重心が存在しないため、処理はステップＳ１０５２へ進む。グループＧ３とグループＧ６が選択された場合、若しくはグループＧ１０とグループ１３が選択された場合、重心ＧＣ３、ＧＣ６を含む直線上にはグループＧ４の重心ＧＣ４とグループＧ５の重心ＧＣ５の２つの重心が存在し、重心ＧＣ１０、ＧＣ１３を含む直線上にはグループＧ１１の重心ＧＣ１１とグループＧ１２の重心ＧＣ１２の２つの重心が存在する。従って処理はステップＳ１０４８へ進む。グループＧ１とグループＧ３が選択された場合、若しくはグループＧ７とグループＧ９が選択された場合、重心ＧＣ１、ＧＣ３を含む直線上にはグループＧ２の重心ＧＣ２が１つ存在し、重心ＧＣ７、ＧＣ９を含む直線上には、グループＧ８の重心ＧＣ８が１つ存在する。従って処理はステップＳ１０５０へ進む。
【００６８】
ステップＳ１０４８では、４つのグループが同一直線上に並列する４点列として登録され、ステップＳ１０５０では、３つのグループが同一直線上に並列する３点列として登録される。例えば図１３の各グループは、グループＧ１、Ｇ２、Ｇ３と、グループＧ６、Ｇ７、Ｇ８がそれぞれ３点列として登録され、グループＧ４、Ｇ５、Ｇ６、Ｇ７と、グループＧ１０、Ｇ１１、Ｇ１２、Ｇ１３がそれぞれ４点列として登録される。
【００６９】
ステップＳ１０５２では、２グループの全ての組合せについて処理が行なわれたか否かがチェックされ、全ての組合せについて処理が終了していない場合はステップＳ１０４０〜Ｓ１０５０の処理が繰り返される。２グループの全ての組合せについて処理が終了したら、図１５のステップＳ１０５４へ進む。
【００７０】
ステップＳ１０５４では、それぞれ同一直線上にある３点列と４点列が１つづつ選択され、それぞれの端点に相当するグループの重心が一致するか否かがチェックされる。それぞれの端点に相当するグループの重心が一致する場合ステップＳ１０５８へ進み、端点に相当するグループの重心が一致しない場合、ステップＳ１０５４へ戻り、他の組合せの３点列と４点列が選択される。図１３において、グループＧ３は、グループＧ１、Ｇ２、Ｇ３の３点列の端点であり、かつグループＧ３、Ｇ４、Ｇ５、Ｇ６の４点列の端点である。従って、３点列と４点列の組合せとして、グループＧ１、Ｇ２、Ｇ３と、グループＧ３、Ｇ４、Ｇ５、Ｇ６が選択される。
【００７１】
ステップＳ１０５８では、４点列を構成するグループのうち、３点列の端点と一致する端点と反対側の端点に相当するグループの重心の座標が、基準点部材３１の座標を示す変数Ａ１（Ａｘ₁ 、Ａｙ₁ ）に格納され、３点列と４点列の一致する端点に相当するグループの重心の座標が基準点部材３４の座標を示す変数Ａ２（Ａｘ₂ 、Ａｙ₂ ）に格納され、３点列を構成するグループのうち、４点列の端点と一致する端点と反対側の端点に相当するグループの重心の座標が、基準点部材３６の座標を示す変数Ａ３（Ａｘ₃ 、Ａｙ₃ ）に格納される。すなわち、図１３の重心ＧＣ６の座標が変数Ａ１に格納され、重心ＧＣ３が変数Ａ２に格納され、重心ＧＣ１が変数Ａ３に格納される。
【００７２】
次いでステップＳ１０６０において、変数Ａ２から変数Ａ１へ向かうベクトルに対する、変数Ａ２から変数Ａ３へ向かうベクトルの成す角度がチェックされる。変数Ａ２から変数Ａ１へ向かうベクトルに対する、変数Ａ２から変数Ａ３へ向かうベクトルの成す反時計回りの角度θが、下限値Ｔｈ３より大きくかつ上限値Ｔｈ４より小さいか否かがチェックされる。本実施形態では、下限値Ｔｈ３は０度、上限値Ｔｈ４は１８０度に設定されている。ステップＳ１０６０の条件を満たす場合、選択された３点列と４点列の組合せがターゲット１０の３つの基準点部材３１、３４、３６、補助点部材３２、３３、３５の相対的関係を満たす組合せであるので、グループの組み合わせ抽出処理を終了し、図１１のステップＳ１０１６へ進む。
【００７３】
角度θが下限値Ｔｈ３と上限値Ｔｈ４の範囲内にない場合、ステップＳ１０６２へ進み、３点列と４点列の全ての組合せについて処理が行なわれたか否かがチェックされる。ステップＳ１０５６〜Ｓ１０６０までの処理が行なわれていない３点列と４点列の組合せが存在する場合は、ステップＳ１０５４へ戻り、ステップＳ１０６０までの処理を繰り返す。
【００７４】
すなわち、基準点部材３１、３４、３６、補助点部材３２、３３、３５の相対的関係を満たす３点列と４点列の組合せが検出されるまでステップＳ１０５４〜Ｓ１０６２の処理が繰り返される。基準点部材と補助点部材の相対的関係を満たす３点列と４点列の組合せが検出されたら、他の３点列と４点列の他の組合せの処理は行なわずに処理を終了する。
【００７５】
図１３の３点列と４点列の組合せ、グループＧ１、Ｇ２、Ｇ３と、グループＧ３、Ｇ４、Ｇ５、Ｇ６に関して、変数Ａ２から変数Ａ１へ向かうベクトルに対する変数Ａ２から変数Ａ３へ向かうベクトルの反時計回りの角度は、図１３から明らかなように、下限値Ｔｈ３と上限値Ｔｈ４の範囲内にある。従って、グループＧ１、Ｇ２、Ｇ３と、グループＧ３、Ｇ４、Ｇ５、Ｇ６は、基準点部材３１、３４、３６、補助点部材３２、３３、３５の相対的関係を満たす３点列と４点列の組合せとして検出される。
【００７６】
ステップＳ１０１６において、カメラ位置の算出処理が行われる。図１６は、カメラ位置算出の処理手順を示すフローチャートである。ステップＳ１０７０で、上述の（１）式の変数ΔＸ、ΔＹ、ΔＺ、α、β、γを初期化する。次いで、ステップＳ１０７２で、（１）式のＰｓ１、Ｐｓ２、Ｐｓ３にそれぞれ（−Ｌ，０，０）、（０，０，０）、（０，０，Ｌ）を代入し、かつ変数ΔＸ、ΔＹ、ΔＺ、α、β、γに適当な値を代入し、カメラ座標系における基準点部材３１、３４、３６の各座標値Ｐｃ１（Ｐｃｘ₁ 、Ｐｃｙ₁ 、Ｐｃｚ₁ ）、Ｐｃ２（Ｐｃｘ₂ 、Ｐｃｙ₂ 、Ｐｃｚ₂ ）、Ｐｃ３（Ｐｃｘ₃ 、Ｐｃｙ₃ 、Ｐｃｚ₃ ）を算出する。
【００７７】
ステップＳ１０７４で、上述の（２）式にＰｃ１、Ｐｃ２、Ｐｃ３の各座標値を代入し、写真座標系における基準点部材３１、３４、３６の各座標値Ｄ１（Ｄｘ₁ 、Ｄｙ₁ ）、Ｄ２（Ｄｘ₂ 、Ｄｙ₂ ）、Ｄ３（Ｄｘ₃ 、Ｄｙ₃ ）を算出する。
【００７８】
ステップＳ１０７６において、撮影画像から自動抽出した基準点部材３１、３４、３６の写真座標系における各座標値Ａ１、Ａ２、Ａ３と、上述の（１）式および（２）式を用いて算出した座標値Ｄ１、Ｄ２、Ｄ３とを比較する。すなわち、各基準点部材の座標値Ｄ１、Ｄ２、Ｄ３とＡ１、Ａ２、Ａ３との差分の絶対値をそれぞれ算出し、その合計値Ｓを算出する。
【００７９】
ステップＳ１０７８で、合計値Ｓが最小であるか否かを確認し、最小でない場合は、変数ΔＸ、ΔＹ、ΔＺ、α、β、γの値を変えてステップＳ１０７２からＳ１０７６の処理を繰り返す。すなわち、合計値Ｓが最小となる変数ΔＸ、ΔＹ、ΔＺ、α、β、γを最小二乗法により算出する。
【００８０】
ステップＳ１０７８で合計値Ｓが最小であることが確認されたら、ステップＳ１０８０において合計値Ｓが最小となった場合の変数ΔＸ、ΔＹ、ΔＺ、α、β、γの値をカメラ位置Ｍを示すパラメータとして登録し、カメラ位置算出処理を終了する。
【００８１】
以上のように本実施形態によれば、ターゲット１０において基準点部材３１、３４の間に設けられた補助点部材と、基準点部材３４、３６の間に設けられた補助点部材の数が異なり、かつ各基準点部材および各補助点部材に反射シートが貼付されているため、撮影画像内において輝度情報により各基準点部材、補助点部材の位置を特定することが容易である。従って、撮影画像内における基準点部材３１、３４、３６の座標値が自動的に抽出される。
【００８２】
また、本実施形態によれば、ターゲット１０の基準点部材３１、３４の間の長さと基準点部材３４、３６の間の長さ、および第１の柱状部材１２と第２の柱状部材１４の角度が既知であるため、シーン座標系における各基準点部材の座標値が所定の値に定まっており、その座標値に基づいて写真座標系における各基準点部材の座標値を算出できる。従って、撮影画像内で自動抽出した基準点部材の座標値と比較することにより、カメラ位置を自動算出することができる。
【００８３】
尚、図１１、１２、１４、１５、１６に示すフローチャートの処理は、測量現場のデジタル画像をモニタに表示し、モニタ上でターゲットの基準点及び任意の測量点を指定し、所定の測量演算等が実行される写真測量システムにおいて、ターゲットの基準点の２次元座標の算出及びカメラ位置の算出を実行する部分である。即ち、本実施形態によれば、モニタに表示されたデジタル画像内のターゲットの基準点をマウスにより指定するという作業が省略されると共に、担当者の熟練度に左右されない正確なカメラ位置のデータが常時得られ、写真測量システムにおける以降の処理段階において精度の高い演算処理が実行される。
【００８４】
【発明の効果】
以上のように、本発明によれば、写真測量においてカメラ位置を自動的に算出することができる。
【図面の簡単な説明】
【図１】本発明の実施形態に適用される写真測量用ターゲットを示す平面図である。
【図２】図１に示す写真測量用ターゲットの側面図である。
【図３】図１のＶＩＩＩ−ＶＩＩＩ線断面におけるターゲットの断面図である。
【図４】図１に示すターゲットの無反射部材の第２の柱状部材側の面を示す平面図である。
【図５】図１に示すターゲットの制御部筐体の近傍を拡大して示す平面図であり、一部破断して示す図である。
【図６】図５のＸＩ−ＸＩ線における断面図であり、制御部筐体の構成を簡略化して示す図である。
