JP2004272756A - Device for investigating congestion degree - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a congestion degree investigating device for reducing an error in a congestion degree investigation. <P>SOLUTION: The device retrieves an image by an image input means 20 whenever the image obtained by photographing a congestion degree investigation object part through the use of an ITV camera 10 is updated. A plurality of change detecting means 30 detect the change of the image, based on the input of the image at each small area. The times of detecting the existence of change are detected by a plurality of change times accumulating means 40. A local change per time ratio calculating means 50 acquires a local change per time ratio as a congestion degree, which is the rate of the accumulated times with the existence of change in a fixed period to the number of the whole images in the fixed period. An outputting/recording means 60 outputs and records the congestion degree. When the largeness of a change detection part with the change-detected small areas connected therein is smaller than a predetermined largeness, a change detecting means 30 determines that the image is not changed in the small areas included in the part, and excludes it from change times accumulation. Thus, an object smaller than an inspection object is prevented from being erroneously detected. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

で表現される。ここで、ｍはｍの拡張ベクトル、ＭはＭの拡張ベクトル、ｓはスカラで、Ｐは、
【数２】

で表される変換行列である。ここで、Ａはカメラ内部パラメータで、
【数３】

【００６３】
Ｒはθのみのカメラの回転行列で、
【数４】

であるから（１）式は、
【数５】

となり、これを展開すると、
【数６】

【数７】

【数８】

となる。
【００６４】
Ｘは監視対象の水平面からのカメラ高さで一定であるからｍとＭは一対一に対応し、
【数９】

【数１０】

で表される。
【００６５】
つまり（９）（１０）式は、ＩＴＶカメラ映像中の点（ｕ，ｖ）が監視対象の平面上の点（Ｚ，Ｙ）との投影変換の式である。例えば、図７に示す検出画像を平面座標系に座標変換して、図１０に示す検出画像を得ることができる。
【００６６】
基本構成要素のみからなる単純な混雑度調査装置では、混雑度の出力をＩＴＶカメラ映像の各小領域ごとに出力していたので、ＩＴＶカメラ１０の光軸が斜めの場合には、平面的な注目地点の混雑度を反映しない場合があったのに対して、第３の実施の形態の混雑度出力は、調査対象箇所の平面上の注目点ごとに混雑度を出力するので、調査対象箇所における混雑度を直接評価することが可能になる。
例えば交差点の映像を使った混雑度調査を地図と照合するとか、店舗内の映像を使った混雑度調査を店内配置図と照合するなどの作業が容易になる。
【００６７】
［第４の実施の形態］
本発明の第４の実施の形態では、前述した第３の実施の形態に係る混雑度調査装置を備え、この複数セットの混雑度測定装置を使用し、複数カメラによる混雑度を平面座標系で統合して出力するものである。
【００６８】
即ち、第４の実施の形態に係る混雑度調査装置では、大きな調査対象地域に対して第３の実施の形態に係る混雑度調査装置を複数セット使用し、各混雑度測定装置より得られる各ＩＴＶカメラの撮像範囲に対応する平面上の各地点ごとの混雑度を、一つの大きな平面図に統合して出力するものである。
【００６９】
例えば、交差点を俯瞰するカメラを対向する角にそれぞれ１台ずつ設置してそれぞれに対応する第３の実施の形態の混雑度調査装置を構成し、これらより得られる各ＩＴＶカメラの撮像範囲に対応する平面上の各地点ごとの混雑度を一つの交差点の平面図に統合して出力する。
【００７０】
これにより、混雑度調査装置は複数のＩＴＶカメラを使用して広い領域の監視を行うような場合にもデータの一元管理が可能となる効果があり、さらに１台のカメラでは死角となっていた地点を他のカメラで調査することにより高範囲にわたる地域の混雑度調査が一つの平面図で実施できる利点がある。
【００７１】
例えば、図１１は交差点（十字路）の角に設置した１台の監視カメラ（ＩＴＶカメラ１０）の撮像範囲（台形部分）で交差点の全体を監視できないのに対して、図１２のように２台の監視カメラ（ＩＴＶカメラ１０，１０）を対向する角に設置した場合の監視エリアは交差点の全体をカバーする。
【００７２】
［第５の実施の形態］
本発明の第５の実施の形態では、複数ビデオ映像による混雑度を平面座標系で統合して出力するものである。
【００７３】
第５の実施の形態では、第４の実施の形態における複数セットの「第３の実施の形態の混雑度調査装置」に代えて、複数のＩＴＶカメラに対応する映像記録装置を使用して記録し、これを１台の混雑度調査装置で順次再生入力して個々の撮影範囲における混雑度調査出力を一つの平面に統合して出力するものである。
【００７４】
これにより、第４の実施の形態と同一の効果が得られるのに加えて、構成は混雑度調査装置が１セットになるので装置の節約が図れる利点がある。
【００７５】
［第６の実施の形態］
本発明の第６の実施の形態は、旋回動作するカメラによる混雑度をパノラマ画像で統合して出力するものである。
【００７６】
第６の実施の形態に係る混雑度調査装置では図１に示す基本構成要素を使用するが、その際に次のような撮影動作及び画像処理動作を行う。即ち、ＩＴＶカメラ１０は水平面内で旋回動作する雲台の上に設置して使用し、雲台の動作はあらかじめ設定する複数の地点で一定時間ずつ停止しながら旋回動作を繰り返す。変化検出手段３０と変化回数積算手段４０は雲台が停止している時間内に停止位置ごとに個別に小領域ごとの変化検出と変化回数積算を行う。局所変化時間率計算手段５０は停止位置ごとに個別に小領域ごとの局所変化時間率を計算し、出力・記録手段６０はこれら停止位置ごとの局所変化時間率を一つのパノラマ画像に統合して混雑度として出力または記録する。図１３はパノラマ画像の結合を説明する図である。
【００７７】
１台のＩＴＶカメラ１０を旋回させて広い領域の監視を行う目的で設置された監視カメラを使用して監視地域の混雑度調査を行う場合に、従来方法ではＩＴＶカメラの映像自体が変化するので、映像中の変化領域を検出して混雑度とする方法が適用できなかったが、第６の実施の形態の方法によれば監視カメラの動作範囲の全体をカバーする地域における混雑度を調査可能である。
【００７８】
また、停止位置ごとのＩＴＶカメラ映像を一つに統合してパノラマ画像を生成しておき、これに本実施例の混雑度調査結果出力をオーバーレイ表示すれば、注目地点と混雑度の対応が直感的に理解しやすい。
【００７９】
［第７の実施の形態］
本発明の第７の実施の形態では、旋回動作するカメラによる混雑度を平面座標系で統合して出力する物である。
【００８０】
第７の実施の形態では、第６の実施の形態の混雑度調査装置において、出力・記録手段６０は停止位置ごとの調査対象箇所の平面に投影変換し、これらより得られる各停止位置の撮像範囲に対応する平面上の各地点ごとの混雑度を一つの大きな平面図に統合して出力するものである。
【００８１】
これにより、監視カメラの撮像範囲の全体をカバーする地域における混雑度を調査可能であるとともに、調査対象箇所における混雑度を平面図と照合して直接評価することが可能になる。また、第４，第５の実施の形態の複数のカメラを使用する方法と比べて使用するカメラが１台なので構成が簡単である。
【００８２】
［第８の実施の形態］
本発明の第８の実施の形態は、旋回動作する複数のカメラによる混雑度を平面座標系で統合して出力するものである。
【００８３】
第８の実施の形態では、第７の実施の形態の混雑度調査装置を複数セット使用し、これらより得られる各ＩＴＶカメラの撮像範囲に対応する平面上の各地点ごとの混雑度をさらに一つの大きな平面図に統合して出力するものである。
【００８４】
これにより、複数の監視カメラの撮像範囲の全体をカバーするさらに広い地域における混雑度を調査可能であるとともに、調査対象箇所における混雑度を平面図と照合して直接評価することが可能になる。
【００８５】
また、第４，第５の実施の形態の複数のカメラを使用する方法と比べれば使用するカメラが少なくてすむ利点があるとともに、第６，第７の実施の形態の１台の旋回カメラを使用する方法では死角となっていた地点を他の旋回カメラで調査することにより高範囲にわたる地域の混雑度調査が一つの平面図で実施できる利点がある。
【００８６】
［第９の実施の形態］
本発明の第９の実施の形態は、混雑度結果を部分表示で出力するものである。
【００８７】
第９の実施の形態では、第１〜第８の実施の形態に係る混雑度調査装置における混雑度の出力・記録手段６０において、局所変化時間率が同一の小領域に対応する領域を同一の色で描画するいわゆるカラーチャート表示とするときに、対応する領域の一部分のみを描画してＩＴＶカメラ入力画像（ＩＴＶカメラ１０で得た画像のこと）またはＩＴＶカメラ入力画像を調査対象平面上に投影変換した画像にオーバーレイ表示する。つまり、対象領域の全体を着色するのではなく、図１４に示すように、例えば調査対象の小領域の周辺部のみ、中心部のみ、１／４右上部のみに着色することにより、図１５に示すように、調査対象箇所のＩＴＶカメラ映像をあわせて参照可能な図を出力するものである。
【００８８】
これにより、従来はＩＴＶカメラ映像にオーバーレイ表示することで注目地点と混雑度の対応が理解しやすいものの、カラーチャート表示では色表示に隠れて注目地点を視認しにくいとの欠点があったが、第９の実施の形態ではこのような欠点が排除され、注目地点と混雑度の対応が視覚的にも直感的に理解しやすい利点を持つ。
【００８９】
［第１０の実施の形態］
本発明の第１０の実施の形態は、混雑度結果をバー表示で出力するものである。
【００９０】
第１０の実施の形態では、第９の実施の形態の部分的なカラーチャートの代わりに、局所変化時間率を集計した小領域ごとに、図１６に示すような局所変化時間率に比例する長さの棒を、図１７に示すように、ＩＴＶカメラ入力画像またはこれを調査対象平面上に投影変換した画像にオーバーレイ表示する。
【００９１】
例えば局所変化時間率が８０％の小領域には長さが８画素の棒を、局所変化時間率が１０％の小領域には長さが１画素の棒を対応する領域に表示し、局所変化時間率が０％の小領域には棒を対応する領域に表示しない。
【００９２】
これにより、従来方法の等高線は混雑度を読むのが煩雑で、カラーチャート表示では色表示と混雑度の対応を知る必要があるため混雑度の度合いについては理解がしにくい問題があったが、本実施の形態ではこれを棒グラフで表示することにより小領域ごとの局所的な混雑度の度合いについて直感的に理解できる。また、大局的にも混雑度調査結果出力の領域全体の中で棒グラフの占める面積が大きい部分は混雑度が高く、棒グラフの面積が少ない部分は車両や人の動きが少ない地点であることなどの結果を把握しやすい利点がある。
【００９３】
［第１１の実施の形態］
第１１の実施の形態では、混雑度結果を動きバー、静止をバー表示で出力するものである。
【００９４】
第１１の実施の形態では、混雑度調査装置に用いる変化検出手段３０として、特開２００２−２４８０８で開示した「静止障害物検出方式」を採用する。この公開技術の内容については後述する。そして第１１の実施の形態では、小領域ごとに検出される変化状態と静止状態を一定時間計数して小領域ごとの局所変化時間率と局所滞留時間率を計算して出力する混雑度調査装置において、出力・記録手段６０は集計した小領域ごとに局所変化時間率と局所滞留時間率のそれぞれに比例する長さの２本の棒をＩＴＶカメラ入力画像またはこれを調査対象平面上に投影変換した画像にオーバーレイ表示する。
【００９５】
例えば、図１８に示すように、小領域ごとに局所変化時間率を示す棒グラフを緑色で、局所滞留時間率を示す棒グラフを赤色で表示する。
【００９６】
これにより、小領域ごとに動きの多かった部分と滞留する時間の多かった部分を含む局所的な混雑度の度合いについて直感的に理解することが可能となり、例えば交差点の利用状況を人や車の流れと滞留状況を含み動的に観察することができる。また、動きと滞留の状態を棒グラフの色で区別すれば大局的にも領域全体の中で緑色の棒グラフの占める面積が大きい部分は混雑度が高いが流れており、赤色の棒グラフの面積が大きい部分は車両や人の動きが停滞する地点であることなどの結果を把握しやすい利点がある。
【００９７】
［公開技術の説明］
ここで特開２００２−２４８０８で開示した公開技術「静止障害物検出方式」の内容について説明をしておく。
【００９８】
上述した公開技術（特開２００２−２４８０８）の要旨は、動画像からサンプリングした画像を小領域に切り分け、一つ一つの小領域について予め準備した移動物体を含まず障害物の存在しない背景画像の対応する領域との正規化相関値の計算を一定時間ごとにサンプリングされる画像に対して繰り返し行って、その相関値の変化状態を閾値で判定し、また、好ましくは、その判定に基づいて背景画像の更新も適宜行いながら、静止障害物を検出することにある。
【００９９】
この公開技術について更に詳述する。公開技術では、図１９に示すように、移動物体を含む動画像からサンプリングした画像（現在の画像）を小領域に切り分け、一つ一つの小領域について予め準備した移動物体を含まず、障害物の存在しない背景画像の対応する小領域との正規化相関を計算することを、一定時間ごとにサンプリングされる画像に対して繰り返す。
図１９は、画像を縦横に分割して小領域（高さＬ、幅Ｋ）としたものである。
正規化相関の計算式は、例えば、式（１１）がある。
【０１００】
【数１１】

