JP2004024853A - Device for supporting walking - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a walking support device for reporting walking position information to a visually challenging person without using sound, and guiding the challenging person in a walking direction. <P>SOLUTION: This device includes: a position determining apparatus for determining a position; an input apparatus for inputting the fact that places are the start and terminal ends of a path, and the kind of the path; and a means for obtaining the positions of the start and terminal ends from the position determining device in response to the input of the fact that the place is the start or terminal end for each path and describing the obtained position and the inputted path kind for each path to generate a route information file. On the basis of the generated route information, the walking direction is guided through the use of a wrist band provided with vibrators or electric stimulators arranged in a matrix shape. <P>COPYRIGHT: (C)2004,JPO

Description

【表２】

【００３３】
表１、２に示す通り、制御パターンは、歩行支援装置が正常に動作し、通知を開始するとき（コード２２）、目的地に到達したとき（コード５５）、点字ブロック又は横断歩道が検出できないとき（コード２４６２４６）、故障のとき（コード１２３４５６１２３４５６）等である。
【００３４】
データパターンは、例えば点字ブロックに沿って直進を指示し、右側に車道等があり危険であることを知らせる（コード２５３６３６）、点字ブロックに沿って直進すると間近に左右に分岐点がることを知らせる（コード２５１２３）、点字ブロックに沿って斜め右方向に進めを知らせる（コード３５）、近くに障害物があることを知らせる（コード１２３１２３）等である。
【００３５】
５．ティーチング
図４にティーチングのフローチャートを示す。オペレータは携帯端末３にキーボード４２を接続し、電源をオンにする。携帯端末４に表示される画面からティーチングモードを選択する（Ｓ１）。
【００３６】
初期値としてパス番号に”０”を入力する（Ｓ１）。次に始端番号に１を加え、始端の位置情報として携帯端末４に接続するＧＰＳ８から位置情報（グローバル座標と位置情報）、又は無線タグから位置情報を取得し、歩数計５のカウンター値として”０”を入力する（Ｓ２）。次に、パスの種類、例えば点字ブロックである旨を入力し（Ｓ３）、パスのメッセージ（例えば右側には側溝がある）があれば入力する（Ｓ４）。
【００３７】
オペレータは経路に沿って歩行を開始する（Ｓ５）。オペレータが点字ブロックの分岐や横断歩道に到達したときの地点がパスの終端となる。パス終端の位置情報はＧＰＳ８から取得するグローバル座標、又は無線タグから取得する位置情報と、方位、及び歩数計５のカウンター値である。これらの値は画像処理装置１０に自動的に入力される（Ｓ６）。
【００３８】
次に、予告振動パターンを入力し（Ｓ７）、パス終端に名称がある場合、例えば山梨大学前といった名称を必要に応じて入力する（Ｓ８）。予告振動パターンは、パスの終端にきたとき、次にどの方向に進むべきか、左右どちら側が危険か、一次停止すべき等の情報を利用者に知らせるためにパスの終端に近づいたときに提示する。パスの終端は次のパス始端となり、Ｓ２からＳ９を目的地点まで繰り返す。目的地点の到達した場合はキーボード４２から終了キーを入力し、ティーチングを終了する（Ｓ９）。
【００３９】
６．誘導処理
図５に誘導処理のフローチャートを示す。視覚障害者は歩行支援システムの電源をオンし、携帯端末４から音声アナウンスに従い目的地を選択する（Ｓ２１）。出発地点のグローバル座標と方位はＧＰＳから自動取得される。選択された目的地と出発地点とから経路情報が自動作成される（Ｓ２２）。経路情報は画像処理装置１０のメモリ内と管理センター７のデータベースを検索し行われる。
【００４０】
経路情報が生成されると、その旨がスピーカにより視覚障害者に通知され、視覚障害者はスタートボタン１５を押す（Ｓ２３）。画像処理装置１０はパス番号１を読み取り（Ｓ２４）、パスの始端に関連付けられたメッセージがあればスピーカ４１により通知する。
【００４１】
ビデオカメラ２により現在地の画像を取得し、経路情報のパスの種類に基づき画像処理して振動パターン（観測振動パターン）を生成する（Ｓ２５）。経路情報から得られる振動パターンと観測振動パターンとが一致するまで、所定の回数処理を繰り返す。振動パターンが一致するとリストバンドによりその振動パターンが提示され（Ｓ２６）、所定回数処理を繰り返しても、なお、振動パターンが一致しない場合は故障パターン（１２３４５６１２３４５６）を提示する。
【００４２】
図６は誘導処理の観測パターン（Ｓ４１〜Ｓ４５）を提示するフローチャートである。ソナー６により近くに障害物があるかどうかを検出し（Ｓ４１）、直前に障害物がある場合は振動パターン２２を提示する（Ｓ４２）。近くに障害物がある場合は振動パターン（コード１２３１２３）を提示する（Ｓ４３）。
【００４３】
パスの種類（点字ブロック又は横断歩道）を判別し、点字ブロックの場合は点字ブロックである旨を音声で通知し、横断歩道の場合はその旨を音声で通知する（Ｓ４４，Ｓ４５）。
【００４４】