【図７】本発明の実施形態である写真測量用ターゲットとカメラとの位置関係を示す斜視図である。
【図８】図７のカメラで撮影した画像を模式的に示す図である。
【図９】基準点部材と、その像点と、カメラの撮影レンズの後側主点位置との位置関係を３次元座標で示す図である。
【図１０】シーン座標系とカメラ座標系を相対的に示す図である。
【図１１】本発明の実施形態のカメラ自動算出処理の手順を示すフローチャートである。
【図１２】撮影画像内で高い輝度情報を有するグループの重心を算出する手順を示すフローチャートである。
【図１３】グループの重心座標の算出が行なわれた後の撮影画像の各画素の状態を模式的に示す図である。
【図１４】グループの組み合わせ抽出処理の前半を示すフローチャートである。
【図１５】グループの組み合わせ抽出処理の後半を示すフローチャートである。
【図１６】カメラ位置算出の処理手順を示すフローチャートである。
【図１７】写真測量装置の座標算出システムのブロック図である。
【符号の説明】
１０ ターゲット
１２ 第１の柱状部材
１４ 第２の柱状部材
１５ ヒンジ
１６ ステー
２０ 制御部筐体
２４ ＩＣカード
３１、３４、３６ 基準点部材
３２、３３、３５ 補助点部材
４１、４２、４３、４４、４５、４６ 無反射部材
９２ ステー用ヒンジ
９４ ロックヒンジ
１８４ コンピュータ
１８６ ＩＣカードリーダ
１８８ ＩＣカードドライバ
１９７ ＣＤ−ＲＯＭリーダ
１９８ ＣＤ−ＲＯＭ
１９９ ＣＤ−ＲＯＭドライブユニット
１９０ モニタ
１９２ キーボード
１９４ マウス
１９６ プリンタ[0001]
BACKGROUND OF THE INVENTION
The present invention relates to calculation of a camera position in photogrammetry.
[0002]
[Prior art]
Conventionally, in photogrammetry such as traffic accident investigation, survey by stereo photography is known. In surveying by stereo photography, the site is photographed simultaneously by two cameras at a predetermined distance, and the stereo image obtained as a result is stereoscopically viewed with an optical machine as a point in the three-dimensional image. The coordinates of each survey point are calculated. In order to calculate the survey point, information on the position of the two cameras that have captured the captured image, that is, information on the distance from the camera to the subject, the angle of the camera with respect to the subject, and the like is necessary. Therefore, the person in charge must always measure and record the position information of the camera at the shooting site.
[0003]
However, it is cumbersome to measure the distance and angle of the camera with respect to the subject every time it is taken, and it takes time and effort. Therefore, it is always necessary to obtain accurate camera position information in accident investigations where quick processing is required. It was difficult. In addition, operations such as stereoscopic viewing of stereo images with optical machines and designation of points in the three-dimensional image depend on the skill level of the person in charge, so in terms of survey work efficiency and survey accuracy There was also a problem.
[0004]
In order to solve such a problem, a photogrammetry system using a target having three reference point members has been proposed. A plurality of photographs are taken at different camera positions and angles with the above-mentioned target installed at the surveying site, and stored as digital images in a computer memory. A suitable pair of digital images is selected from a plurality of digital images and displayed on a monitor connected to a computer, and a target reference point member is designated in each image using a mouse or the like connected to the computer. To do. Camera position data is calculated based on the coordinate values of the designated reference point member. Furthermore, an arbitrary survey point in the image on the monitor is designated, and an appropriate calculation process is executed based on the coordinate value of the designated survey point and the camera position data, and a white map of the survey site is created and displayed on the monitor. .
[0005]
[Problems to be solved by the invention]
However, the operation of operating the mouse while designating the digital image displayed on the monitor and designating the reference point member largely depends on the operation technique of the operator. In addition, depending on the situation of the surveying site, there are cases where a large number of photographs are taken, and designation of a reference point member with a mouse imposes a great deal of labor on the operator. That is, the designation of the reference point member is not always accurately performed. When the reference point member is not specified correctly, the camera position data with high accuracy is not calculated. As a result, the calculation processing based on the camera position data is affected, and the white map is not created with high accuracy. .