【０１０１】
ある小領域についてその相関値の時間変化をグラフにした例を図２０に示す。図２０の区間Ａに示すように、ある小領域についてこの相関値の変化量が大きければ、移動物体が通過中であると判断し、また、図２０の区間Ｂ、Ｃに示すように、相関値の変化量が小さい状態が一定回数以上連続すれば、その小領域に写っているのは背景であると判断できる。
このようにして背景と見なされた小領域の相関値が、図２０の区間Ｂに示すように、１に近い値であれば、予め準備した背景画像の対応する小領域と同じ画像であることになる。
【０１０２】
しかし、図２０の区間Ｃに示すように、ある閾値より小さければ、この小領域には予め準備した背景画像には写っていない静止物体が存在すると判断できる。
【０１０３】
更に、相関値の変化量が小さい状態が一定回数以上連続し相関値が１に近い値となった小領域について、現在の画像の小領域を背景画像の対応する小領域と置き換えることにより、背景画像を更新する。
【０１０４】
［第１２の実施の形態］
本発明の第１２の実施の形態は、混雑度を局所変化時間率の最大値で正規化してバー表示で出力するものである。
【０１０５】
第１２の実施の形態では、第１０及び第１１の実施の形態の混雑度調査装置の棒グラフ表示出力について、小領域ごとに表示する棒グラフの長さを、対応する小領域の局所変化時間率を調査領域全体の局所変化時間率の最大値に対する割合に比例する長さとする。例えば、調査領域全体の局所変化時間率の最大値が２０％のときにその最大の小領域には長さが１０画素の棒を、局所変化時間率が１０％の小領域には長さが５画素の棒を、局所変化時間率が４％の小領域には長さが２画素の棒を対応する領域に表示する。
【０１０６】
第１０及び第１１の実施の形態のように、単に棒グラフ出力をしたのでは、比較的変化の少ない調査場所や時間帯には棒グラフ長さが全体に短く、場合によっては全ての小領域が最小長さとなり小領域ごとの混雑度の比較ができないことも想定されるが、第１２の実施の形態によれば映像中に動きが少ない地点および時間帯においても局所変化時間率の最大値を基準に棒グラフ表示するので小領域ごとに局所変化時間率に違いがあればこれを反映した棒グラフ表示となるため、局所的な混雑度の度合いについて直感的に理解できる。
【０１０７】
また、大局的にも混雑度調査結果出力の領域全体の中で棒グラフの占める面積が大きい部分は混雑度が高いほうで、棒グラフの面積が少ない部分は車両や人の動きが少ない地点であることなどの結果を把握できる。
【０１０８】
［第１３の実施の形態］
本発明の第１３の実施の形態は、混雑度を局所変化時間率の最大値で正規化し更に関数変換してバー表示で出力するものである。
【０１０９】
第１３の実施の形態では、第１０及び第１１の実施の形態の混雑度調査装置の棒グラフ表示出力について、小領域ごとに表示する棒グラフの長さを、対応する小領域の局所変化時間率を調査領域全体の局所変化時間率の最大値に対する割合に基づきあらかじめ設定する線形変換式により計算する長さとする。
【０１１０】
例えば、変換式をｙ＝ｘ^３で調査領域全体の局所変化時間率の最大値が２０％のときにその最大の小領域には長さが１０画素の棒を、局所変化時間率が９０％の小領域には長さが７画素（１０画素×０．９^３）の棒を、局所変化時間率が６０％の小領域には長さが２画素（１０画素×０．６^３）の棒を対応する領域に表示するものとする。
【０１１１】
これにより、上記の例の場合、第１０の実施の形態では局所変化時間率６０％と９０％の間で棒グラフの長さの差が３画素だったのに対して、第１３の実施の形態では５画素となり、特に各小領域の局所変化時間率がこの間（つまり６０％から９０％の間で調査範囲における混雑度が高めの部分が多い場合）に集中しているようなケースに違いを比較しやすい。
【０１１２】
また、変換式をｙ＝ｘ^１／３とした場合に、例えば調査領域全体の局所変化時間率の最大値が２０％のときにその最大の小領域には長さが１０画素の棒を、局所変化時間率が６％の小領域には長さが７画素（１０画素×０．３^１／３）の棒を、局所変化時間率が１％の小領域には長さが３画素（１０画素×０．０５^１／３）の棒を対応する領域に表示する。
【０１１３】
これにより、第１０の実施の形態では局所変化時間率０％と３０％の間で棒グラフの長さの差が３画素だったのに対して、第１３の実施の形態では７画素となり、特に各小領域の局所変化時間率が３０％以下の小さいデータに集中しているようなケースに違いを比較しやすい。
【０１１４】
【発明の効果】
本発明では、基本構成要素を共通にしており、この共通の基本構成要素により次に示す共通の効果を奏することができる。
【０１１５】
▲１▼ 局所的な混雑度の機械による調査が可能。
▲２▼ 調査員の疲労等による計数ミスなどのばらつきがない混雑度データが得られる。
▲３▼ 長時間にわたる連続する混雑度調査が可能なので、混雑度の１日の変動、１週間の変動、１ヶ月の変動、１年間の変動を容易に得ることができる。
▲４▼ 多地点の同時調査が可能になるので、地域的な混雑度の違いや通行人の流れなどを把握できる。
▲５▼ ポイントの混雑度のみでなくＩＴＶカメラで撮影される範囲の領域（道路、通路、歩道、広場、門、交差点、地下街、等々）の各部分における混雑度を調査できる。道路や交差点を通過する車両の調査においては車線をまたがる車両を含む混雑度の調査が可能。
▲６▼ 人も車も同じ装置を適用可能。
▲７▼ トラフィックカウンタや専用ゲートの設置が不要。
▲８▼ 統計解析に利用可能な混雑度を正しく反映。
▲９▼ ＩＴＶカメラの映像をモニタで観察すれば混雑度の視覚的な監視も可能。
【０１１６】
請求項１の発明では、変化検出時に小さな部分を排除するため、基本構成要素による効果に加えて、更に次のような効果を奏する。
▲１▼ 調査対象の車両や人物の大きさより小さな物体を排除するので、小さな物体が混雑度調査の出力の誤差となる危険性が排除される。
【０１１７】
請求項２の発明では、斜めカメラで変化検出時に上部分をカットするため、基本構成要素による効果に加えて、更に次のような効果を奏する。
▲１▼ 光軸が斜め方向に設置されたＩＴＶカメラを使用した場合に変化検出時に変化検出領域を結合して上部分をカットすることにより、混雑度調査の出力は平面的な注目地点の混雑度を反映するものとなる。
【０１１８】
請求項３の発明では、平面座標系に変換して混雑度を出力するため、基本構成要素による効果に加えて、更に次のような効果を奏する。
▲１▼ 本発明では混雑度出力を平面座標に座標変換することにより調査対象箇所の平面上の注目点ごとに混雑度を出力するので、調査対象箇所における混雑度を直接評価することが可能になる。
【０１１９】
請求項４の発明では、複数カメラによる混雑度を平面座標系で統合して出力するため、基本構成要素による効果に加えて、更に次のような効果を奏する。
▲１▼ 混雑度調査装置は複数のＩＴＶカメラを使用して広い領域の監視を行うような場合にもこれら複数のＩＴＶカメラを使用した混雑度調査結果を一つの平面に統合するので、データの一元管理が可能となる。
▲２▼ さらに１台のカメラでは死角となっていた地点を他のカメラで調査することにより高範囲にわたる地域の混雑度調査が一つの平面図で実施できる。
【０１２０】
請求項５の発明では、複数ビデオ映像による混雑度を平面座標系で統合して出力するため、基本構成要素による効果に加えて、更に次のような効果を奏する。
▲１▼ 基本構成要素および請求項４の発明と同一の効果が得られるのに加えて、構成は混雑度調査結果装置が１セットになるので装置の節約が図れる。
【０１２１】
請求項６の発明では、旋回動作するカメラによる混雑度をパノラマ画像で統合して出力するため、基本構成要素による効果に加えて、更に次のような効果を奏する。
▲１▼ １台のＩＴＶカメラを旋回させて広い領域の監視を行う目的で設置された監視カメラを使用して監視地域の混雑度調査を行う場合にも監視カメラの動作範囲の全体をカバーする地域における混雑度を調査可能である。
▲２▼ 停止位置ごとのＩＴＶカメラ映像を一つに統合してパノラマ画像を生成しておき、これに本実施例の混雑度調査結果出力をオーバーレイ表示すれば、注目地点と混雑度の対応が直感的に理解しやすい。
【０１２２】
請求項７の発明では、旋回動作するカメラによる混雑度を平面座標系で統合して出力するため、基本構成要素による効果に加えて、更に次のような効果を奏する。
▲１▼ 旋回動作する監視カメラの撮像範囲の全体をカバーする地域における混雑度を調査可能である。
▲２▼ 調査対象箇所における混雑度を平面図と照合して直接評価することが可能になる。
▲３▼ 請求項４，５の複数のカメラを使用する方法と比べて使用するカメラが１台なので構成が簡単である。
【０１２３】
請求項８の発明では、旋回動作する複数のカメラによる混雑度を平面座標系で統合して出力するため、基本構成要素による効果に加えて、更に次のような効果を奏する。
▲１▼ 複数の監視カメラの撮像範囲の全体をカバーするさらに広い地域における混雑度を調査可能である。
▲２▼ 調査対象箇所における混雑度を平面図と照合して直接評価することが可能になる。
▲３▼ 請求項４、５の複数のカメラを使用する方法と比べれば使用するカメラが少なくてすむ。
▲４▼ 請求項６、７の１台の旋回カメラを使用する方法では死角となっていた地点を他の旋回カメラで調査することにより、高範囲にわたる地域の混雑度調査が一つの平面図で実施できる。
【０１２４】
請求項９の発明では、混雑度結果を部分表示で出力するため、基本構成要素による効果に加えて、更に次のような効果を奏する。
▲１▼ 対応する領域の一部分のみを描画してＩＴＶカメラ入力画像またはこれを調査対象平面上に投影変換した画像にオーバーレイ表示するので、カラーチャート表示では色表示に隠れて注目地点を視認しにくいとの欠点が排除され、注目地点と混雑度の対応が視覚的にも直感的に理解しやすい。
【０１２５】
請求項１０の発明では、混雑度結果をバー表示で出力するため、基本構成要素による効果に加えて、更に次のような効果を奏する。
▲１▼ 局所変化時間率を棒グラフで表示することにより小領域ごとの局所的な混雑度の度合いについて直感的に理解できる。
▲２▼ 大局的にも混雑度調査結果出力の領域全体の中で棒グラフの占める面積が大きい部分は混雑度が高く、棒グラフの面積が少ない部分は車両や人の動きが少ない地点であるなどの結果を把握しやすい。
【０１２６】
請求項１１の発明では、混雑度結果を動き、静止をバー表示で出力するため、基本構成要素による効果に加えて、更に次のような効果を奏する。
▲１▼ 局所変化時間率を動きと静止で個別の棒グラフで表示するので、出力小領域ごとに動きの多かった部分と滞留する時間の多かった部分を含む局所的な混雑度の度合いについて直感的に理解することが可能となる。
▲２▼ 動きと滞留の状態を棒グラフの色で区別すれば大局的にも領域全体の中で黄色の棒グラフの占める面積が大きい部分は混雑度が高いが流れており、赤色の棒グラフの面積が大きい部分は車両や人の動きが停滞する地点であることなどの結果を把握しやすい利点がある。
【０１２７】
請求項１２の発明では、混雑度を局所変化時間率の最大値で正規化してバー表示で出力するため、基本構成要素による効果に加えて、更に次のような効果を奏する。
▲１▼ 混雑度を局所変化時間率の最大値で正規化して出力するので、映像中に動きが少ない地点および時間帯においても局所変化時間率の最大値を基準に棒グラフ表示するので小領域ごとに局所変化時間率に違いがあればこれを反映した棒グラフ表示となるため、局所的な混雑度の度合いについて直感的に理解できる。
▲２▼ 大局的にも混雑度調査結果出力の領域全体の中で棒グラフの占める面積が大きい部分は混雑度が高いほうで、棒グラフの面積が少ない部分は車両や人の動きが少ない地点であることなどの結果を把握できる。
【０１２８】
請求項１３の発明では、混雑度を局所変化時間率の最大値で正規化し更に関数変換してバー表示で出力するため、基本構成要素による効果に加えて、更に次のような効果を奏する。
▲１▼ 対応する小領域の局所変化時間率を調査領域全体の局所変化時間率の最大値に対する割合に基づきあらかじめ設定する線系変換式により計算する長さで表示するので、特に各小領域の局所変化時間率が高いほうまたは低いほうに集中しているようなケースに局所的な変化時間率の違いを比較しやすい。
▲２▼ 関数変換の変換式を適当に選択することにより、局所変化時間率が高いほうに集中している場合と低いほうに集中している場合のいずれにも対応出来る。
【図面の簡単な説明】
【図１】本発明の実施の形態に係る混雑度調査装置を示す構成図。
【図２】混雑度調査場所のＩＴＶカメラ映像例を示す画像図である。
【図３】変化検出手段の一例を示す構成図である。
【図４】カラーチャート表示した混雑度調査結果例を示す画像図である。
【図５】混雑度超過場所に斜め方向に設置されたＩＴＶカメラ映像例を示す画像図である。
【図６】ＩＴＶカメラ映像中の変化検出例を示す画像図である。
【図７】変化検出の上部分カットの例を示す画像図である。
【図８】空間座標系とカメラ座標系との関係を示す説明図である。
【図９】監視カメラのカメラパラメータを示す説明図である。
【図１０】平面座標系への変換結果例を示す画像図である。
【図１１】交差点のＩＴＶカメラとその監視エリアを示す画像図である。
【図１２】２台のＩＴＶカメラとその監視エリアを示す画像図である。
【図１３】パノラマ画像を示す説明図である。
【図１４】混雑度調査結果の部分表示例を示す画像図である。
【図１５】部分表示による混雑度調査結果表示例を示す画像図である。
【図１６】混雑度調査結果の棒グラフ表示例を示す画像図である。
【図１７】棒グラフ表示による混雑度調査結果例を示す画像図である。
【図１８】動きと滞留を区別した混雑度調査結果の棒グラフ表示の例を示す画像図。
【図１９】対応する小領域の正規化相関を計算する手法を示す説明図。
【図２０】背景画像との正規化相関値の時間推移を示す特性図。
【符号の説明】
１０ ＩＴＶカメラ
２０ 画像入力手段
３０ 変化検出手段
４０ 変化回数積算手段
５０ 局所変化時間率計算手段
６０ 出力・記録手段[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a congestion degree investigation device. More specifically, the present invention relates to an apparatus for examining the degree of congestion of vehicles and pedestrians on roads, intersections, stores, and the like. BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a congestion degree survey device that measures and records the degree of congestion of a vehicle or a pedestrian passing or staying in a video image temporally and locally.
[0002]
[Prior art]
[Background and purpose]
Grasp the users of public facilities such as stations, investigate the use of underground malls and shopping streets, investigate the traffic volume of roads and passages, conduct preliminary surveys to predict the number of users when opening stores such as convenience stores, Traffic volume is being investigated at roads and intersections for the purpose of predictive surveys on the effects of widening and land readjustment, but more detailed information on local and temporal There is a need to understand such fluctuations. For example, in order to examine the configuration of lanes on roads, to install lanes dedicated to turning left and right at intersections, and to control the spacing of traffic lights, and to examine product placement by grasping customer movements in stores, There is a need for a means for investigating which part in the investigation area is congested or at what time and how much there is no movement.
[0003]
In order to investigate such a degree of congestion, an attempt has been made to use image processing in addition to a technique of visually counting by a researcher and a technique of counting vehicles by a traffic counter. In other words, research reports from academic societies and the like often use approaches to detect and track vehicles and detect and track pedestrians by installing cameras at survey locations, photographing intersections and paths, and processing the images. It has been reported.
For example, Non-Patent Document 1 describes detection of a vehicle traveling on a highway, and Non-Patent Document 2 describes a research report on detection of pedestrian behavior.
[0004]
[Non-patent document 1]
Shinji Ozawa, "Recognition and Understanding in ITS Road Images", IEICE Technical Report, PRMU 98-91-105, pp. 139-157. 99-104, 1998.
[Non-patent document 2]
Takeuchi, Kaneko, Igarashi, Sato, Hane, "Image analysis of walking behavior based on robust background subtraction and region extraction," Journal of the Institute of Image Electronics Engineers of Japan, vol. 31, no. 2, pp. 193-201, 2002.
[0005]
However, the above-described image processing attempts have the following disadvantages.
{Circle around (1)} Various methods have been proposed, but there is no decisive answer.
{Circle around (2)} Originally, image processing is easily affected by changes in lighting conditions and survey targets.
{Circle around (3)} Because there are many mistakes in cutting out passers-by or vehicles, the reliability of measured values is not high.
(4) There are some research cases such as installing a camera at the top to eliminate the overlap of pedestrians, but measurement errors remain.
[0006]
Because of the above drawbacks in image processing attempts, the following problems exist even if an attempt is made to conduct a congestion degree survey using image processing.
(1) In particular, a technique for correctly counting the number of crowds has not been realized.
{Circle around (2)} In the case where passersby do not overlap in the video, many methods have been proposed but cannot be applied to crowds.
{Circle around (3)} Since each method is a method of counting vehicles or passersby at a specific measurement target point, the congestion degree of the target point can be calculated by totaling the measurement results in time series.
{Circle around (4)} However, in order to investigate the distribution of the local congestion degree as described above, it is necessary to arrange the measuring means of the conventional method at each target point, for example, to investigate the movement of customers in the store. Such cases are complicated and difficult to realize.
[0007]
[Prior Invention]