予告地点に到達すると予告パターン（コード２２）が通知される（Ｓ２８）。予告地点はパスの終端の少し手前に設定する。例えば、パス１の区間は歩数計５のカウンター値から一定値を引き算した地点とする。例えばカウンター値が１００であった場合には９０を予告地点とする。また、パスの終端のグローバル座標値から数メートル手前の座標値を予告地点としてもよい。どちらを取るかは周囲の環境条件により決定する。
【００４５】
予告地点に到達したら、予告振動パターン（例えば、点字ブロックの分岐を知らせるコード２５１２３）と同一のパターンをビデオカメラ２により検出する。
【００４６】
振動パターンの指示した方向に歩行、徐行または停止する（Ｓ３０）。検出できない場合は経路から外れた可能性があり、管理センター７に音声で通知する（Ｓ３２）。
【００４７】
視覚障害者は観測振動パターンに従い歩行支援装置が誘導する方向に進み、提示された振動パターンに相当する、例えば点字ブロックの分岐点等を白杖で見つけ、歩行を継続する。これを目的地点まで繰り返す。
【００４８】
８．経路情報データ構造
図７に経路情報のデータ構造の一例を示す。経路情報はパスの集合体である。パスにはパス番号（英数字）が付与される。経路情報１はパス番号（Ｐ１）からパス番号４（Ｐ４）で構成される。Ｐ１を例にとって説明する。種類は点字ブロックか横断歩道かの識別を表し整数値である。始端、終端の位置情報は、ＧＰＳ８からＸＹ座標として実数値、歩数計からはパス区間の歩数値（整数）である。オペレータにより入力された終端の予告振動パターンは整数、メッセージは文字列である。なお、ＧＰＳ８から取得する位置情報に代えて、無線タグから取得する位置情報であっても良い。
【００４９】
パスの終端番号は空白又は整数、パスの終端名称は空白又は文字列で記録される。Ｐ２〜Ｐ４も同様であり、Ｐ１〜Ｐ４の集合が経路情報１となる。
【００５０】
経路情報２はパス番号１（Ｐ１）、パス番号２（Ｐ２）、パス番号１２（Ｐ１２）、パス番号１３（Ｐ１３）で構成される。経路情報２は経路情報１のＰ１とＰ２を使い、Ｐ２の終端から分岐するＰ１２とＰ１３で構成されている。経路情報は出発地点と目的地点とから、管理センター７のデータベースから経路を探索しパスを系列として生成する。
【００５１】
９．画像処理
（１）点字ブロック、滑り止めブロックの検出
図８に点字ブロックの検出フローチャート図、図９（ａ）に点字ブロック画像、図９（ｂ）Ｕ分布、図９（ｃ）Ｖ分布、図９（ｄ）Ｕ−Ｖ分布を示す。図１０は２値化画像から振動パターンの生成する図、図１１に点字ブロックに沿って敷設されている滑り止めから振動パターンを生成する図を示す。表３、表４は２値化した画像を９ブロックに分割し特徴マトリクスとし、これを振動パターンに変換する特徴マトリクス・振動パターン変換テーブルである。
【００５２】
【表３】

【表４】

【００５３】
▲１▼点字ブロックの検出
点字ブロックを検出し、これを振動パターンに変換するには、先ず、ビデオカメラ２から点字ブロックの画像（原画像）を取り込み（Ｓ５１）、図９（ａ）に示すように点字ブロック画像の下部（白線）枠で示した部分）に閾値検出用ウィンドウを設け、図９（ｂ）赤色分布（又はＵ分布）と図９（ｃ）青色分布（又はＶ分布）を求め、赤色分布（又はＵ分布）から青色分布（Ｖ分布）を引き算した差分布を求めると図９（ｄ）になる。次に、黄色点字ブロックを検出するための差分閾値を求める（Ｓ５２）。原画分閾値で２値化し黄色点字ブロックを表す２値画像を求める。点字ブロックが見つからないときは、原画像の読み込みに戻る（Ｓ５２’）。
【００５４】
次に、２値化画像を図１０（ｃ）に示すように９ブロックに分割し、分割ブロック内に黒がある部分については”１”を、白の部分は”０”とする特徴マトリクスを作成する（Ｓ５３）。特徴マトリクスを表１、表２に示す特徴マトリクス・振動パターン変換テーブルにより、振動パターンに変換する（Ｓ５４）。
【００５５】
▲２▼滑り止めブロックの検出
滑り止めが視覚障害者の歩行方向に向かって左右のいずれにあるかは、滑り止め領域に特徴的に現れている平行の複数のラインを検出することで行う。例えば、ホームの微分画像は図１１にようになる。滑り止め領域とアスファルト領域の微分値の度数分布は図１１（ｂ）に示すように、滑り止め領域の微分値が大きな値となる（Ｓ５６）。そこで、図１１に示すように点字ブロック画像を９分割し、大きな微分値を持つ領域を滑り止めがある領域とする（Ｓ５７）。検出した結果を特徴マトリクスとし、これを特徴マトリクス・振動パターン変換テーブルにより振動パターンに変換する。以上により点字ブロック、滑り止めブロックを検出し、検出した点字ブロックの方向や滑り止の位置をリストバンド３により視覚障害者に通知し（Ｓ５９）、歩行誘導を行う。
【００５６】
（２）横断歩道の検出
図１２は横断歩道の検出フローチャート、図１３は横断歩道の画像、画像処理により作成した濃淡ヒストグラム、および特徴マトリクス、振動パターンを示したものである。
【００５７】
先ず、横断歩道の原画像を取り込み（Ｓ６１）、点字ブロックの検出と同様に検出ウィンドウにより画像全部について濃淡ヒストグラムを作成し（Ｓ６２）、判別分析法により横断歩道マーク検出閾値を算出する（Ｓ６２）。点字ブロックが見つからないときは、原画像の読み込みに戻る（Ｓ６２’）。
【００５８】
閾値から横断歩道の２値化画像を作成し（Ｓ６３）、横断歩道マークの幅Ｗとギャップ間隔Ｇとを算出する（Ｓ６４）。振動パターンの生成は、図１３に示すように、横断歩道の原画像を２値化画像とし、２値化画像を９個ブロックに分割し、その分割領域が白と黒の占有率から特徴マトリクスを作成し、これを特徴マトリクス・振動パターン変換テーブルから振動パターンを生成する（Ｓ６５〜Ｓ６８）。
【００５９】
１０．視覚障害者の誘導の例示
図１４は駅ホームを視覚障害者が歩行する際、歩行支援装置がどのように誘導するのかを示したものである。図１４において、視覚障害者は図の左端から右に向かって歩行している。このとき、視覚障害者の眼鏡に設置してあるビデオカメラ２は、前方１ｍ〜１０ｍの範囲を捉えている。ビデオカメラ２には警告点字ブロック、線路等が写っている。前述した方法により歩行方向に対し、警告点字ブロックの位置を判別し、リストバンド３に設置されている振動子３１により振動パターンを通知する。
【００６０】
図１５は駅ホームにおける歩行誘導の振動パターンを示したものである。視覚障害者は点字ブロックに沿って直進する。しかし歩行方向に向かって右側に線路があることから右側が危険であることを通知することになる。そこで、振動パターンは表１に示すコード２５１４１４であり、その振動周期は図１５に示す通りである。１周期２回振動が振動子▲１▼と振動子▲２▼、１周期１振動が振動子▲３▼と振動子▲５▼である。
【００６１】