[0006]
The present invention solves the above-described problems, and an object thereof is to provide a photogrammetry apparatus that automatically calculates a camera position from a captured image.
[0007]
[Means for Solving the Problems]
In the photogrammetric target reference point calculation apparatus according to the present invention, first, second, and third reference points whose distances are known are provided, and the first reference point and the second reference point are determined. The first straight line that connects, the second straight line that connects the second reference point, and the third reference point have a predetermined angle, and there are different auxiliary points on the first straight line and the second straight line, respectively. An image acquisition means for photographing a surveying object with a camera together with a photogrammetry target, and the number of auxiliary points on the first straight line and a second straight line in a photographed image obtained by the image acquisition means By comparing the number of auxiliary points, the positions of the first reference point, the second reference point, and the third reference point in the captured image are specified, and the photographic coordinates that are two-dimensional orthogonal coordinates in the captured image Secondary to obtain the coordinate values of the first reference point, the second reference point, and the third reference point in the system Characterized by comprising a coordinate acquisition unit.
[0008]
Preferably, in the photogrammetry target, the first reference point, the second reference point, the third reference point, and the auxiliary point have reflection information that is relatively higher than other subjects in the captured image. The two-dimensional coordinate acquisition means compares the luminance information of each pixel constituting the captured image with a predetermined threshold value, and each pixel has two types of pixels having high luminance information and pixels having low luminance information. Binarization processing means, group extraction means for extracting a group in which a plurality of pixels having high luminance information in a captured image are consecutive within a predetermined range, luminance information and photographic coordinate system of each pixel constituting the group Group center-of-gravity coordinate value calculating means for calculating the coordinate value of the center of gravity of the group based on the coordinate value at, a first reference point, a second reference point, and a first reference point in the captured image A first point sequence extracting means for extracting a first point sequence in which the same number of centroids as the total number of auxiliary points on the line are arranged on the same line; and a second reference point and a third in the photographed image. A second point sequence extracting means for extracting a second point sequence in which the same number of centroids as the total number of reference points and the number of auxiliary points on the second straight line are arranged on the same straight line; A selection unit that selects a combination of the first point sequence and the second point sequence sharing the center of gravity of one end, and a second point selected by the selection unit in the first point sequence selected by the selection unit The center of gravity corresponding to the end opposite to the center of gravity shared with the point sequence is identified as the first reference point, and the center of gravity shared by the first point sequence and the second point sequence selected by the selection means is the second. A first point sequence selected by the selection unit in the second point sequence selected by the selection unit The center of gravity corresponding to the end portion of the center of gravity and the opposite with and a third reference point specifying means for specifying the reference point.
[0009]
In the photogrammetry target, for example, two auxiliary points are provided on the first straight line, one auxiliary point is provided on the second straight line, and the first point sequence is on the same straight line in the photographed image. And the second point sequence is a three-point sequence in which three centroids are arranged on the same straight line.
[0010]
In addition, in the method for calculating the reference point of the photogrammetric target according to the present invention, the first reference point and the second reference point are provided with first, second, and third reference points whose distances are known. The first straight line that connects the second reference point and the second straight line that connects the second reference point and the third reference point have a predetermined angle, and the number of assists on the first straight line and the second straight line are different. A first step of photographing a surveying object with a camera together with a photogrammetric target provided with a point, and the number of auxiliary points on the first straight line and a second straight line in a photographed image photographed by the camera Based on the difference from the number of auxiliary points, the first reference point, the second reference point, and the third reference point are specified, and the first reference point in the photographic coordinate system that is a two-dimensional orthogonal coordinate in the photographed image. And a second step of acquiring coordinate values of the reference point, the second reference point, and the third reference point. To.
[0011]
The recording medium storing the reference point calculation program for the photogrammetric target according to the present invention is provided with first, second and third reference points whose distances are known, and the first reference point is provided. The first straight line connecting the point and the second reference point and the second straight line connecting the second reference point and the third reference point have a predetermined angle, and the first straight line and the second straight line The number of auxiliary points on the first straight line and the auxiliary points on the second straight line in the photographed image obtained by photographing the surveying object with the photogrammetry target on which the different number of auxiliary points are provided. The first reference point, the second reference point, and the third reference point are identified based on the difference in the number of the first reference point, the first reference point in the photographic coordinate system that is a two-dimensional orthogonal coordinate in the photographed image, the first reference point, A two-dimensional coordinate acquisition routine for acquiring the coordinate values of the second reference point and the third reference point is provided. To.
[0012]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 shows a partially broken view of a photogrammetry target applied to an embodiment according to the present invention. FIG. 2 shows a side view of the photogrammetric target. The target 10 has an L-shape and includes two columnar members 12 and 14. The first columnar member 12 and the second columnar member 14 are made of a metal material, have a hollow rectangular column shape inside, and a non-reflective sheet is pasted on the entire outer peripheral surface thereof. Both the first columnar member 12 and the second columnar member 14 have the same length L. _w And each thickness is the same length L _H It is.
[0013]
A rectangular parallelepiped control unit housing 20 is integrally fixed to one end 12 a of the first columnar member 12. The control unit housing 20 is made of a metal material, and a non-reflective sheet is attached to the entire outer peripheral surface thereof. In the control unit housing 20, the thickness is the same length L as the thickness of the first columnar member 12. _H And the width thereof is the width L of the first columnar member 12. _w About twice as much. The control unit housing 20 is provided such that the side surface 20b thereof is located on the same plane as the side surface 12b of the first columnar member 12, and the side surface 20c of the control unit housing 20 is the side surface of the first columnar member 12. It protrudes from 12c.
[0014]
One end portion 14 a of the second columnar member 14 is rotatably attached to the side surface 20 c of the control unit housing 20 by a hinge 15. The side surface 14b of the second columnar member 14 is located on the same plane as the end surface 20d on the opposite side of the surface on which the first columnar member 12 of the control unit housing 20 is provided.
[0015]
The angle α formed by the side surface 14c of the second columnar member 14 and the side surface 20c of the control unit housing 20, that is, the axial centers A and B of the two columnar members 12 and 14 (indicated by a two-dot chain line in the figure). A stay 16 that is a fixing member is connected to the acute angle side that is formed, whereby the two columnar members 12 and 14 are fixed to each other. The width and thickness of the stay 16 is the width L of the first and second columnar members 12 and 14. _w And thickness L _H Smaller than. The length of the stay 16 in the longitudinal direction is shorter than the length of the first columnar member 12 in the longitudinal direction.
[0016]
The stay 16 is provided so as to be inclined with respect to the two columnar members 12, 14. At this time, the acute angle α formed with respect to the first columnar member 12 of the second columnar member 14 is 90 degrees. The stay 16 is pivotally fixed to the first columnar member 12 by a stay hinge 92, and is detachable from the second columnar member 14 by a lock hinge 94.
[0017]
Three reference point members 31, 34, 36 and three auxiliary point members 32, 33, 35 are the same on the upper surface of the target 10, that is, the upper surfaces of the two columnar members 12, 14 and the control unit housing 20. Provided on a plane. The reference point member 31 (first reference point) and auxiliary point members 32 and 33 are provided on the upper surface 12 e of the first columnar member 12, and the reference point member 34 (second reference point) is provided on the control unit housing 20. The auxiliary point member 35 and the reference point member 36 (third reference point) are provided on the upper surface 20 e of the second columnar member 14. That is, two auxiliary point members 32 and 33 are provided on a straight line (first straight line) connecting the reference point member 31 and the reference point member 34, and a straight line connecting the reference point member 34 and the reference point member 36 (second line). The auxiliary point member 35 is provided on the straight line).