Therefore, the inventor of the present application invented a “congestion degree investigation device” for investigating which part in the investigation target area is congested or at what time and in which time, and has previously been patented as Japanese Patent Application No. 2002-259636. Filed (filing date: October 25, 2002). The “congestion degree investigation device” of Japanese Patent Application No. 2002-259636 is referred to as “prior application invention” in this specification.
[0008]
The configuration of the "congestion degree investigating apparatus" according to the invention of the preceding application is composed of an ITV camera having a photographing range for a congestion degree investigation target, an image input means, and a pixel (or a small image obtained by appropriately dividing the image) in the image. Change detection means independent for each area, change number integration means for integrating the number of times a change is detected for each pixel (or for each small area), and a ratio between the number of change detections and the total number of images in a given time. A local change time rate calculating means for calculating, and an output / recording means for outputting or recording the calculated ratio with a degree of congestion, wherein the video input means each time a video taken by the ITV camera is updated. An image is input, and the change detection means examines whether or not there is a change for each pixel (or for each small area), and accumulates the number of changes detected within a predetermined time set by the local change time rate calculation means, and Local change time By calculating the ratio of the total number of images predetermined time period by the calculating means is characterized in that the output or recorded by the output and recording means the ratio as the congestion degree.
Further, instead of the video captured by the ITV camera, a configuration device that uses a video reproduced by a video reproducing unit that reproduces a video of a target location of the congestion degree survey may be used.
[0009]
Since the basic configuration of the “congestion degree investigation device” according to the invention of the previous application is used as a basic configuration requirement of the present invention, the details thereof will be described later.
[0010]
According to the invention of this earlier application, a target location for which the degree of congestion is to be investigated is photographed by an ITV camera, and a portion having a change in the video is detected by an image processing method. Focusing on the parts that have changed in the video are pedestrians and traveling vehicles, the presence / absence of changes at the point of interest (pixel of interest) is integrated for a certain period of time, and the ratio of the change to the certain period of time (this Is referred to as a local change time rate), and this is used as an index of the congestion degree. Although the local change time rate does not count the number of direct passers or vehicles, it is enough information to consider the less crowded and less crowded, and is used in the survey and statistics of the crowdedness described above. It can be an indicator.
[0011]
Further, as an embodiment of the invention of the earlier application, only the above-described method of evaluating a point of interest in a small area instead of each pixel of an image, and the method of using a video image reproduced by a video reproducing unit instead of an image of an ITV camera are only used. Rather, a method of detecting a change by a background difference or a spatial difference or a time correlation change analysis, a method of detecting a still state together with a video change by a time correlation change analysis, a method of eliminating a spatial or temporal isolated point A method of outputting the number of vehicles or the number of people as a congestion degree by a conversion formula, and a method of displaying the output of the congestion degree in a contour line display or a color chart display corresponding to an ITV camera image have been proposed.
[0012]
[Problems to be solved by the invention]
However, as a result of further research and development, it was found that the prior invention had the following problems to be improved.
[0013]
1) In the technique of the invention of the prior application, the local change time rate is calculated by integrating the number of pixels or small areas where a change is detected in the video. For this reason, since a change is detected for each pixel or small area of the ITV camera image, an object smaller than the survey target may be detected, and there is a risk that an output error of the congestion degree survey may occur. For example, if the traveling vehicle is the subject of the survey in the video of the intersection, a small falling object unrelated to the survey is also detected, and the output is equivalent to that of the stopped vehicle.
[0014]
2) According to the technique of the invention of the prior application, the output of the result of the congestion degree survey is distribution data viewed from the viewpoint of the ITV camera image. Although the correspondence between the point of interest and the degree of congestion can be intuitively understood by overlaying the image on the ITV camera image in this way, when the optical axis of the ITV camera is oblique rather than vertically downward, In some cases, the degree of congestion at a planar point of interest is not reflected. For example, when a large vehicle stops in front of a sidewalk and blocks the sidewalk in an image of observing an intersection obliquely from above, it is difficult to distinguish whether the position of a stationary object detected there is a roadway or a sidewalk.
[0015]
3) The congestion degree investigating device according to the invention of the earlier application examines the congestion degree in the video of one ITV camera. Therefore, in the case of monitoring a wide area using a plurality of ITV cameras, a video is acquired for each of the installed ITV cameras and the congestion degree survey is individually performed. Data are individually recorded and stored in correspondence with the ITV cameras of the prior art, so that data cannot be centrally managed.
[0016]
4) The congestion degree investigating device according to the invention of the earlier application detects a change area in the video and sets it as the congestion degree. Therefore, when a wide area is monitored by turning one ITV camera, the image itself of the ITV camera changes. Therefore, in the prior application invention, the change area in the image is detected to determine the congestion degree. Cannot check the congestion degree.
[0017]
5) The results of the congestion survey according to the prior application invention are displayed in a contour display and a color chart display. Therefore, although the correspondence between the point of interest and the degree of congestion can be easily understood intuitively by overlay display on the ITV camera image, it is difficult to intuitively understand the degree of congestion. It is difficult to read the degree of congestion on the contour line, and it is difficult to visually recognize the point of interest by hiding the color display in the color chart display.
[0018]
6) In the prior invention, an overlay display is performed at a local change time ratio obtained by dividing a time at which a change has occurred by a certain measurement time. Such a local change time rate is effective as an index indicating a difference in the degree of congestion in a scene having a relatively large amount of motion in a video. It is assumed that the user concentrates on the minimum unit of the contour line display or the color chart display. In this case, it is not possible to recognize the difference in the local congestion degree. Conversely, in survey places and time zones where there are relatively many changes, it is assumed that the concentration will be concentrated on the maximum unit of the contour line display or color chart display that indicates the local change time rate. I can't tell.
[0019]
The present invention has been made in view of the above-described conventional technology, and can accurately detect the degree of congestion, can output the degree of congestion reflecting the degree of congestion at a planar attention point, and perform wide-range monitoring. It is an object of the present invention to provide a congestion degree investigation device capable of performing the following.
[0020]
[Means for Solving the Problems]
A configuration of the present invention that solves the above-described problem, a television camera that outputs an image by capturing an image of a target location of a congestion degree survey as an imaging range,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
The change detection means sets the size of the survey object in advance, and when detecting a change in the image for each small region in the image, the small region in which the change is detected is detected in the vicinity of the small region in which the change is detected. The size of the entire change detection portion is calculated by coupling, and if the size of the entire change detection portion is smaller than the size of the survey object set in advance, the size of the small region included in the entire change detection portion is determined. It is characterized in that there is no change in the image and the image is excluded from the change count integration.
[0021]
Further, the configuration of the present invention is a television camera that outputs an image by photographing a target location of the congestion degree survey as a photographing range,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
The change detecting means sets in advance the number of upper exclusion small areas based on the height of the survey target in the video, and when detecting a change in the video for each of the small areas in the video, the change detection means The small areas where a change is detected around the vertical direction are connected, and the small area included in the range of the number of the excluded small areas from the upper end of the connected small areas is excluded from the change number accumulation as there is no change in the image. It is characterized by doing.
[0022]
Further, the configuration of the present invention is a television camera that outputs an image by photographing a target location of the congestion degree survey as a photographing range,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