図１６は視覚障害者が駅の階段からホームに入り、電車の乗るまでの誘導処理の状況を示したものである。図１７は視覚障害者が横断歩道を渡るときの誘導処理を示したものである。図１８は視覚障害者が日常生活において訪れるであろう百貨店、喫茶店、市役所、病院を訪れるときの誘導処理を示したものである。
【００６２】
【発明の効果】
以上説明したとおり、本発明によれば音声を使わずに視覚障害者に歩行位置の情報や歩行方向を通知することができる。これにより視覚障害者の駅ホーム内の転落事故、交差点内における衝突事故等を防止すできる。更に、本発明は低廉な価格で視覚障害者の歩行を支援する歩行支援装置を提供する特有の効果がある。
【図面の簡単な説明】
【図１】歩行支援装置の構成を示した図。
【図２】リストバンドの構成を示した図。
【図３】リストバンドの断面を示した図。
【図４】ティーチングのフローチャート図。
【図５】誘導処理のフローチャート図。
【図６】誘導処理の観測パターンを提示するフローチャート図。
【図７】経路情報のデータ構造の一例を示した図。
【図８】点字ブロックの検出フローチャート図。
【図９】点字ブロックの検出方法を示した図。
【図１０】点字ブロックの２値化画像を振動パターンに変換する方法を示した図。
【図１１】滑り止めブロックを検出して振動パターンを生成する方法を示した図。
【図１２】横断歩道の検出フローチャート図。
【図１３】横断歩道を検出して振動パターンを生成する方法を示した図。
【図１４】駅ホームを視覚障害者が歩行する際、歩行支援装置がどのように誘導するのかを示した図。
【図１５】駅ホームにおける歩行誘導の振動パターンを示した図。
【図１６】視覚障害者が駅の階段からホームに入り、電車の乗るまでの誘導処理の状況を示した図。
【図１７】視覚障害者が横断歩道を渡るときの誘導処理を示した図。
【図１８】視覚障害者が日常生活において訪れるであろう百貨店、喫茶店、市役所、病院を訪れるときの誘導処理を示した図。
【符号の説明】
１　　歩行支援装置
２　　ビデオカメラ
３　　リストバンド
４　　携帯端末（ＰＤＡ）
５　　歩数計
６　　ソナーシステム
７　　管理センター
８　　ＧＰＳ
１０　画像処理装置
１１　ＣＰＵ
１２　メモリ
１３　画像処理チップ
１４　入出力チップ
１５　スタートボタン
３１　振動子
４１　スピーカ
４２　キーボード[0001]
【Technical field】
The present invention relates to a device that assists a visually impaired person when walking alone using a white cane, and in particular, detects a Braille block, a pedestrian crossing, a guide block of a station platform, and a track by image processing, and notifies by tactile sensation. The present invention relates to a walking support device for the visually impaired.
[0002]
[Prior art]
As a device to assist the walking of the visually impaired, a chip that transmits position information is incorporated in a Braille block, and when a special stick is brought close to the Braille block, information on the current location is wirelessly received from the chip, and a dedicated mobile terminal outputs synthesized speech. There is a device that gives directions and distance to the destination.
[0003]
As an apparatus for assisting the crossing of an intersection, there is an infrared voice information guidance system (talking sign) for giving a direction of walking by sound or notifying current position information by voice.
[0004]
As a method for preventing a visually impaired person from falling off the platform, a partition wall with a door is installed between the platform and the track in some cases.
[0005]
However, a device that transmits information by voice undesirably interferes with the sound localization of the visually impaired. In addition, it does not directly tell the position of the dangerous place in the walking direction of the visually impaired person. Furthermore, large-scale construction and installation of systems are expensive and time-consuming to spread nationwide.
[0006]
Therefore, an object of the present invention is to provide a walking assistance device that notifies a visually impaired person of walking position information without using a voice as much as possible and supports walking. Another object of the present invention is to provide a walking support device that directly informs the position of a track on a station platform and the walking direction when crossing an intersection with respect to the traveling direction of a visually impaired person. Another object of the present invention is to provide a walking assistance device having these functions at a low price.