[0018]
Each reference point member 31, 34, 36 and each auxiliary point member 32, 33, 35 have a disk shape, all have the same diameter, and the width L of the columnar members 12, 14. _w Smaller than.
[0019]
The reference point members 31 and 34 and the auxiliary point members 32 and 33 are provided on a straight line (first straight line) connecting the reference point member 31 and the reference point member 34 parallel to the axis A direction. 36 and the auxiliary point member 35 are provided on a straight line (second straight line) connecting the reference point member 34 and the reference point member 36 parallel to the axis B direction. That is, two auxiliary point members 32 and 33 are provided on a straight line connecting the reference point member 31 and the reference point member 34, and one auxiliary point member 35 is provided on a straight line connecting the reference point member 34 and the reference point member 36. And the number of auxiliary point members arranged on each straight line is different. The distances between the reference point member 31 and the auxiliary point member 32, between the auxiliary point member 32 and the auxiliary point member 33, and between the auxiliary point member 33 and the reference point member 34 are the same. , The distance between the auxiliary point member 35 and the reference point member 36 is equal. Further, the distance from the reference point member 31 to the reference point member 34 is equal to the distance from the reference point member 34 to the reference point member 36.
[0020]
The reference plane members 31, 34, 36 and the auxiliary point members 32, 33, 35 define a photogrammetric reference plane, and at the same time, the side length of an isosceles triangle having the reference point members 31, 34, 36 as vertices is determined. It is defined as a reference scale. That is, the distance from the reference point member 31 to the reference point member 34, the distance from the reference point member 34 to the reference point member 36, or the distance from the reference point member 36 to the reference point member 31 is already known. Used as a measure for photogrammetry.
[0021]
As is apparent from FIG. 1, the number of auxiliary point members is different on two sides having the same length of an isosceles triangle, so that the orientation of the target 10 can be easily determined, and there are many captured images with the same subject. In addition, the camera position can be easily assumed.
[0022]
The reference plane defined by the reference point members 31, 34, 36 and the auxiliary point members 32, 33, 35 is preferably parallel to the road surface. The control unit housing 20 is provided on the first columnar member 12 side, and when only the end of the second columnar member 14 is connected by the hinge 15, the second columnar member 12 side is heavy and the second columnar member 12 side is heavy. The columnar member 14 floats from the road surface, and the reference point member 31 and auxiliary point members 32 and 33 on the first columnar member 12 side, and the auxiliary point member 35 and reference point member 36 on the second columnar member 14 side are on the same plane. There is a possibility that it cannot be located above and the reference plane is not parallel to the road surface.
[0023]
For this reason, the middle of the columnar members 12 and 14 are connected to each other by the stay 16 to prevent the second columnar member 14 from rising from the road surface, and the reference plane is parallel to the road surface to improve the accuracy of photogrammetry. It is improving.
[0024]
Further, a sheet-like elastic member 19 (see FIG. 5) is provided in the gap between the side surface 20c of the control unit housing 20 and the end surface 14a of the second columnar member 14 generated by the attachment of the hinge 15. The backlash of the columnar member 14 is prevented. The elastic member 19 is formed of rubber, sponge, or the like, and is closely fixed to the end surface 14 a of the second columnar member 14. A spring member may be provided instead of the sheet-like elastic member 19.
[0025]
Reflective sheets are affixed to the reference point members 31, 34, 36 and the auxiliary point members 32, 33, 35, and the reference point members 31, 34, 36 and the auxiliary point members 32, 33, 35 are black. Non-reflective members 41, 42, 43, 44, 45, and 46, which are disc members, are provided. Thereby, the reference point members 31, 34, and 36 and the auxiliary point members 32, 33, and 35 in the captured image can be easily identified, and the accuracy of photogrammetry can be improved.
[0026]
The target 10 includes three legs 18 on the surface opposite to the surface on which the reference point members 31, 34, 36 and the auxiliary point members 32, 33, 35 are provided. The legs 18 are provided corresponding to the reference point members 31, 34, and 36, so that the target 10 is placed away from the road surface by the length of the legs 18.
[0027]
The configuration of the auxiliary point member 35 and the non-reflecting member 45 will be described with reference to FIGS. 3 is a cross-sectional view of the target 10 taken along the line VIII-VIII in FIG. FIG. 4 is a plan view showing a surface of the non-reflective member 45 on the second columnar member 14 side. The other reference point members 31, 34, 36, auxiliary point members 32, 33 and non-reflective members 41, 42, 43, 44, 46 have the same configuration as the auxiliary point member 35 and non-reflective member 45, so here Description is omitted.
[0028]
A magnet holding member 62 is provided on the upper surface 14 e of the second columnar member 14, and an annular magnet 64 is accommodated in the magnet holding member 62. The outer diameter of the magnet holding member 62 is the width L of the second columnar member 14. _w Is approximately the same length. The magnet 64 is fixed together with the magnet holding member 62 to the second columnar member 14 by a screw member 66. A reflective sheet 68 is attached to the head 67 of the screw member 66. The auxiliary point member 35 is configured by the magnet holding member 62, the magnet 64, the screw member 66, and the reflection sheet 68.
[0029]
The non-reflective member 45 includes a disk 72 made of a material that can transmit radio waves, for example, resin or rubber. When the material of the disc 72 is rubber, the non-reflective member 45 is prevented from being damaged by dropping. A non-reflective sheet 74 is attached to one surface of the disc 72. In this embodiment, the diameter of the nonreflective member 45 is about 7 times the diameter of the auxiliary point member 35, that is, the screw member head 67. Further, the thickness of the non-reflective member 45 is slightly smaller than the thickness of the head 67 of the screw member 66.
[0030]
A fitting hole 76 having substantially the same diameter as the head 67 of the screw member 66 is formed in the center of the non-reflective member 45. An annular iron plate 78 is embedded around the fitting hole 76 on the surface of the nonreflective member 45 where the nonreflective sheet 74 is not provided. The inner diameter of the iron plate 78 is substantially equal to the diameter of the fitting hole 76, and the outer diameter thereof is substantially equal to the outer diameter of the magnet holding member 62.
[0031]
The non-reflective member 45 is detachable from the auxiliary point member 35. When the target 10 is used, the head 67 of the screw member 66 and the fitting hole 76 are fitted, and at this time, the iron plate 78 is closely fixed to the head 67 of the screw member 66 or the magnet holding member 62 by the magnetic force of the magnet 64. Is done. As is apparent from FIG. 3, in the state where the non-reflective member 45 is attached to the auxiliary point member 35, the reflective sheet 68 and the non-reflective sheet 74 are substantially on the same plane. When the target 10 is not used, the non-reflective member 45 is manually detached from the auxiliary point member 35.
[0032]
Thus, by making the non-reflective member 45 detachable from the auxiliary point member 35, the portability of the target 10 is improved. Further, by providing the non-reflective sheet 74 around the reflective sheet 68, the area of the auxiliary point member 35 becomes clear, and shooting at a dark place due to nighttime or rain, or in a place where the road surface is likely to be reflected. Even when shooting conditions such as shooting are poor, it is easy to identify the auxiliary point member 35 in the shot image.
[0033]
In addition, the ratio of the diameter of the auxiliary point member 35 and the non-reflective member 45, that is, the size of the region of the reflective sheet 68 and the non-reflective sheet 74 is not particularly limited to this embodiment. 68 should just be a magnitude | size which can fully recognize. Further, the shapes of the auxiliary point member 35 and the non-reflective member 45 are not limited to a circle.