The output / recording unit is characterized in that the imaging range of the television camera is projected and converted into a plane of a congestion degree investigation target place, and the congestion degree is output as a congestion degree for each point on the plane in the investigation target place.
[0023]
Further, the configuration of the present invention includes a plurality of sets of congestion degree investigation devices that output the congestion degree as the degree of congestion at each point on the plane of the survey target location, and the imaging of each television camera obtained from each congestion degree measurement device. It is characterized in that the congestion degree for each point on the plane corresponding to the range is integrated into one large plan view and output.
[0024]
Further, the configuration of the present invention is a television camera that outputs an image by photographing a target location of the congestion degree survey as a photographing range,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
In place of the television camera, the video camera further includes a video playback unit that plays back a video of a target portion of the congestion degree survey, and the video input unit inputs a video every time a video output from the video playback unit is updated. ,
The output / recording unit projects the imaging range onto the plane of the congestion degree investigation target location, and outputs the congestion degree as the congestion degree for each point on the plane in the investigation target location,
A plurality of videos photographed and recorded by a plurality of television cameras for a large survey area are reproduced by one video reproducing unit, and the output / recording unit outputs congestion degree survey output in each photographing range to one. It is characterized in that it is integrated and output on a plane.
[0025]
Further, the configuration of the present invention is a television camera that outputs an image by capturing an image of a target location of a congestion degree survey as a capturing range,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
The television camera is installed on a pan head that turns in a horizontal plane, and the pan head repeats the turning operation while stopping for a predetermined time at a plurality of preset points, the change detection unit and the change number accumulation unit. Performs change detection for each small region and integration of the number of changes individually for each stop position during the time when the camera platform is stopped, and the local change time rate calculating means individually performs local change for each small region for each stop position. A change time rate is calculated, and the output / recording means integrates these local change time rates for each stop position into one panoramic image and outputs or records the congestion degree.
[0026]
Further, the configuration of the present invention is a television camera that outputs an image by capturing an image of a target location of a congestion degree survey as a capturing range,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
The television camera is installed on a pan head that turns in a horizontal plane, and the pan head repeats the turning operation while stopping for a predetermined time at a plurality of preset points, the change detection unit and the change number accumulation unit. Performs change detection for each small region and integration of the number of changes individually for each stop position during the time when the camera platform is stopped, and the local change time rate calculating means individually performs local change for each small region for each stop position. The output / recording means calculates the rate of change, and performs projection conversion to the plane of the inspection target location for each stop position, and obtains the congestion degree for each point on the plane corresponding to the imaging range of each stop position obtained from these. It is characterized in that it is integrated and output into one large plan view.
[0027]
Further, the configuration of the present invention is provided with a plurality of sets of the congestion degree investigation device,
It is characterized in that the congestion degree at each point on the plane corresponding to the imaging range of each television camera obtained from each congestion degree investigation device is further integrated into one large plan view and output.
[0028]
Also, the configuration of the present invention, in the congestion degree investigation device,
The output / recording unit draws only a part of the corresponding area and draws only a part of the area when the area corresponding to the small area having the same local change time rate is displayed as a color chart that draws the same color. The obtained color chart display is overlaid on an image obtained by a television camera or an image obtained by projecting and converting this image on a plane to be investigated.
[0029]
Also, the configuration of the present invention, in the congestion degree investigation device,
The output / recording means is an image obtained by a television camera or an image obtained by projecting and converting this image on a plane to be investigated on a bar having a length proportional to the local change time rate for each small area in which the local change time rate is totaled. And an overlay display.
[0030]
Further, the configuration of the present invention is a television camera that outputs an image by capturing an image of a target location of a congestion degree survey as a capturing range,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
The change detection means calculates a local change time rate and a local residence time rate for each small area by counting the change state and the stationary state detected for each small area for a certain period of time,
The output / recording means sets two bars having a length proportional to each of the local change time rate and the local residence time rate for each of the small areas, and displays the image obtained by the television camera or the image on the survey target plane. The projection-converted image is displayed as an overlay.
[0031]
Also, the configuration of the present invention, in the congestion degree investigation device,
Regarding the bar graph display output by the output / recording means, the length of the bar graph displayed for each small area is such that the local change time rate of the corresponding small area is proportional to the ratio of the local change time rate of the entire survey area to the maximum value. It is characterized by the above.
[0032]
Also, the configuration of the present invention, in the congestion degree investigation device,
For the bar graph display output by the output / recording means, the length of the bar graph displayed for each small area is set in advance based on the ratio of the local change time rate of the corresponding small area to the maximum value of the local change time rate of the entire survey area. The length is calculated by a linear conversion equation.
[0033]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Although there are a plurality of embodiments of the present invention as will be described later, common components are adopted as basic components. This common basic component employs the basic configuration of the “congestion degree investigation device” according to the above-mentioned invention. Therefore, prior to the description of each embodiment, common basic components of the present invention will be described.
[0034]
[Basic components of the present invention]
The basic components of the present invention will be described. In the basic components employed in the present invention, a target location for which the degree of congestion is to be investigated is photographed with an ITV camera (television camera), and a portion that changes in the video is detected by an image processing method. ) At a point of interest (pixel or small area: hereinafter referred to as a “small area”), the presence or absence of a change in the video is integrated for a certain period of time, and the ratio of the time of the change to the certain time (this is the local change time rate This local change time ratio is output as an index of the congestion degree. Although the local change time rate does not count the number of direct passers or vehicles, it is enough information to consider the less crowded and less crowded. It can be an indicator.
[0035]
As shown in FIG. 1, the basic components of the present invention include an ITV camera 10, a video input unit 20, and a number corresponding to the number of small areas set in a video (video screen) (the same number as the set number of small areas). ), A plurality of change frequency integrating means 40, a plurality of local change time rate calculating means 50, and a congestion degree output / recording means 60.
[0036]
The operation of the basic components of the present invention will be described. An ITV camera 10 that captures an image of a target location of the congestion degree survey is set. Each time a video image captured and output by the ITV camera 10 is updated, a video image as shown in FIG. 2 is input (taken) by the image input means 20.
[0037]
The change detection means 30 are arranged in the video (in the video screen) by the number corresponding to the number of preset small areas, and each of the small areas in the video captured (input) by the image input means 20 is provided. Are individually input to each of the change detecting means 30. For example, the image of the small area on the first surface is input to the first change detection means 30, the image of the small area on the second surface is input to the second change detection means 30, and the image of the third small area is input to the second change detection means 30. The image of the area is input individually while being input to the third change detecting means 30. Then, each of the change detection units 30 examines whether or not there is a change in the input video divided for each small area.
[0038]
Each change number integrating means 40 is connected to each change detecting means 30 in a one-to-one correspondence, and calculates the number of times of detection of a change in an image for each small area detected by the correspondingly connected change detecting means 30. The predetermined period of time is set as one period, and is integrated for each period (every constant period).