[0007]
[Means for Solving the Problems]
The present invention provides a position determination device that determines a position, an input device that receives input of the start and end of a path and an input of a type of a path, and that the start and end of a path are input for each path. And a path information creating means for acquiring the start and end positions from the position determination device in accordance with the acquired position and describing the acquired position and the type of the input path for each path to create a path information file. According to the present invention, a healthy person determines the path while moving this walking support device, and automatically obtains at least the start and end points and the start and end position information for each path, without fail and without error. Route information can be easily created.
[0008]
The input device accepts input of information related to a message associated with a start and / or end, and the path information creating unit describes information related to the input message. Information about the message is input to the route information, and the visually impaired person is notified of the start and end points. Thereby, the visually impaired can easily grasp the surrounding situation.
[0009]
Path reading means for reading the start and end positions and path types described for each path in the path information file; read start and end positions and path types; current position and current position obtained from the input device; The walking direction is derived from the type of the path at the position. The start and end of the path acquired from the path information created by teaching, the type of path, the position acquired from the position determination device (start end, end, route en route), and the image acquired from the image processing device included in the input device By comparing with the type of path and guiding appropriately, the visually impaired can definitely walk on the intended path. If the vehicle deviates from the route, a warning is given to the visually impaired person by alarm.
[0010]
The walking direction inducing means arranges the vibrators or the electric stimulators in a matrix, changes the vibration pattern or the electric stimulus pattern according to the contents of the guidance, and guides by a sense of touch. Since visually impaired persons judge the surrounding situation from hearing, when walking assistance is carried out by voice, the grasp of the surrounding situation becomes confused. By using the tactile sense, walking of a visually impaired person can be supported more securely and reliably.