[0034]
FIG. 5 is an enlarged plan view showing the vicinity of the control unit housing 20 in FIG. 1 and is a partially broken view. FIG. 6 is a cross-sectional view taken along line XI-XI in FIG. 5, and is a diagram showing a simplified configuration of the control unit housing 20.
[0035]
A battery storage chamber 83 is provided on the end surface 20 d side of the control unit housing 20. A battery 87 serving as a power source is stored in the battery storage chamber 83 and supplies power to the target 10. The battery storage chamber 83 has an opening on the end face 20d side and is closed by a lid portion 83a. A switch 85 is integrally provided on the end surface 20 d of the control unit housing 20, and the power is turned on and off by manual operation of the switch 85. When the power is turned on, various sensors such as an inclination angle sensor and an azimuth sensor (not shown) are activated, and position information of the target 10 is obtained.
[0036]
The upper surface 20 e of the control unit housing 20 has an opening 81, and the opening 81 is closed by a cover 82. The cover 82 is made of a material that can transmit radio waves, such as a resin.
[0037]
Such a target 10 is photographed by a camera after being arranged at a predetermined position in a surveying site.
[0038]
Next, a procedure for calculating the camera position of the present embodiment using the target 10 will be described. FIG. 7 is a diagram illustrating the positional relationship between the target 10 and the camera 100. The target 10 is photographed at the camera position M by the camera 100. The camera position M is defined as the rear principal point position of the taking lens of the camera 100. The optical axis at the camera position M is indicated by a two-dot chain line O. As described above, the three reference point members and the three auxiliary point members are provided on the target 10, but only the three reference point members that are the vertices are used for explanation in order to avoid complication of the drawing. In FIG. 7, the plane determined by the reference point members 31, 34, and 36 is a reference plane, the distance between the reference point member 31 and the reference point member 34, and the reference point member 34 and the reference point member 36. The distance between them is shown as a reference scale L.
[0039]
As is apparent from FIG. 7, the camera 100 is equipped with an IC card 24 that is a readable recording medium. The IC card 24 is detachable from the camera 100. The photographed image data is recorded on the IC card 24. The photogrammetry based on the image data is performed after the IC card 24 is mounted on a photogrammetry apparatus described later and the image data recorded on the IC card 24 is loaded into the memory of the photogrammetry apparatus.
[0040]
FIG. 8 shows a captured image at the camera position M. The photographed image has a two-dimensional orthogonal coordinate system (Xp _, Yp) is set, and the coordinate origin is set as the imaging center CP. As is clear from FIG. 8, in the captured image, the image points of the reference point members 31, 34, and 36 are represented by coordinates P1 (Xp ₁ , Yp ₁ ), P2 (Xp ₂ , Yp ₂ ), P3 (Xp _Three , Yp _Three ). In the present specification, the two-dimensional orthogonal coordinates of the photographed image are referred to as “photographic coordinate system”.
[0041]
FIG. 9 is a diagram relatively showing a positional relationship between an image captured by the camera 100 and the target 10. A reference plane defined by the reference point members 31, 34 and 36 is indicated by a hatched area in the figure. In FIG. 9, the three-dimensional orthogonal coordinates of the camera position M are indicated by (Xc, Yc, Zc). The coordinate origin coincides with the rear principal point position of the taking lens of the camera 100 at the camera position M, and the Zc axis thereof coincides with the optical axis O of the taking lens at the camera position M. The Xc axis is parallel to the Xp axis of the two-dimensional orthogonal coordinate system of the captured image shown in FIG. 8, and the Yc axis is parallel to the Yp axis of the two-dimensional orthogonal coordinate system. In this specification, the three-dimensional orthogonal coordinates of the camera position M are referred to as a “camera coordinate system”.
[0042]
Here, in order to specify the camera position M based on the captured image, a right-handed three-dimensional orthogonal coordinate system (Xs, Ys, Zs) shown in FIG. 10 is appropriately set. The coordinate origin of the three-dimensional orthogonal coordinate system (Xs, Ys, Zs) is made to coincide with the reference point member 34 of the target 10. The Zs axis is a direction from the reference point member 34 to 36, passes through the reference point member 34, is orthogonal to the Zs axis and is on the reference plane as an Xs axis, passes through the reference point member 34, and enters the reference plane. An axis that is orthogonal, that is, an axis that is orthogonal to the paper surface is defined as a Ys axis. Since the distances between the reference point member 31 and the reference point member 34 of the target 10 and between the reference point member 34 and the reference point member 36 are L as described above, the image points of the reference point members 31, 34 and 36 are coordinates. Ps1 (-L, 0, 0), Ps2 (0, 0, 0), Ps3 (0, 0, L). In this specification, this three-dimensional orthogonal coordinate is referred to as a “scene coordinate system”.
[0043]
The camera position M is indicated by the relative relationship between the scene coordinate system and the camera coordinate system. That is, the movement distance of the origin of the camera coordinate system from the origin of the scene coordinate system, the rotation angle of the Xs axis around the Xc axis, the rotation angle of the Ys axis around the Yc axis, and the rotation angle of the Zs axis around the Zc axis Thus, the camera position M is defined.
[0044]
The coordinate value Psi (Psxi, Psyi, Pszi) (i = 1, 2, 3) of the reference point in the scene coordinate system and the coordinate value Pci (Pcxi, Pcyi, Pczi) (i = 1) of the reference point Pi in the camera coordinate system 2, 3) is expressed by equation (1).
[0045]
[Expression 1]

[0046]
In equation (1), α is the rotation angle of the Xs axis around the Xc axis, β is the rotation angle of the Ys axis around the Yc axis, γ is the rotation angle of the Zs axis around the Zc axis, and R is α, β, A rotation matrix of γ, Δ is a vector from scene coordinates to camera coordinates.
[0047]
Further, the coordinate value Pci (Pcxi, Pcyi, Pczi) (i = 1, 2, 3) of the reference point in the camera coordinate system and the coordinate value Di (Dxi, Dyi) (i = 1, of the reference point in the photographic coordinate system). The relationship with (2, 3) is expressed by equation (2). Note that “f” in equation (2) is the focal length of the taking lens of the camera 100.
[0048]
[Expression 2]

[0049]
As described above, the coordinate value in the photographic coordinate system is calculated from the coordinate value in the scene coordinate system of the reference point by the above-described equations (1) and (2). On the other hand, the coordinate value of the reference point in the photographic coordinate system is automatically extracted from the captured image as will be described later. Accordingly, by comparing the coordinate value of the reference point calculated from the equations (1) and (2) with the coordinate value automatically extracted from the photographed image, ΔX, ΔY, ΔZ, α, β defining the camera position. , Γ can be calculated.
[0050]
FIG. 17 is a block diagram of a coordinate calculation system of a photogrammetry apparatus in which photogrammetry is performed based on image data stored in the IC card 24. As shown in FIG. 17, the coordinate calculation system is configured by a computer system including a computer 184. The computer 184 includes a central processing unit (CPU) 184A, an operating system, a read only memory (ROM) 184B in which constants are stored, a write / readable memory (RAM) 184C in which data and the like are temporarily stored, and a CPU 184A. The hard disk 184D and the input / output interface (I / O) 184E for storing the calculation results executed in the above are included.
[0051]
The computer system also includes an IC card reader 186 connected to the computer 184 via the I / O 184E. The IC card reader 186 has a slot in which the IC memory card 24 is inserted, and has an IC card driver 188 for reading a digital image and other data for one frame.