[0039]
Each of the local change time rate calculating means 50 is connected to each of the change number accumulating means 40 in a one-to-one correspondence, and a fixed time (1) for each small area detected by the correspondingly connected change number accumulating means 40. The ratio (local change time ratio) of the number of times of change detection of the video in the period) to the total number of images within a certain time is calculated.
[0040]
The output / recording means 60 records each local change time rate calculated by each local change time rate calculation means 50 as the congestion degree and outputs the same to the outside.
[0041]
As an example of the change detecting means 30, there is a method based on time difference processing. That is, as shown in FIG. 3, the brightness of the small region of the previous image is temporarily held by the brightness holding circuit 31, and the brightness of the current small region is changed by the comparison circuit 32 when the brightness is different from the previous brightness (image change). The presence is detected. Note that a video signal delay circuit may be employed instead of the luminance holding circuit 31.
[0042]
The effect of the basic constituent elements of the present invention is that the present apparatus can be used to investigate the tendency of local congestion in the imaging range, so that an investigator is not required, and the investigator performs the inspection visually. The advantage is that reliable data acquisition with less variation and less errors due to differences in investigators can be performed, continuous long-term congestion surveys can be conducted, and multi-site surveys can be easily realized by installing a camera. There is.
[0043]
Also, compared to road traffic vehicle measurement using a traffic counter, etc., the installation work of the sensor is easier and it is possible to grasp the congestion degree of the entire monitoring area instead of each lane, especially including vehicles traveling across lanes Congestion degree investigation becomes possible.
[0044]
Furthermore, the same device can be used for both pedestrians and vehicles, and there is no need to install a dedicated gate for traffic volume survey.
[0045]
Also, the local change time rate output and recorded by this device does not count the number of direct passers and vehicles, but correctly reflects the degree of congestion for the purpose of examining the degree of congestion. Information that can be used as an indicator for surveys and statistics on congestion.
[0046]
In addition, in the basic components of the present invention, the monitoring target area may be the entire range captured by the ITV camera 10, or the congestion degree may be checked only for a part of the video.
[0047]
A method of using a video image instead of the ITV camera 10; a method of detecting a video change by a background difference or a spatial difference or a time correlation change analysis; and a method of detecting a still state together with a video change by a time correlation change analysis. A method for eliminating spatial or temporal isolated points, a method for outputting the number of vehicles or the number of people as a congestion degree by a conversion formula, displaying the congestion degree on an ITV camera image by contour lines, or a color chart as shown in FIG. A method of displaying an overlay on a display has been proposed, but the congestion degree survey device for basic components of the present invention described in this specification includes all of them.
[0048]
In FIG. 4, the drawing is drawn by monotone, black diagonal lines, and black cross hatch, but the actual display is color display, and each color is displayed in the color of the characters shown in the figure. 14 to 18 used in the following description, color display is shown in the same drawing state.
[0049]
Further, the congestion degree survey device of the present invention can realize the image input means and the following by software of a computer. This is the same in all the embodiments described below.
[0050]
[First Embodiment]
In the first embodiment of the present invention, a small portion is removed when a change in an image is detected, while using the above-described basic components of the present invention.
[0051]
That is, in the congestion degree investigation device according to the first embodiment, the size (height and width) of the investigation target is set in advance in the change detection unit 30. The change detecting means 30 connects the small areas where a change has been detected and the small areas where changes have been detected around the small area where the change has been detected when detecting a change in the image for each small area in the image. And the width) are calculated, and when the size of the entire change detection portion is smaller than the preset size (height and width) of the survey target, these small regions included in the entire change detection portion are calculated. It is assumed that there is no change in the image in and is excluded from the change count integration.
[0052]
Accordingly, the congestion degree research apparatus according to the first embodiment does not detect an object smaller than the size of the survey target set in advance, thereby eliminating the risk that a small object may cause an error in the output of the congestion degree research. Is done.
[0053]
For example, if the traveling vehicle is the target of the survey in the video of the intersection, small falling objects unrelated to the survey are excluded from the change detection, so it is possible to output a highly reliable congestion degree survey result that depends only on the traveling vehicle. Become.
[0054]
[Second embodiment]
In the second embodiment of the present invention, while using the above-described basic components of the present invention, when the optical axis of the ITV camera 10 is oblique, when the change in the image is detected, the upper part of the image is displayed. The video is cut to detect a change in image.
[0055]
That is, in the congestion degree investigation device according to the second embodiment, when the optical axis of the ITV camera 10 to be used is not in the vertical downward direction but in the oblique direction, the change detection means 30 transmits the camera image of the investigation object in advance. The “number of upper exclusion small areas” based on the medium height is set. The change detecting means 30 connects the small areas for which the change has been detected in the up-down direction to the small areas for which the change has been detected at the time of detecting the change of the image for each small area in the image. From the upper end, it is determined that there is no change in the image in the small areas included in the set number of upper excluded small areas, and is excluded from the change count integration.
[0056]
For example, when an image captured by the ITV camera 10 installed obliquely at the congestion degree investigation site is as shown in FIG. 5, an image of a change detection result obtained by detecting a change in the image by the change detection unit is displayed. In the case as shown in FIG. 6, as shown in FIG. 7, the upper part is cut out of the part (the part corresponding to the inspection object) in which the small areas where the change is detected in the vertical direction are connected. is there.
[0057]
With this, in the congestion degree survey device according to the second embodiment, since a predetermined size of the upper part of the survey target set in advance is cut, the output of the congestion degree survey indicates the congestion degree of the planar attention point. It will reflect. For example, when a large vehicle stops at the front of the sidewalk and blocks the sidewalk with an image of observing the intersection obliquely from above, the fact that the stationary object detected there is stationary on the roadway is correctly output.
[0058]
[Third Embodiment]
In the third embodiment of the present invention, while using the above-described basic components of the present invention, when the optical axis of the ITV camera 10 is oblique, an image in which a change in video is detected is converted to a plane coordinate system. It converts and outputs the congestion degree.
[0059]
That is, in the congestion degree survey device according to the third embodiment, when the optical axis of the ITV camera 10 to be used is not in the vertical downward direction but in the oblique direction, the congestion degree output / recording means 60 is provided by the ITV camera 10. Is projected onto the plane of the congestion degree investigation target location, and the congestion degree output is output as the congestion degree for each point on the plane of the investigation target location.
[0060]
Here, the conversion formula of the projection conversion to the survey target plane will be described. Camera parameters that determine the relationship between the position (u, v) in the ITV camera image as shown in FIG. 8 and the spatial position (x, y, z) in the real world are:
(1) Six external parameters relating to the camera installation position (X, Y, Z) and camera posture (θ, φ, ψ),
{Circle around (2)} There are a total of 11 parameters including the lens focal length f of the camera itself, the image center (u0, v0), the CCD element unit length ratio, and five internal parameters of the axis intersection angle (references: excluding reference Tsuji, 3D Vision, Kyoritsu Shuppan, 1998).
[0061]
Among them, in the general ITV camera 10, the CCD element unit length ratio is adjusted to 1: 1 and the crossing angle of the axes is adjusted to 90 degrees. Since the ITV camera 10 is installed horizontally, ψ = 0. Since φ is irrelevant for the calculation of coordinates on the plane, φ = 0 can be set. Further, if the origin of the camera coordinate system is u0 = v0 = 0 at the camera lens center, and the origin of the spatial coordinate system is also X = Y = Z = 0 at the camera lens center, the remaining parameters are as shown in FIG. As shown in FIG. 7, there are two ratios, α, which represent the ratio between the lens focal length and the CCD size, and the camera elevation angle θ.
[0062]
Assuming that a certain measurement point is M (X, Y, Z) and an image point of M on the CCD is m (u, v), the relationship between M and m is as follows.
(Equation 1)