[0011]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, an embodiment in which the present invention is applied to a walking assistance device that assists a visually impaired person to reach a destination by relying on a white stick will be described in detail.
[0012]
1. Definition of terms The definitions of terms used in this specification are as follows.
Position determination device: Determines the position of the walking support device from latitude / longitude information obtained from GPS, position information obtained from a wireless tag, azimuth information obtained from a compass, and distance information obtained from a pedometer. apparatus.
[0013]
Route: The path from the starting point to the destination for the visually impaired.
[0014]
Pass type: Braille block, pedestrian crossing, or guide block on station platform.
[0015]
Path: A section between the start end and the end. In the case where the starting end is the starting point, the first end is the point where the Braille block branches or starts crossing the crosswalk. The second start end is the first end. The second end point is where the next Braille block branches or begins to cross the pedestrian crossing. The beginning and end are repetitions of this. The start and end of the path are determined by the person who creates the path information (operator).
[0016]
Route information: A sequence of information related to a path constituting a route from a starting point to a destination.
[0017]
Walking support: Refers to processing for assisting a visually impaired person in a walking direction from a starting point to a destination based on route information. Support in the walking direction is provided by tactile sense or voice at the beginning and end of the path.
[0018]
Teaching: A process in which an operator uses a walking support device to obtain route information while walking on a route. It is desirable that the operator is a healthy person who is familiar with the operation method of the walking assist device and at what time it is desired to notify the visually impaired person of what kind of information.
[0019]
Message: Information entered in connection with the path. For example, there is a gutter on one side of the path.
[0020]
2. FIG. 1 shows the configuration of a walking support device.
[0021]
The walking support device 1 includes an image processing device 10, a video camera 2, a wristband 3, a personal digital assistant (PDA) 4, a pedometer 5, a sonar system 6, a management center 7, a GPS 8, and the like. The individual modules in each section and their functions are as follows.
[0022]
Image processing device 10: Input / output interface with external devices such as an image processing chip 13 for processing images of braille blocks, images of pedestrian crossings, guidance blocks and railroad tracks of station platforms, and wrist bands input from video camera 2. , An input / output chip 14, a memory 12 for storing information indicating the start and end of the path and position information, and a CPU 11 for processing data and the like in the memory 12 and generating a signal for the wristband.
[0023]
Video camera 2: A camera that is installed on glasses, hats, etc. of the visually impaired and captures images such as braille blocks around the walking path.
[0024]
Wristband 3: A device for arranging transducers on a matrix and guiding the walking direction by touch. The vibrator is vibrated in a fixed pattern, and the direction of the Braille block, the direction of the guide block on the station platform, and the like are notified to the visually impaired.
[0025]
Mobile terminal (PDA) 4: a device that has a wireless communication function and a GPS function, or a function of reading position information from a wireless tag, and inputs and outputs data between the image processing apparatus 10 and the management center 7. Input of path information and the like is received from a keyboard 42 used during teaching, and a message is output from a speaker 41 by voice.
[0026]
Pedometer 5: The number of steps from the starting point is counted at each of the start and end points, and the position of the pedestrian is calculated. The count value is transmitted to the image processing device 10.
[0027]
Sonar system 6: a device that detects an obstacle on a walking route.
[0028]
Management center 7: Data and the like transmitted from image processing apparatus 10 are stored in a database as path information and stored, and path information is transmitted in response to a request from image processing apparatus 10. It also responds to emergency calls from pedestrians.
[0029]
GPS8: A device for acquiring the latitude and longitude of the start and end of the path, including a compass, and acquiring the position and orientation of the walking support device.
[0030]
3. Configuration of Wristband 3 FIG. 2 shows a configuration of a wristband using an electric vibrator. The wristband 3 is worn on the left and right wrists. If you have a white stick on your right arm, wear it on your left arm wrist. Six electric vibrators are arranged in a matrix at the center of the wristband, and a visually impaired person is guided by transmitting instructions of a walking direction such as straight ahead, right turn, left turn, and the like according to a vibration pattern. FIG. 3 shows a cross section of the wristband 3. The wristband is secured to the wrist with tape. A protruding needle is provided at the center to sharpen the vibration pattern.
[0031]
4. Vibration pattern tables 1 and 2 show the vibration patterns of the wristband 3. The vibration patterns are a total of 23 patterns of (1) control patterns (5 patterns) and (2) data patterns (18 patterns). The number of vibrations per period of the vibrator is (1) one time (*) and (2) two times (*).
[0032]
[Table 1]

[Table 2]

[0033]
As shown in Tables 1 and 2, the control pattern cannot detect a braille block or a pedestrian crossing when the walking assist device operates normally and starts notification (code 22), when it reaches the destination (code 55). Time (code 246246), failure (code 123456123456), and the like.