[0052]
The computer system also includes a CD-ROM reader 197 connected to the computer 184 via the I / O 184E. The CD-ROM reader 197 includes a tray for mounting the CD-ROM 198 and has a CD-ROM drive unit 199. A target reference point and a camera position calculation program, which will be described later, are stored in the CD-ROM 198. This calculation program is read from the CD-ROM 198 and stored in the hard disk 184D.
[0053]
Further, the computer system can reproduce a captured image based on one frame of digital image data read from the IC memory card 24, a monitor 190 for reproducing a survey map created by the computer 184, and various commands. A keyboard 192 for inputting signals and data to the computer 184, a mouse 194 for operating a cursor displayed on the monitor 190, and a printer 196 for printing a survey map as necessary are provided.
[0054]
FIG. 11 is a flowchart showing the procedure for calculating the camera position of the present embodiment, which is executed by the CPU 184A. In step S1000, a digital image of the surveying site photographed by setting the above-described target 10 is loaded into the photogrammetry apparatus of the present embodiment, and then gray scale processing is performed in step S1002. Gray scale processing is processing that removes color components from a color image and extracts only luminance information. The luminance information of 256 gradations is extracted for every pixel constituting the digital image, and stored in the array format variable Img. That is, the luminance information of each pixel is represented by Img (j, k). j and k are two-dimensional coordinate values (coordinate values of a photographic coordinate system) in a digital image of a pixel.
[0055]
In step S1004, the luminance information extracted for each pixel is binarized. The luminance information stored in the variable Img is compared with a predetermined threshold, and if the luminance information is high and exceeds the threshold, “1” is set, and if the luminance information is low and below the threshold, “0” is set in the same arrangement format as Img. Stored in the variable Bin. That is, data obtained by binarizing the luminance information of each pixel is represented by Bin (j, k). As described above, the reflection sheet is affixed to the reference point members 31, 34, and 36 and the auxiliary point members 32, 33, and 35 of the target 10 used in the present embodiment, and the black non-reflective member 41 is disposed around the reflection sheet. , 42, 43, 44, 45, 46 are provided. Therefore, the binarized data of the pixels corresponding to the reference point members 31, 34, 36 and the auxiliary point members 32, 33, 35 are “1”, and the surrounding non-reflective members 41, 42, 43, 44, 45, 46 The binarized data of the pixel corresponding to “0” is “0”.
[0056]
Next, labeling is performed in step S1006. Labeling is a process of extracting, as a group, a portion in which pixels having binarized data “1” are continuous, and a region having high luminance information is extracted from a digital image. For each group, the number of pixels constituting the group and the respective two-dimensional coordinate values are stored in a predetermined variable.
[0057]
Next, in step S1008, the area of each group, that is, the number of pixels is calculated. In step S1010, it is determined whether or not the group corresponds to a reference point member or an auxiliary point member based on the number of pixels in each group. Specifically, it is checked whether or not the number of pixels is within the range between the lower limit value Th1 and the upper limit value Th2. The lower limit value Th1 and the upper limit value Th2 are such that the range of the number of pixels defined thereby sufficiently includes the number of pixels of the group corresponding to the reference point member and auxiliary point member in the digital image, and is the minimum range. Is set. Accordingly, the processing in step S1010 corresponds to a group having about 1 or 2 pixels corresponding to the noise of the digital image, or a region having high luminance information over a wide area such as paint applied on the road of the surveying site. Group to be removed.
[0058]
Next, in step 1012, the center of gravity is calculated in consideration of the luminance information of each group. FIG. 12 is a flowchart illustrating a processing procedure for calculating the center of gravity. In step S1022, variables SumX, SumY, and SumImg are initialized. The variable SumX stores the product of the luminance information of each pixel constituting the group and the x coordinate value, and the variable SumY stores the product of the luminance information of each pixel constituting the group and the y coordinate value. Accumulated and stored. In addition, the variable SumImg stores the sum obtained by adding the luminance information of each pixel constituting the group.
[0059]
In step S1026, the product of the luminance information (Img (j, k)) of the pixel and the x coordinate value (j) of the pixel is calculated, added to the value stored in the variable SumX at that time, and then added to the variable SumX. Stored. Next, in step S1028, the product of the luminance information of the pixel and the y coordinate value (k) of the pixel is calculated, added to the value stored in the variable SumY at that time, and stored in the variable SumY. In step S1030, the luminance information of the pixel is added to the value stored in the variable SumImg at that time, and stored in the variable SumImg. The processes in steps S1026 to S1030 described above are repeated for all the pixels constituting the group.
[0060]
That is, the variable SumX stores the sum of values obtained by weighting the luminance information of each pixel based on the coordinate value in the x direction, and the variable SumY stores the luminance information of each pixel based on the coordinate value in the y direction. Stores the sum of weighted values. The variable SumImg stores the sum of luminance information of each pixel.
[0061]
In step S 1032, it is checked whether or not the above-described luminance information accumulation calculation processing has been performed on all the pixels constituting the group. When the processing of steps S1026 to S1030 is performed for all the pixels constituting the group, the process proceeds to step S1034. In step S1034, a process of dividing the sum of values obtained by weighting the luminance information of each pixel based on the coordinate value in the x direction by the sum of the luminance information of each pixel is performed, and the x coordinate value of the center of gravity of the group is calculated. Is done. In step S1036, a process of dividing the sum of values obtained by weighting the luminance information of each pixel based on the coordinate value in the y direction by the sum of the luminance information of each pixel is performed, and the y coordinate value of the center of gravity of the group is calculated. Is done.
[0062]
Next, in step S1038, it is checked whether or not the coordinates of the center of gravity in consideration of the luminance information have been calculated for all groups in the captured image. If the calculation processing of the center-of-gravity coordinates for all the groups has not been completed, the processing from step S1020 to S1036 is repeated. When the calculation process of the coordinates of the center of gravity for all the groups is completed, the process proceeds to step S1014 in FIG.
[0063]
FIG. 13 is a diagram schematically showing the state of each pixel of the captured image after the calculation of the barycentric coordinates of each group. Pixels with low luminance information are hatched. As illustrated in FIG. 13, groups G1 to G13 are extracted, and the respective centers of gravity are GC1 to GC13.
[0064]
In step S1014, a group combination that satisfies the relative relationship between the reference point members 31, 34, and 36 and the auxiliary point members 32, 33, and 35 of the target 10 is extracted. FIG. 14 is a flowchart showing the first half of the group combination extraction process, and FIG. 15 is a flowchart showing the second half. Hereinafter, the procedure of the group combination extraction process will be described with reference to FIG.
[0065]
In step S1040, two arbitrary groups are selected from the groups in the image. In step S1042, a straight line including the center of gravity of the selected two groups is calculated. Next, in step S1044, a group having a centroid between the centroids of the two selected groups is detected on the straight line.
[0066]
In step S1046, the number of groups detected in step S1044 is checked. If there are four groups having the center of gravity on the straight line including the two groups selected in step S1040, the process proceeds to step S1048. If three groups, the process proceeds to step S1050. Otherwise, the process proceeds to step S1052.
[0067]
For example, when the group G1 and the group G2 are selected, since the centroids of other groups do not exist on the straight line including the centroids GC1 and GC2, the process proceeds to step S1052. When the group G3 and the group G6 are selected, or when the group G10 and the group 13 are selected, two centroids of the centroid GC4 of the group G4 and the centroid GC5 of the group G5 exist on the straight line including the centroids GC3 and GC6. On the straight line including the centroids GC10 and GC13, there are two centroids, the centroid GC11 of the group G11 and the centroid GC12 of the group G12. Accordingly, the process proceeds to step S1048. When the group G1 and the group G3 are selected, or when the group G7 and the group G9 are selected, one centroid GC2 of the group G2 exists on the straight line including the centroids GC1 and GC3, and includes the centroids GC7 and GC9. There is one center of gravity GC8 of the group G8 on the straight line. Accordingly, the process proceeds to step S1050.