Is represented by here, m Is an extended vector of m, M Is an extended vector of M, s is a scalar, and P is
(Equation 2)

Is a transformation matrix represented by Here, A is a camera internal parameter,
[Equation 3]

[0063]
R is the rotation matrix of the camera only for θ,
(Equation 4)

Therefore, equation (1) is
(Equation 5)

And expand this,
(Equation 6)

(Equation 7)

(Equation 8)

It becomes.
[0064]
Since X is constant at the camera height from the horizontal plane to be monitored, m and M correspond one-to-one,
(Equation 9)

(Equation 10)

Is represented by
[0065]
That is, the expressions (9) and (10) are the expressions for the projection transformation of the point (u, v) in the ITV camera image with the point (Z, Y) on the plane to be monitored. For example, the detection image shown in FIG. 10 can be obtained by performing coordinate transformation of the detection image shown in FIG. 7 into a plane coordinate system.
[0066]
In a simple congestion level survey device including only basic components, the output of the congestion level is output for each small area of the ITV camera image. Therefore, when the optical axis of the ITV camera 10 is oblique, a flat In contrast to the case where the congestion degree at the point of interest is not reflected, the congestion degree output according to the third embodiment outputs the congestion degree for each point of interest on the plane of the target point. It is possible to directly evaluate the degree of congestion in.
For example, operations such as collating a congestion degree survey using an image of an intersection with a map or collating a congestion degree survey using a video in a store with an in-store layout diagram are facilitated.
[0067]
[Fourth Embodiment]
The fourth embodiment of the present invention includes the congestion degree survey device according to the third embodiment described above, uses the plurality of sets of congestion degree measuring devices, and measures the congestion degree by a plurality of cameras in a plane coordinate system. The output is integrated.
[0068]
That is, in the congestion degree survey device according to the fourth embodiment, a plurality of sets of the congestion degree survey devices according to the third embodiment are used for a large survey target area, and each congestion degree measurement device obtained from each congestion degree measurement device is used. The congestion degree at each point on the plane corresponding to the imaging range of the ITV camera is integrated and output in one large plan view.
[0069]
For example, a camera for bird's-eye view of an intersection is installed at each of the opposing corners, and a congestion degree survey device according to the third embodiment is configured to correspond to each camera. The congestion degree at each point on the plane to be integrated is integrated into a plan view of one intersection and output.
[0070]
As a result, the congestion level survey device has the effect that data can be centrally managed even when a large area is monitored using a plurality of ITV cameras, and one camera has a blind spot. There is an advantage that a congestion degree survey of a region over a large area can be performed in one plan view by investigating a point with another camera.
[0071]
For example, in FIG. 11, the entire intersection cannot be monitored in the imaging range (trapezoidal portion) of one surveillance camera (ITV camera 10) installed at the corner of the intersection (crossroad). The surveillance area when the surveillance cameras (ITV cameras 10, 10) are installed at opposite corners covers the entire intersection.
[0072]
[Fifth Embodiment]
In the fifth embodiment of the present invention, the congestion degree of a plurality of video images is integrated and output in a plane coordinate system.
[0073]
In the fifth embodiment, recording is performed using a video recording device corresponding to a plurality of ITV cameras, instead of a plurality of sets of the “congestion degree investigation device according to the third embodiment” in the fourth embodiment. Then, these are sequentially reproduced and input by one congestion degree investigation device, and the congestion degree investigation outputs in the individual photographing ranges are integrated and output on one plane.
[0074]
Thereby, in addition to obtaining the same effects as the fourth embodiment, there is an advantage that the congestion degree investigation device is configured as one set, so that the device can be saved.
[0075]
[Sixth Embodiment]
In the sixth embodiment of the present invention, the degree of congestion by a turning camera is integrated and output as a panoramic image.
[0076]
The congestion degree research apparatus according to the sixth embodiment uses the basic components shown in FIG. 1, and performs the following photographing operation and image processing operation at that time. That is, the ITV camera 10 is installed and used on a pan head that moves in a horizontal plane, and the pan operation repeats the turning operation while stopping for a predetermined time at a plurality of preset points. The change detecting means 30 and the change number accumulating means 40 individually perform the change detection for each small area and the change number accumulating for each stop position within the time when the camera platform is stopped. The local change time rate calculation means 50 individually calculates the local change time rate for each small area for each stop position, and the output / recording means 60 integrates these local change time rates for each stop position into one panoramic image. Output or record as congestion degree. FIG. 13 is a diagram illustrating the combination of panoramic images.
[0077]
When a congestion degree survey of a monitoring area is performed using a monitoring camera installed for the purpose of turning one ITV camera 10 to monitor a wide area, the image of the ITV camera itself changes in the conventional method. Although the method of detecting the change area in the video and determining the degree of congestion could not be applied, according to the method of the sixth embodiment, the degree of congestion in an area covering the entire operation range of the surveillance camera can be investigated. It is.
[0078]
In addition, if the panoramic image is generated by integrating the ITV camera images for each stop position into one and the congestion degree survey result output of the present embodiment is overlaid on the panoramic image, the correspondence between the point of interest and the congestion degree is intuitive. Easy to understand.
[0079]
[Seventh Embodiment]
In the seventh embodiment of the present invention, the degree of congestion by a turning camera is integrated and output in a plane coordinate system.
[0080]
In the seventh embodiment, in the congestion degree survey device according to the sixth embodiment, the output / recording means 60 performs projection conversion to a plane of a survey target location for each stop position, and obtains an image of each stop position obtained from these. The congestion degree for each point on the plane corresponding to the range is integrated into one large plan view and output.
[0081]
This makes it possible to investigate the congestion degree in an area covering the entire imaging range of the surveillance camera, and to directly evaluate the congestion degree in the investigation target location by comparing it with a plan view. Further, the configuration is simpler because one camera is used as compared with the method of using a plurality of cameras of the fourth and fifth embodiments.
[0082]
[Eighth Embodiment]
In the eighth embodiment of the present invention, the congestion degrees of a plurality of turning cameras are integrated and output in a plane coordinate system.
[0083]
In the eighth embodiment, a plurality of sets of the congestion degree investigating apparatus of the seventh embodiment are used, and the congestion degree at each point on the plane corresponding to the imaging range of each ITV camera obtained from the set is further reduced. The output is integrated into two large plan views.
[0084]
This makes it possible to investigate the congestion degree in a wider area covering the entire imaging range of a plurality of monitoring cameras, and to directly evaluate the congestion degree at the investigation target location by comparing it with a plan view.
[0085]
Further, compared with the method of using a plurality of cameras according to the fourth and fifth embodiments, there is an advantage that a smaller number of cameras are used, and one turning camera according to the sixth and seventh embodiments can be used. The method used has an advantage that a congestion degree survey of a wide area can be conducted in one plan view by investigating a spot which has been a blind spot with another turning camera.
[0086]
[Ninth embodiment]
The ninth embodiment of the present invention is to output the congestion degree result in partial display.
[0087]
In the ninth embodiment, in the congestion degree output / recording means 60 in the congestion degree survey devices according to the first to eighth embodiments, the area corresponding to the small area having the same local change time rate is set to the same area. When a so-called color chart display in which colors are drawn is used, only a part of a corresponding area is drawn and an ITV camera input image (an image obtained by the ITV camera 10) or an ITV camera input image is projected on a survey target plane. Display an overlay on the converted image. That is, instead of coloring the entire target region, as shown in FIG. 14, for example, only the peripheral portion, the center portion, and only the 1/4 upper right portion of the small region to be investigated are colored as shown in FIG. As shown in the figure, a diagram that can be referred to is output together with the ITV camera image of the survey target location.
[0088]
As a result, although the correspondence between the point of interest and the degree of congestion can be easily understood by overlaying the image on the ITV camera image in the past, the color chart display has a drawback that it is difficult to visually recognize the point of interest because it is hidden by the color display. In the ninth embodiment, such disadvantages are eliminated, and there is an advantage that the correspondence between the point of interest and the degree of congestion can be easily visually and intuitively understood.
[0089]
[Tenth embodiment]
In the tenth embodiment of the present invention, the result of the congestion degree is output as a bar display.
[0090]
In the tenth embodiment, instead of the partial color chart of the ninth embodiment, a length proportional to the local change time rate as shown in FIG. As shown in FIG. 17, the stick is overlaid on the input image of the ITV camera or an image obtained by projecting and transforming the input image on the survey plane.
[0091]
For example, a bar having a length of 8 pixels is displayed in a small region having a local change time rate of 80%, and a bar having a length of 1 pixel is displayed in a small region having a local change time ratio of 10%. The bar is not displayed in the corresponding area in the small area where the change time ratio is 0%.
[0092]
This makes it difficult to read the degree of congestion for the contour lines of the conventional method, and it is difficult to understand the degree of congestion because it is necessary to know the correspondence between the color display and the degree of congestion in the color chart display. In the present embodiment, by displaying this in a bar graph, the degree of local congestion for each small area can be intuitively understood. Also, in the overall area of the congestion degree survey result output, the area where the bar graph occupies a large area has a high congestion degree, and the area where the bar graph has a small area is a point where the movement of vehicles and people is small. There is an advantage that the result can be easily grasped.
[0093]
[Eleventh embodiment]
In the eleventh embodiment, a congestion degree result is output as a motion bar, and stillness is output as a bar.
[0094]
In the eleventh embodiment, the “stationary obstacle detection method” disclosed in JP-A-2002-24808 is adopted as the change detection unit 30 used in the congestion degree investigation device. The contents of the disclosed technology will be described later. In the eleventh embodiment, a congestion degree research apparatus that counts a change state and a stationary state detected for each small area for a certain period of time and calculates and outputs a local change time rate and a local residence time rate for each small area. In the above, the output / recording means 60 projects two sticks having a length proportional to the local change time rate and the local dwell time rate for each totaled small area by projecting and transforming the ITV camera input image or the same onto the survey plane. Overlay on the selected image.
[0095]
For example, as shown in FIG. 18, a bar graph indicating the local change time ratio for each small region is displayed in green, and a bar graph indicating the local residence time ratio is displayed in red.
[0096]
As a result, it is possible to intuitively understand the degree of local congestion including the portion where there is much movement and the portion where there has been a lot of time in each small area. Dynamic observations can be made, including flow and retention conditions. Also, if the state of the movement and the staying state are distinguished by the color of the bar graph, the portion where the area occupied by the green bar graph is large in the entire area is high in congestion degree, but the area of the red bar graph is large. The part has an advantage that it is easy to grasp the result such as the point where the movement of the vehicle or the person stagnates.
[0097]
[Description of open technology]
Here, the contents of the disclosed technique “stationary obstacle detection method” disclosed in JP-A-2002-24808 will be described.
[0098]
The gist of the above-mentioned public technique (Japanese Patent Application Laid-Open No. 2002-24808) is that an image sampled from a moving image is divided into small regions, and a background image without an obstacle and no moving object prepared for each small region is prepared. The calculation of the normalized correlation value with the corresponding region is repeatedly performed on the image sampled at regular intervals, and the change state of the correlation value is determined by a threshold, and the background is preferably determined based on the determination. The object of the present invention is to detect a stationary obstacle while appropriately updating an image.
[0099]
This disclosed technique will be described in more detail. In the disclosed technology, as shown in FIG. 19, an image (current image) sampled from a moving image including a moving object is divided into small regions, and each small region does not include a moving object prepared in advance, and does not include an obstacle. The calculation of the normalized correlation with the corresponding small area of the background image where no is present is repeated for images sampled at regular intervals.
FIG. 19 shows an image divided vertically and horizontally into small areas (height L, width K).
The formula for calculating the normalized correlation is, for example, formula (11).
[0100]
[Equation 11]