[0034]
The data pattern indicates, for example, that the vehicle is going straight along a braille block and informs that there is a roadway or the like on the right side and is dangerous (code 253636). (Code 25123), a notification indicating that the vehicle is proceeding diagonally right along the braille block (code 35), and a notification that an obstacle is nearby (code 123123).
[0035]
5. Teaching FIG. 4 shows a flowchart of the teaching. The operator connects the keyboard 42 to the mobile terminal 3 and turns on the power. The teaching mode is selected from the screen displayed on the mobile terminal 4 (S1).
[0036]
"0" is input as a pass number as an initial value (S1). Next, 1 is added to the start end number, and position information (global coordinates and position information) from the GPS 8 connected to the mobile terminal 4 or position information from the wireless tag is acquired as start position information, and the position information is obtained as a counter value of the pedometer 5 as “ "0" is input (S2). Next, the type of the path, for example, the fact that the block is a Braille block is input (S3), and if there is a path message (for example, there is a gutter on the right side), it is input (S4).
[0037]
The operator starts walking along the route (S5). The point at which the operator arrives at the branch of the Braille block or at the crosswalk is the end of the path. The position information at the end of the path is global coordinates obtained from the GPS 8 or position information obtained from the wireless tag, the direction, and the counter value of the pedometer 5. These values are automatically input to the image processing device 10 (S6).
[0038]
Next, a notice vibration pattern is input (S7), and if there is a name at the end of the path, a name, for example, in front of Yamanashi University is input as necessary (S8). The notice vibration pattern is presented when approaching the end of the path to inform the user when the vehicle arrives at the end of the path, in which direction to proceed next, which side is dangerous, whether to temporarily stop, etc. I do. The end of the path is the start of the next path, and S2 to S9 are repeated to the destination. If the destination has been reached, the end key is input from the keyboard 42 to end the teaching (S9).
[0039]
6. Guidance Process FIG. 5 shows a flowchart of the guidance process. The visually impaired person turns on the power of the walking support system and selects a destination from the portable terminal 4 according to the voice announcement (S21). The global coordinates and orientation of the departure point are automatically acquired from GPS. Route information is automatically created from the selected destination and starting point (S22). The route information is obtained by searching the memory of the image processing apparatus 10 and the database of the management center 7.
[0040]
When the route information is generated, the fact is notified to the visually impaired person by the speaker, and the visually impaired person presses the start button 15 (S23). The image processing apparatus 10 reads the pass number 1 (S24), and notifies the speaker 41 of a message associated with the start of the pass, if any.
[0041]
An image of the current location is acquired by the video camera 2, and image processing is performed based on the type of path in the route information to generate a vibration pattern (observation vibration pattern) (S25). The process is repeated a predetermined number of times until the vibration pattern obtained from the route information matches the observed vibration pattern. If the vibration patterns match, the wristband presents the vibration pattern (S26), and if the vibration pattern does not match even after repeating the process a predetermined number of times, a failure pattern (123456123456) is presented.
[0042]
FIG. 6 is a flowchart showing the observation patterns (S41 to S45) of the guidance process. The sonar 6 detects whether or not there is an obstacle nearby (S41). If there is an obstacle immediately before, the vibration pattern 22 is presented (S42). If there is an obstacle nearby, a vibration pattern (code 123123) is presented (S43).
[0043]
The type of path (Braille block or pedestrian crossing) is determined, and if it is a Braille block, it is notified by voice that it is a Braille block, and if it is a pedestrian crossing, it is notified by voice (S44, S45).
[0044]
Upon reaching the notice point, a notice pattern (code 22) is notified (S28). The notice point is set slightly before the end of the pass. For example, the section of pass 1 is a point obtained by subtracting a certain value from the counter value of pedometer 5. For example, if the counter value is 100, 90 is set as the notice point. Alternatively, a coordinate value several meters before the global coordinate value at the end of the path may be set as the notice point. Which one to take depends on the surrounding environmental conditions.
[0045]
When reaching the notice point, the video camera 2 detects the same pattern as the notice vibration pattern (for example, the code 25123 for informing the branch of the braille block).
[0046]
Walk, slow down or stop in the direction indicated by the vibration pattern (S30). If it cannot be detected, it may have deviated from the route and notifies the management center 7 by voice (S32).
[0047]
The visually impaired person proceeds in the direction guided by the walking assist device in accordance with the observed vibration pattern, finds, for example, a branch point of a Braille block, which corresponds to the presented vibration pattern, with a white cane, and continues walking. This is repeated up to the destination.