[0068]
In step S1048, the four groups are registered as a four-point sequence parallel on the same straight line, and in step S1050, the three groups are registered as a three-point sequence parallel on the same straight line. For example, in each group in FIG. 13, groups G1, G2, and G3, and groups G6, G7, and G8 are registered as a three-point sequence, and groups G4, G5, G6, and G7, and groups G10, G11, G12, and G13 are registered. Each is registered as a 4-point sequence.
[0069]
In step S1052, it is checked whether or not processing has been performed for all combinations of the two groups. If processing has not been completed for all combinations, the processing in steps S1040 to S1050 is repeated. When the processing is completed for all combinations of the two groups, the process proceeds to step S1054 in FIG.
[0070]
In step S1054, a three-point sequence and a four-point sequence that are on the same straight line are selected one by one, and it is checked whether the centroids of the groups corresponding to the respective end points match. If the centroids of the groups corresponding to the respective end points match, the process proceeds to step S1058. If the centroids of the groups corresponding to the end points do not match, the process returns to step S1054 to select another combination of the three-point string and the four-point string. . In FIG. 13, a group G3 is an end point of a three-point sequence of groups G1, G2, and G3, and an end point of a 4-point sequence of groups G3, G4, G5, and G6. Therefore, groups G1, G2, and G3 and groups G3, G4, G5, and G6 are selected as combinations of the three-point string and the four-point string.
[0071]
In step S1058, the coordinates of the center of gravity of the group corresponding to the end point opposite to the end point that coincides with the end point of the three-point sequence among the groups constituting the 4-point sequence are the variables A1 (Ax) indicating the coordinates of the reference point member 31. ₁ , Ay ₁ ) And the coordinates of the center of gravity of the group corresponding to the end points where the three-point string and the four-point string coincide are the variables A2 (Ax ₂ , Ay ₂ ) And the coordinates of the center of gravity of the group corresponding to the end point on the opposite side of the end point that coincides with the end point of the 4-point sequence among the groups constituting the 3-point sequence are variables A3 ( Ax _Three , Ay _Three ). That is, the coordinates of the center of gravity GC6 in FIG. 13 are stored in the variable A1, the center of gravity GC3 is stored in the variable A2, and the center of gravity GC1 is stored in the variable A3.
[0072]
Next, in step S1060, the angle formed by the vector from variable A2 to variable A3 with respect to the vector from variable A2 to variable A1 is checked. It is checked whether the counterclockwise angle θ formed by the vector from the variable A2 to the variable A3 with respect to the vector from the variable A2 to the variable A1 is larger than the lower limit value Th3 and smaller than the upper limit value Th4. In the present embodiment, the lower limit value Th3 is set to 0 degrees, and the upper limit value Th4 is set to 180 degrees. If the condition of step S1060 is satisfied, the combination of the selected 3-point sequence and 4-point sequence satisfies the relative relationship of the three reference point members 31, 34, 36 and auxiliary point members 32, 33, 35 of the target 10. Therefore, the group combination extraction process ends, and the process proceeds to step S1016 in FIG.
[0073]
If the angle θ is not within the range between the lower limit value Th3 and the upper limit value Th4, the process proceeds to step S1062, and it is checked whether or not processing has been performed for all combinations of the three-point string and the four-point string. If there is a combination of a three-point sequence and a four-point sequence for which the processing from steps S1056 to S1060 has not been performed, the process returns to step S1054 and the processing up to step S1060 is repeated.
[0074]
That is, the processes of steps S1054 to S1062 are repeated until a combination of a three-point string and a four-point string satisfying the relative relationship between the reference point members 31, 34, 36 and the auxiliary point members 32, 33, 35 is detected. When a combination of a three-point string and a four-point string that satisfies the relative relationship between the reference point member and the auxiliary point member is detected, the process ends without performing the other combinations of the three-point string and the four-point string. .
[0075]
The combination of the three-point sequence and the four-point sequence in FIG. 13, the groups G1, G2, G3, and the groups G3, G4, G5, G6, the inverse of the vector from the variable A2 to the variable A3 with respect to the vector from the variable A2 to the variable A1 As is apparent from FIG. 13, the clockwise angle is within the range between the lower limit value Th3 and the upper limit value Th4. Accordingly, the groups G1, G2, and G3 and the groups G3, G4, G5, and G6 are a three-point row and a four-point row that satisfy the relative relationship between the reference point members 31, 34, and 36 and the auxiliary point members 32, 33, and 35. Are detected as a combination.
[0076]
In step S1016, camera position calculation processing is performed. FIG. 16 is a flowchart illustrating a processing procedure for camera position calculation. In step S1070, the variables ΔX, ΔY, ΔZ, α, β, and γ in the above equation (1) are initialized. Next, in step S1072, (−L, 0, 0), (0, 0, 0), (0, 0, L) are substituted for Ps1, Ps2, and Ps3 in the equation (1), respectively, and a variable ΔX, Substituting appropriate values into ΔY, ΔZ, α, β, γ, the coordinate values Pc1 (Pcx) of the reference point members 31, 34, 36 in the camera coordinate system ₁ , Pcy ₁ , Pcz ₁ ), Pc2 (Pcx ₂ , Pcy ₂ , Pcz ₂ ), Pc3 (Pcx _Three , Pcy _Three , Pcz _Three ) Is calculated.
[0077]
In step S1074, the coordinate values of Pc1, Pc2, and Pc3 are substituted into the above equation (2), and the coordinate values D1 (Dx (Dx) of the reference point members 31, 34, and 36 in the photographic coordinate system are substituted. ₁ , Dy ₁ ), D2 (Dx ₂ , Dy ₂ ), D3 (Dx _Three , Dy _Three ) Is calculated.
[0078]
In step S1076, the coordinates calculated using the coordinate values A1, A2, A3 in the photographic coordinate system of the reference point members 31, 34, 36 automatically extracted from the captured image, and the above-described equations (1) and (2). The values D1, D2, and D3 are compared. That is, the absolute value of the difference between the coordinate values D1, D2, and D3 of each reference point member and A1, A2, and A3 is calculated, and the total value S is calculated.
[0079]
In step S1078, it is confirmed whether or not the total value S is the minimum. If not, the values of the variables ΔX, ΔY, ΔZ, α, β, and γ are changed, and the processing of steps S1072 to S1076 is repeated. That is, the variables ΔX, ΔY, ΔZ, α, β, γ that minimize the total value S are calculated by the least square method.
[0080]
If it is confirmed in step S1078 that the total value S is minimum, the parameters ΔX, ΔY, ΔZ, α, β, and γ when the total value S is minimum in step S1080 are parameters indicating the camera position M. And the camera position calculation process ends.
[0081]
As described above, according to the present embodiment, the number of auxiliary point members provided between the reference point members 31 and 34 and the number of auxiliary point members provided between the reference point members 34 and 36 in the target 10 are different. In addition, since the reflection sheet is affixed to each reference point member and each auxiliary point member, it is easy to specify the positions of each reference point member and auxiliary point member from the luminance information in the captured image. Accordingly, the coordinate values of the reference point members 31, 34, and 36 in the captured image are automatically extracted.
[0082]
Further, according to the present embodiment, the length between the reference point members 31 and 34 of the target 10 and the length between the reference point members 34 and 36, and the first columnar member 12 and the second columnar member 14. Since the angle is known, the coordinate value of each reference point member in the scene coordinate system is set to a predetermined value, and the coordinate value of each reference point member in the photographic coordinate system can be calculated based on the coordinate value. Therefore, the camera position can be automatically calculated by comparing with the coordinate value of the reference point member automatically extracted in the captured image.