[0101]
FIG. 20 shows an example in which the change over time of the correlation value of a certain small area is graphed. As shown in the section A of FIG. 20, if the change amount of the correlation value is large for a certain small area, it is determined that the moving object is passing, and as shown in the sections B and C of FIG. If the state in which the amount of change in the value is small continues for a certain number of times or more, it can be determined that the background is reflected in the small area.
If the correlation value of the small area regarded as the background in this way is a value close to 1 as shown in the section B of FIG. 20, the image is the same as the corresponding small area of the background image prepared in advance. become.
[0102]
However, as shown in a section C of FIG. 20, if the threshold value is smaller than a certain threshold value, it can be determined that there is a stationary object not shown in the background image prepared in advance in this small area.
[0103]
Further, for a small area in which the state in which the amount of change in the correlation value is small continues for a certain number of times or more and the correlation value is close to 1, the small area in the current image is replaced with the corresponding small area in the background image, Update the image.
[0104]
[Twelfth embodiment]
In the twelfth embodiment of the present invention, the congestion degree is normalized by the maximum value of the local change time ratio and is output in a bar display.
[0105]
In the twelfth embodiment, for the bar graph display output of the congestion degree survey devices of the tenth and eleventh embodiments, the length of the bar graph to be displayed for each small region, the local change time rate of the corresponding small region The length is proportional to the ratio of the local change time rate to the maximum value of the entire investigation area. For example, when the maximum value of the local change time rate of the entire survey area is 20%, a bar having a length of 10 pixels is provided in the largest small area, and the length is provided in a small area having a local change time rate of 10%. A 5-pixel bar is displayed in a region corresponding to a small pixel having a local change time ratio of 4% and a 2-pixel bar in length.
[0106]
As in the case of the tenth and eleventh embodiments, simply outputting the bar graph makes the bar graph length shorter as a whole at survey places and time zones where there is relatively little change. Although it is supposed that the congestion degree cannot be compared for each small area due to the length, according to the twelfth embodiment, the maximum value of the local change time rate is used even at a point and a time zone where the motion is small in the video. If the local change time ratio is different for each small area, the bar graph display reflects the difference, so that the degree of local congestion can be intuitively understood.
[0107]
Also, in the overall area of the congestion degree survey result output, the part where the area occupied by the bar graph is large is the congestion degree, and the part where the bar graph is small is the point where the movement of vehicles and people is small. You can grasp the results such as.
[0108]
[Thirteenth embodiment]
In the thirteenth embodiment of the present invention, the degree of congestion is normalized by the maximum value of the local change time ratio, further converted into a function, and output as bar display.
[0109]
In the thirteenth embodiment, for the bar graph display output of the congestion degree survey devices of the tenth and eleventh embodiments, the length of the bar graph to be displayed for each small region, the local change time rate of the corresponding small region, The length is calculated by a linear conversion formula set in advance based on the ratio of the local change time rate to the maximum value of the entire investigation area.
[0110]
For example, the conversion formula is y = x ³ When the maximum value of the local change time rate of the entire investigation area is 20%, a bar having a length of 10 pixels is used for the maximum small area, and the length is set to 7 for a small area whose local change time rate is 90%. Pixel (10 pixels x 0.9 ³ ), A small region having a local change time rate of 60% is 2 pixels long (10 pixels × 0.6 pixels). ³ ) Shall be displayed in the corresponding area.
[0111]
Accordingly, in the case of the above example, in the tenth embodiment, the difference in the length of the bar graph between the local change time ratios of 60% and 90% is 3 pixels, whereas in the thirteenth embodiment, In particular, the difference is in the case where the local change time rate of each small area is concentrated during this period (that is, when the congestion degree in the investigation range is high between 60% and 90%). Easy to compare.
[0112]
Also, the conversion equation is y = x ^1/3 For example, when the maximum value of the local change time rate of the entire investigation area is 20%, a bar having a length of 10 pixels is set in the maximum small area and a small area having a local change time rate of 6% is set in the small area. Is 7 pixels long (10 pixels × 0.3 ^1/3 ) Is applied to a small area having a local change time ratio of 1% and a length of 3 pixels (10 pixels × 0.05). ^1/3 ) Is displayed in the corresponding area.
[0113]
As a result, while the difference in the length of the bar graph between the local change time ratio of 0% and 30% was 3 pixels in the tenth embodiment, it was 7 pixels in the thirteenth embodiment. It is easy to compare the difference in a case where the local change time rate of each small area is concentrated on small data of 30% or less.
[0114]
【The invention's effect】
In the present invention, the basic components are shared, and the following common effects can be obtained by the common basic components.
[0115]
(1) Investigation of local congestion degree by machine is possible.
{Circle around (2)} Congestion degree data without variation such as counting errors due to fatigue of the researcher can be obtained.
{Circle around (3)} Since a continuous congestion degree survey over a long period of time is possible, a daily fluctuation, a weekly fluctuation, a monthly fluctuation, and a yearly fluctuation of the congestion degree can be easily obtained.
(4) Simultaneous surveys at multiple points are possible, so that it is possible to grasp differences in the degree of local congestion and the flow of pedestrians.
(5) In addition to the point congestion degree, it is possible to investigate the congestion degree in each part of an area (a road, a passage, a sidewalk, a plaza, a gate, an intersection, an underground mall, etc.) in a range photographed by an ITV camera. In the investigation of vehicles passing through roads and intersections, it is possible to investigate the degree of congestion including vehicles crossing lanes.
(6) The same device can be applied to both people and cars.
(7) No need to install traffic counters or dedicated gates.
(8) Correctly reflect the degree of congestion available for statistical analysis.
(9) Visual observation of the degree of congestion is possible by observing the video of the ITV camera on the monitor.
[0116]
According to the first aspect of the present invention, small portions are excluded when a change is detected, so that the following effects are further obtained in addition to the effects of the basic components.
{Circle around (1)} Since objects smaller than the size of the vehicle or person to be investigated are excluded, the danger of small objects becoming errors in the output of the congestion degree investigation is eliminated.
[0117]
According to the second aspect of the present invention, the upper portion is cut when a change is detected by the oblique camera, so that the following effects are further obtained in addition to the effects of the basic components.
{Circle around (1)} When using an ITV camera with the optical axis set in an oblique direction, the change detection area is combined and the upper part is cut when a change is detected, so that the output of the congestion degree survey is a flat congestion at the point of interest. It reflects the degree.
[0118]
According to the third aspect of the present invention, since the congestion degree is output after conversion into the plane coordinate system, the following effects are further obtained in addition to the effects of the basic components.
{Circle around (1)} In the present invention, since the congestion degree is output for each point of interest on the plane of the survey target place by converting the congestion degree output into coordinate plane coordinates, the congestion degree at the survey target place can be directly evaluated. Become.
[0119]
According to the fourth aspect of the present invention, the congestion degrees of a plurality of cameras are integrated and output in a plane coordinate system, so that the following effects are further obtained in addition to the effects of the basic components.
{Circle around (1)} Even when monitoring a wide area using a plurality of ITV cameras, the congestion degree survey device integrates the congestion degree survey results using the plurality of ITV cameras into one plane, so that the data Centralized management becomes possible.
{Circle around (2)} Further, by using another camera to investigate a spot that was in a blind spot with one camera, it is possible to conduct a congestion degree survey of a wide area in one plan view.
[0120]
According to the fifth aspect of the present invention, the congestion degree of a plurality of video images is integrated and output in a plane coordinate system, so that the following effects are further obtained in addition to the effects of the basic components.
{Circle around (1)} In addition to obtaining the same effects as the basic constituent elements and the invention of claim 4, since the configuration has one set of the congestion degree survey result apparatus, the apparatus can be saved.
[0121]
According to the sixth aspect of the present invention, the congestion degree of the turning camera is integrated and output as a panoramic image. Therefore, in addition to the effects of the basic components, the following effects are further achieved.
(1) The entire operation range of the surveillance camera is covered even when the congestion degree of the surveillance area is investigated using the surveillance camera installed for the purpose of turning one ITV camera to monitor a wide area. The degree of congestion in the area can be investigated.
{Circle around (2)} A panorama image is generated by integrating the ITV camera images for each stop position into one, and the congestion degree survey result output of the present embodiment is overlaid on the panoramic image. Easy to understand intuitively.
[0122]
According to the seventh aspect of the present invention, the congestion degree of the turning camera is integrated and output in the plane coordinate system. Therefore, in addition to the effects of the basic components, the following effects are further exhibited.
{Circle around (1)} It is possible to investigate the degree of congestion in an area covering the entire imaging range of the turning surveillance camera.
{Circle around (2)} It is possible to directly evaluate the congestion degree at the survey target location by comparing it with the plan view.
{Circle around (3)} As compared with the method using a plurality of cameras according to claims 4 and 5, the configuration is simple because one camera is used.
[0123]
According to the eighth aspect of the present invention, the congestion degrees of the plurality of turning cameras are integrated and output in the plane coordinate system. Therefore, in addition to the effects of the basic components, the following effects are further obtained.
{Circle around (1)} It is possible to investigate the congestion degree in a wider area covering the entire imaging range of a plurality of monitoring cameras.
{Circle around (2)} It is possible to directly evaluate the congestion degree at the survey target location by comparing it with the plan view.
{Circle around (3)} As compared with the method using a plurality of cameras according to claims 4 and 5, the number of cameras to be used can be reduced.
{Circle around (4)} In the method using one turning camera according to claims 6 and 7, the blind spot is investigated by another turning camera, so that the congestion degree survey of a high area can be performed in one plan view. Can be implemented.
[0124]
According to the ninth aspect of the present invention, since the congestion degree result is output in partial display, the following effect can be obtained in addition to the effect of the basic components.
(1) Since only a part of the corresponding area is drawn and overlay-displayed on the ITV camera input image or an image obtained by projecting and transforming the image onto the survey plane, the color chart display hides the color display and makes it difficult to visually recognize the point of interest. And the correspondence between the point of interest and the degree of congestion can be visually and intuitively understood.
[0125]
According to the tenth aspect of the present invention, since the congestion degree result is output as a bar display, the following effect is further obtained in addition to the effect of the basic components.
{Circle around (1)} By displaying the local change time rate in a bar graph, the degree of local congestion for each small area can be intuitively understood.
(2) Overall, the area where the bar graph occupies a large area in the whole area of the congestion degree survey result output area has a high congestion degree, and the area where the bar graph has a small area is a point where the movement of vehicles and people is small. It is easy to grasp the result.
[0126]
According to the eleventh aspect of the present invention, since the congestion degree result is moved and the stillness is output as a bar display, the following effects are further obtained in addition to the effects of the basic components.
(1) Since the local change time rate is displayed as a separate bar graph for moving and stationary, intuitively as to the degree of local congestion, including the portion with much movement and the portion with long staying time, for each output small area. It becomes possible to understand.
(2) If the state of the movement and the staying state are distinguished by the color of the bar graph, the portion where the area occupied by the yellow bar graph in the whole area is large is highly congested, but the area of the red bar graph is large. The large part has an advantage that it is easy to grasp the result such as the point where the movement of the vehicle or the person stagnates.
[0127]
According to the twelfth aspect of the present invention, the congestion degree is normalized by the maximum value of the local change time ratio and is output in the form of a bar, so that the following effects are obtained in addition to the effects of the basic components.
(1) Since the degree of congestion is normalized by the maximum value of the local change time rate and output, the bar graph is displayed on the basis of the maximum value of the local change time rate even at points and time zones where there is little movement in the video, so that each small area is displayed. If there is a difference in the local change time ratio, a bar graph display reflecting the difference is displayed, so that the degree of local congestion can be intuitively understood.
(2) Overall, the area where the bar graph occupies a larger area in the entire area of the congestion degree survey result output is the higher congestion degree, and the area where the bar graph is smaller is a point where the movement of vehicles and people is less. You can understand the results of things.
[0128]
According to the thirteenth aspect of the present invention, the congestion degree is normalized by the maximum value of the local change time ratio, further converted into a function, and output in the form of a bar. Therefore, the following effects are obtained in addition to the effects of the basic components.
{Circle around (1)} Since the local change time rate of the corresponding small area is displayed with a length calculated by a linear transformation equation set in advance based on the ratio of the local change time rate to the maximum value of the local change time rate of the entire survey area, particularly, It is easy to compare the difference in the local change time rate in a case where the local change time rate is concentrated on the higher or lower one.
{Circle around (2)} By appropriately selecting the conversion formula of the function conversion, it is possible to cope with both the case where the local change time rate is concentrated on the higher side and the case where the local change time rate is concentrated on the lower side.
[Brief description of the drawings]
FIG. 1 is a configuration diagram showing a congestion degree investigation device according to an embodiment of the present invention.
FIG. 2 is an image diagram showing an example of an ITV camera image of a congestion degree investigation place.
FIG. 3 is a configuration diagram illustrating an example of a change detection unit.
FIG. 4 is an image diagram showing an example of a congestion degree survey result displayed in a color chart.
FIG. 5 is an image diagram showing an example of a video image of an ITV camera installed obliquely in a place where the degree of congestion is excessive.
FIG. 6 is an image diagram showing an example of detecting a change in an ITV camera image.
FIG. 7 is an image diagram showing an example of an upper partial cut of change detection.
FIG. 8 is an explanatory diagram showing a relationship between a space coordinate system and a camera coordinate system.
FIG. 9 is an explanatory diagram showing camera parameters of a monitoring camera.
FIG. 10 is an image diagram showing an example of a conversion result to a plane coordinate system.
FIG. 11 is an image diagram showing an ITV camera at an intersection and its monitoring area.
FIG. 12 is an image diagram showing two ITV cameras and their monitoring areas.
FIG. 13 is an explanatory diagram showing a panoramic image.
FIG. 14 is an image diagram showing a partial display example of a congestion degree investigation result.
FIG. 15 is an image diagram showing a display example of a congestion degree investigation result by partial display.
FIG. 16 is an image diagram showing a bar graph display example of the congestion degree investigation result.
FIG. 17 is an image diagram showing an example of a result of a congestion degree investigation by bar graph display.
FIG. 18 is an image diagram showing an example of a bar graph display of a result of a congestion degree investigation in which movement and stay are distinguished.
FIG. 19 is an explanatory diagram showing a method for calculating a normalized correlation of a corresponding small region.
FIG. 20 is a characteristic diagram showing a time transition of a normalized correlation value with a background image.
[Explanation of symbols]
10 ITV camera
20 Image input means
30 Change detection means
40 means for integrating the number of changes
50 Local change time rate calculation means
60 output / recording means