[0048]
8. FIG. 7 shows an example of the data structure of the route information. The route information is a set of paths. A pass number (alphanumeric) is assigned to the pass. The route information 1 is composed of a path number (P1) to a path number 4 (P4). This will be described by taking P1 as an example. The type is an integer value indicating the identification of a braille block or a pedestrian crossing. The start and end position information is a real value as XY coordinates from the GPS 8 and a step value (integer) of the path section from the pedometer. The notice vibration pattern at the end input by the operator is an integer, and the message is a character string. Note that, instead of the position information obtained from the GPS 8, position information obtained from a wireless tag may be used.
[0049]
The end number of the path is recorded as a blank or an integer, and the end name of the path is recorded as a blank or a character string. The same applies to P2 to P4, and a set of P1 to P4 is path information 1.
[0050]
The path information 2 includes a path number 1 (P1), a path number 2 (P2), a path number 12 (P12), and a path number 13 (P13). The route information 2 uses P1 and P2 of the route information 1 and is composed of P12 and P13 branched from the end of P2. As the route information, a route is searched from the database of the management center 7 based on the departure point and the destination point, and the paths are generated as a series.
[0051]
9. Image processing (1) Detection of Braille blocks and non-slip blocks FIG. 8 is a flowchart of detection of Braille blocks, FIG. 9A is a Braille block image, FIG. 9B is a U distribution, FIG. 9C is a V distribution, and FIG. 9 (d) shows a UV distribution. FIG. 10 is a diagram illustrating generation of a vibration pattern from a binarized image, and FIG. 11 is a diagram illustrating generation of a vibration pattern from non-slip laid along a Braille block. Tables 3 and 4 are feature matrix / vibration pattern conversion tables for dividing the binarized image into nine blocks to obtain a feature matrix and converting the feature matrix into a vibration pattern.
[0052]
[Table 3]

[Table 4]

[0053]
(1) Detection of Braille Block To detect a Braille block and convert it into a vibration pattern, first, an image (original image) of the Braille block is fetched from the video camera 2 (S51), and shown in FIG. 9A. As described above, a threshold detection window is provided in the lower part (white line) frame of the Braille block image, and the red distribution (or U distribution) in FIG. 9B and the blue distribution (or V distribution) in FIG. FIG. 9D shows the difference distribution obtained by subtracting the blue distribution (V distribution) from the red distribution (or U distribution). Next, a difference threshold value for detecting a yellow braille block is obtained (S52). A binary image representing the yellow braille block is obtained by binarization using the original fraction threshold value. If no Braille block is found, the process returns to reading the original image (S52 ').
[0054]
Next, the binarized image is divided into nine blocks as shown in FIG. 10C, and a feature matrix in which black portions in the divided blocks are set to “1” and white portions are set to “0”. It is created (S53). The feature matrix is converted into a vibration pattern using the feature matrix / vibration pattern conversion tables shown in Tables 1 and 2 (S54).
[0055]
(2) Detection of non-slip block Whether the non-slip is located in the left or right direction in the walking direction of the visually impaired person is determined by detecting a plurality of parallel lines characteristically appearing in the non-slip region. For example, the home differential image is as shown in FIG. In the frequency distribution of the differential values of the non-slip region and the asphalt region, as shown in FIG. 11B, the differential value of the non-slip region has a large value (S56). Therefore, as shown in FIG. 11, the Braille block image is divided into nine, and an area having a large differential value is set as a non-slip area (S57). The detected result is used as a feature matrix, which is converted into a vibration pattern by a feature matrix / vibration pattern conversion table. As described above, the braille block and the non-slip block are detected, the direction of the detected braille block and the position of the non-slip are notified to the visually impaired person by the wristband 3 (S59), and the walking is guided.
[0056]
(2) Crosswalk Detection FIG. 12 is a crosswalk detection flowchart, and FIG. 13 shows a crosswalk image, a density histogram created by image processing, a feature matrix, and a vibration pattern.
[0057]
First, an original image of a pedestrian crossing is taken in (S61), a grayscale histogram is created for the entire image by a detection window in the same manner as in the detection of a Braille block (S62), and a pedestrian crossing mark detection threshold is calculated by a discriminant analysis method (S62). . If no Braille block is found, the process returns to reading the original image (S62 ').
[0058]
A binarized image of the pedestrian crossing is created from the threshold value (S63), and the width W of the pedestrian crossing mark and the gap interval G are calculated (S64). As shown in FIG. 13, the generation of the vibration pattern is performed by using the original image of the pedestrian crossing as a binarized image, dividing the binarized image into nine blocks, and dividing the divided region into a feature matrix based on the white and black occupation rates. And a vibration pattern is generated from the feature matrix / vibration pattern conversion table (S65 to S68).