[0083]
11, 12, 14, 15 and 16, the digital image of the surveying site is displayed on the monitor, the reference point of the target and an arbitrary surveying point are designated on the monitor, and a predetermined surveying calculation is performed. In the photogrammetry system in which the above and the like are executed, the calculation of the two-dimensional coordinates of the target reference point and the calculation of the camera position are executed. That is, according to the present embodiment, the operation of designating the reference point of the target in the digital image displayed on the monitor with the mouse is omitted, and accurate camera position data independent of the skill level of the person in charge is obtained. It is always obtained, and high-precision arithmetic processing is executed in the subsequent processing steps in the photogrammetry system.
[0084]
【The invention's effect】
As described above, according to the present invention, the camera position can be automatically calculated in the photogrammetry.
[Brief description of the drawings]
FIG. 1 is a plan view showing a photogrammetry target applied to an embodiment of the present invention.
FIG. 2 is a side view of the photogrammetric target shown in FIG.
FIG. 3 is a cross-sectional view of the target taken along the line VIII-VIII in FIG. 1;
4 is a plan view showing a surface on the second columnar member side of the non-reflective member of the target shown in FIG. 1; FIG.
FIG. 5 is an enlarged plan view showing the vicinity of the control unit housing of the target shown in FIG.
6 is a cross-sectional view taken along line XI-XI in FIG. 5, and is a diagram showing a simplified configuration of a control unit housing.
FIG. 7 is a perspective view showing a positional relationship between a photogrammetry target and a camera according to an embodiment of the present invention.
8 is a diagram schematically showing an image photographed by the camera of FIG. 7. FIG.
FIG. 9 is a diagram illustrating a positional relationship between a reference point member, its image point, and a rear principal point position of a photographing lens of a camera in three-dimensional coordinates.
FIG. 10 is a diagram relatively showing a scene coordinate system and a camera coordinate system.
FIG. 11 is a flowchart illustrating a procedure of camera automatic calculation processing according to the embodiment of this invention.
FIG. 12 is a flowchart showing a procedure for calculating a center of gravity of a group having high luminance information in a captured image.
FIG. 13 is a diagram schematically illustrating a state of each pixel of a captured image after calculation of the barycentric coordinates of the group.
FIG. 14 is a flowchart showing the first half of group combination extraction processing;
FIG. 15 is a flowchart showing the second half of the group combination extraction processing;
FIG. 16 is a flowchart illustrating a processing procedure for camera position calculation.
FIG. 17 is a block diagram of a coordinate calculation system of the photogrammetry apparatus.
[Explanation of symbols]
10 Target
12 First columnar member
14 Second columnar member
15 Hinge
16 stays
20 Control unit housing
24 IC card
31, 34, 36 Reference point member
32, 33, 35 Auxiliary point member
41, 42, 43, 44, 45, 46 Non-reflective member
92 Stay hinge
94 Lock Hinge
184 computer
186 IC card reader
188 IC card driver
197 CD-ROM reader
198 CD-ROM
199 CD-ROM drive unit
190 Monitor
192 keyboard
194 mouse
196 Printer

Claims (5)

First, second, and third reference points whose distances are known are provided, the first straight line connecting the first reference point and the second reference point, the second reference point, and the Surveying together with a photogrammetric target having a predetermined angle with the second straight line connecting the third reference points, and a different number of auxiliary points provided on the first straight line and the second straight line, respectively. Image acquisition means for capturing an object with a camera;
By comparing the number of auxiliary points on the first straight line and the number of auxiliary points on the second straight line in the captured image obtained by the image acquisition means, the The positions of the first reference point, the second reference point, and the third reference point are specified, and the first reference point and the second reference point in a photographic coordinate system that is a two-dimensional orthogonal coordinate in the captured image. And a two-dimensional coordinate acquisition means for acquiring the coordinate value of the third reference point and a reference point calculation device for a photogrammetric target. In the photogrammetry target, the first reference point, the second reference point, the third reference point, and the auxiliary point have relatively higher luminance information than the other subjects in the captured image. Made of reflective members,
The two-dimensional coordinate acquisition means
The binarization processing means for comparing the luminance information of each pixel constituting the photographed image with a predetermined threshold and classifying each pixel into two types of pixels having high luminance information and pixels having low luminance information When,
A group extracting means for extracting a group in which a plurality of pixels having the high luminance information in the photographed image are consecutive within a predetermined range;
Group centroid coordinate value calculating means for calculating the coordinate value of the centroid of the group based on the luminance information of each pixel constituting the group and the coordinate value in the photographic coordinate system;
In the captured image, a first number of the centroids equal to the total number of the first reference point, the second reference point, and the number of auxiliary points on the first straight line are aligned on the same straight line. First point sequence extraction means for extracting a point sequence;
In the captured image, a second number of the centroids equal to the total number of the second reference point, the third reference point, and the number of the auxiliary points on the second straight line are arranged on the same straight line. Second point sequence extracting means for extracting the point sequence;
Selection means for selecting a combination of the first point sequence and the second point sequence sharing the center of gravity of one end in each point sequence;
In the first point sequence selected by the selection unit, the centroid corresponding to the end opposite to the centroid shared with the second point sequence selected by the selection unit is used as the first reference point. The center of gravity shared by the first point sequence and the second point sequence selected by the selection unit is specified as the second reference point, and the second point selected by the selection unit Reference point specifying means for specifying, as the third reference point, the center of gravity corresponding to the end opposite to the center of gravity shared with the first point sequence selected by the selection means in the point sequence of The apparatus for calculating a reference point of a target for photogrammetry according to claim 1. In the photogrammetry target, two auxiliary points are provided on the first straight line, one auxiliary point is provided on the second straight line, and the first point sequence in the photographed image. 3. The four-point sequence in which the four centroids are arranged on the same straight line, and the second point sequence is a three-point sequence in which the three centroids are arranged on the same straight line. Photogrammetry target reference point calculation device. First, second, and third reference points whose distances are known are provided, the first straight line connecting the first reference point and the second reference point, the second reference point, and the Surveying together with a photogrammetric target having a predetermined angle with the second straight line connecting the third reference points, and a different number of auxiliary points provided on the first straight line and the second straight line, respectively. A first step of photographing an object with a camera;
Based on the difference between the number of the auxiliary points on the first straight line and the number of the auxiliary points on the second straight line in the captured image taken by the camera, the first reference point, The second reference point and the third reference point are specified, and the first reference point, the second reference point, and the third reference point in a photographic coordinate system that is a two-dimensional orthogonal coordinate in the captured image. And a second step of acquiring the coordinate value of the reference point of the photogrammetric target. First, second, and third reference points whose distances are known are provided, the first straight line connecting the first reference point and the second reference point, the second reference point, and the Surveying together with a photogrammetric target having a predetermined angle with the second straight line connecting the third reference points, and a different number of auxiliary points provided on the first straight line and the second straight line, respectively. Based on the difference between the number of the auxiliary points on the first straight line and the number of the auxiliary points on the second straight line in the captured image obtained by photographing the object by the camera, the first reference point The second reference point and the third reference point are specified, and the first reference point, the second reference point, and the third reference point in a photographic coordinate system that is a two-dimensional orthogonal coordinate in the captured image. Photogrammetric target reference equipped with a two-dimensional coordinate acquisition routine to acquire the coordinate value of the reference point Recording medium, wherein the program of calculation is stored.

Images