Claims (13)

A TV camera that shoots the location of the congestion degree survey as a shooting range and outputs video,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
The change detection means sets the size of the survey object in advance, and when detecting a change in the image for each small region in the image, the small region in which the change is detected is detected in the vicinity of the small region in which the change is detected. The size of the entire change detection portion is calculated by coupling, and if the size of the entire change detection portion is smaller than the size of the survey object set in advance, the size of the small region included in the entire change detection portion is determined. A congestion degree research apparatus characterized in that there is no change in the video and the video is excluded from the change count integration. A TV camera that shoots the location of the congestion degree survey as a shooting range and outputs video,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
The change detecting means sets in advance the number of upper exclusion small areas based on the height of the survey target in the video, and when detecting a change in the video for each of the small areas in the video, the change detection means The small areas where a change is detected around the vertical direction are connected, and the small area included in the range of the number of the excluded small areas from the upper end of the connected small areas is excluded from the change number accumulation as there is no change in the image. A congestion degree investigation device characterized in that: A TV camera that shoots the location of the congestion degree survey as a shooting range and outputs video,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
The output / recording means projects the imaging range of the TV camera onto a plane of a congestion degree investigation target location, and outputs the congestion degree as a congestion degree for each point on the plane of the investigation target location. Degree survey device. A plurality of sets of the congestion degree investigation device according to claim 3 are provided, and the congestion degree at each point on a plane corresponding to the imaging range of each television camera obtained from each congestion degree measurement device is integrated into one large plan view. A congestion degree investigation device characterized in that the congestion degree is output. A TV camera that shoots the location of the congestion degree survey as a shooting range and outputs video,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
In place of the television camera, the video camera further includes a video playback unit that plays back a video of a target portion of the congestion degree survey, and the video input unit inputs a video every time a video output from the video playback unit is updated. ,
The output / recording unit projects the imaging range onto the plane of the congestion degree investigation target location, and outputs the congestion degree as the congestion degree for each point on the plane in the investigation target location,
A plurality of videos taken and recorded by a plurality of television cameras for a large survey area are reproduced by one video reproducing unit, and the output / recording unit outputs congestion degree survey output in each photographing range into one. A congestion degree investigation device characterized by integrating and outputting on a plane. A TV camera that shoots the location of the congestion degree survey as a shooting range and outputs video,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
The television camera is installed on a pan head that turns in a horizontal plane, and the pan head repeats a turning operation while stopping at a plurality of preset points for a predetermined time, the change detection unit and the change count integration unit. Performs change detection for each small region and integration of the number of changes individually for each stop position during the time when the camera platform is stopped, and the local change time rate calculating means individually performs local change for each small region for each stop position. A congestion degree surveying apparatus, wherein a change time rate is calculated, and the output / recording means integrates the local change time rates for each stop position into one panoramic image and outputs or records the congestion degree. A TV camera that shoots the location of the congestion degree survey as a shooting range and outputs video,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
The television camera is installed on a pan head that turns in a horizontal plane, and the pan head repeats a turning operation while stopping at a plurality of preset points for a predetermined time, the change detection unit and the change count integration unit. Performs change detection for each small region and integration of the number of changes individually for each stop position during the time when the camera platform is stopped, and the local change time rate calculating means individually performs local change for each small region for each stop position. The output / recording means calculates the rate of change, and performs projection conversion to the plane of the inspection target location for each stop position, and obtains the congestion degree for each point on the plane corresponding to the imaging range of each stop position obtained from these. A congestion degree investigation device characterized in that it is integrated and output in one large plan view. A plurality of sets of the congestion degree survey device according to claim 7 are provided,
A congestion degree investigation apparatus characterized in that the congestion degree for each point on a plane corresponding to the imaging range of each television camera obtained from each congestion degree investigation apparatus is further integrated into one large plan view and output. The congestion degree investigation device according to any one of claims 1 to 8,
The output / recording unit draws only a part of the corresponding area and draws only a part of the area when the area corresponding to the small area having the same local change time rate is displayed as a color chart that draws the same color. A congestion degree surveying apparatus characterized in that the obtained color chart display is overlaid on a picture obtained by a television camera or an image obtained by projecting and transforming this picture on a plane to be examined. The congestion degree investigation device according to any one of claims 1 to 8,
The output / recording means is an image obtained by a television camera or an image obtained by projecting and converting this image on a plane to be investigated on a bar having a length proportional to the local change time rate for each small area in which the local change time rate is totaled. A congestion level investigation device characterized in that an overlay display is provided. A TV camera that shoots the location of the congestion degree survey as a shooting range and outputs video,
Video input means for inputting a video every time a video output from the television camera is updated,
As many as the number of small areas set in the video are arranged, the video of each small area is input individually, and the presence or absence of change of the input video divided for each small area is inspected respectively. A plurality of change detecting means for
The number of changes is integrated in a one-to-one correspondence with each of the change detecting means, and the number of changes detected in the video detected by the correspondingly connected change detecting means is integrated every predetermined fixed time. Means,
It is connected to each of the number-of-changes integrating means in a one-to-one correspondence. A plurality of local change time rate calculating means for calculating a local change time rate which is a ratio with a number;
In a congestion degree investigation device having an output and recording means for recording and outputting the local change time rate calculated by each local change time rate calculation means as the congestion degree to the outside,
The change detection means calculates a local change time rate and a local residence time rate for each small area by counting the change state and the stationary state detected for each small area for a certain period of time,
The output / recording means sets two bars having a length proportional to each of the local change time rate and the local residence time rate for each of the small areas, and displays the image obtained by the television camera or the image on the survey target plane. A congestion degree investigation device characterized by displaying an overlay on an image subjected to projection conversion. The congestion degree investigation device according to claim 10 or 11,
Regarding the bar graph display output by the output / recording means, the length of the bar graph displayed for each small area is such that the local change time rate of the corresponding small area is proportional to the ratio of the local change time rate of the entire survey area to the maximum value. A congestion degree investigation device, characterized in that: The congestion degree investigation device according to claim 10 or 11,
For the bar graph display output by the output / recording means, the length of the bar graph displayed for each small area is set in advance based on the ratio of the local change time rate of the corresponding small area to the maximum value of the local change time rate of the entire survey area. A congestion degree investigation device characterized in that the length is calculated by a linear conversion formula.

Images