[0059]
10. Example of Guidance of Visually Impaired Person FIG. 14 shows how the walking assist device guides a visually impaired person when walking on a station platform. In FIG. 14, the visually impaired person is walking from the left end of the figure to the right. At this time, the video camera 2 installed on the glasses of the visually impaired captures a range of 1 m to 10 m in front. The video camera 2 shows a warning braille block, a track, and the like. The position of the warning braille block is determined with respect to the walking direction by the method described above, and the vibration pattern is notified by the vibrator 31 provided on the wristband 3.
[0060]
FIG. 15 shows a vibration pattern of walking guidance at a station platform. The visually impaired goes straight along the braille block. However, since there is a track on the right side in the walking direction, it is notified that the right side is dangerous. Therefore, the vibration pattern is the code 251414 shown in Table 1, and the vibration period is as shown in FIG. Two vibrations in one cycle are the vibrator (1) and the vibrator (2), and one vibration in one cycle is the vibrator (3) and the vibrator (5).
[0061]
FIG. 16 shows the state of the guidance process until a visually impaired person enters the platform from the stairs of the station and gets on the train. FIG. 17 shows a guidance process when a visually impaired person crosses a crosswalk. FIG. 18 shows a guidance process when a visually impaired person visits a department store, a coffee shop, a city hall, or a hospital in daily life.
[0062]
【The invention's effect】
As described above, according to the present invention, it is possible to notify a visually impaired person of walking position information and walking direction without using voice. As a result, it is possible to prevent a visually impaired person from falling down on the platform of a station, a collision inside an intersection, and the like. Further, the present invention has a specific effect of providing a walking assistance device for assisting a visually impaired person at low cost.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration of a walking support device.
FIG. 2 is a diagram showing a configuration of a wristband.
FIG. 3 is a diagram showing a cross section of a wristband.
FIG. 4 is a flowchart of teaching.
FIG. 5 is a flowchart of a guidance process.
FIG. 6 is a flowchart showing an observation pattern of a guidance process.
FIG. 7 is a diagram showing an example of a data structure of route information.
FIG. 8 is a detection flowchart of a braille block.
FIG. 9 is a diagram showing a method of detecting a Braille block.
FIG. 10 is a diagram showing a method of converting a binarized image of a braille block into a vibration pattern.
FIG. 11 is a diagram illustrating a method of detecting a non-slip block and generating a vibration pattern.
FIG. 12 is a flowchart for detecting a pedestrian crossing.
FIG. 13 is a diagram illustrating a method of detecting a pedestrian crossing and generating a vibration pattern.
FIG. 14 is a diagram showing how the walking support device guides a visually impaired person walking on a station platform.
FIG. 15 is a diagram showing a vibration pattern of walking guidance in a station platform.
FIG. 16 is a diagram showing a state of a guidance process until a visually impaired person enters the platform from the stairs of the station and gets on the train.
FIG. 17 is a diagram showing guidance processing when a visually impaired person crosses a crosswalk.
FIG. 18 is a diagram illustrating a guidance process when a visually impaired person visits a department store, a coffee shop, a city hall, and a hospital in daily life.
[Explanation of symbols]
1 walking support device 2 video camera 3 wristband 4 mobile terminal (PDA)
5 Pedometer 6 Sonar system 7 Management center 8 GPS
10 Image processing device 11 CPU
12 memory 13 image processing chip 14 input / output chip 15 start button 31 vibrator 41 speaker 42 keyboard

Claims (5)

A position determining device for determining a position,
An input device for receiving inputs of the start and end of the path and the type of the path; A walking support device comprising: a path information creating unit that creates and acquires a path information file by describing acquired and acquired positions and the type of input path for each path. The input device is configured to receive input of information on a message associated with a start end and an end,
The walking support device according to claim 1, wherein the route creating means describes information about a message input in connection with a start end and / or an end. Path reading means for reading the start and end positions and the type of path described for each path in the path information file;
3. A walking direction guiding means for guiding a walking direction from the read start and end positions and the type of path, the current position obtained from the input device, and the type of path at the current position. The walking support device according to the above. The said walk direction guidance means arrange | positions a vibrator or an electric stimulator in a matrix form, and changes a vibration pattern or an electric stimulus pattern according to the guidance content, and guides by tactile sense. 3. The walking support device according to claim 1. The walking support device according to any one of claims 1 to 4, wherein the route reading unit further includes a unit that reads the message and notifies the read message by voice.

Images