WO2024134803A1 - Door opening/closing system and sensor for door opening/closing system - Google Patents
Door opening/closing system and sensor for door opening/closing systemDownload PDFInfo
- Publication number
- WO2024134803A1 WO2024134803A1PCT/JP2022/047145JP2022047145WWO2024134803A1WO 2024134803 A1WO2024134803 A1WO 2024134803A1JP 2022047145 WJP2022047145 WJP 2022047145WWO 2024134803 A1WO2024134803 A1WO 2024134803A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- door
- sensor
- closing system
- door opening
- event
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
- E—FIXED CONSTRUCTIONS
- E05—LOCKS; KEYS; WINDOW OR DOOR FITTINGS; SAFES
- E05F—DEVICES FOR MOVING WINGS INTO OPEN OR CLOSED POSITION; CHECKS FOR WINGS; WING FITTINGS NOT OTHERWISE PROVIDED FOR, CONCERNED WITH THE FUNCTIONING OF THE WING
- E05F15/00—Power-operated mechanisms for wings
- E05F15/70—Power-operated mechanisms for wings with automatic actuation
- E05F15/73—Power-operated mechanisms for wings with automatic actuation responsive to movement or presence of persons or objects
- E05F15/74—Power-operated mechanisms for wings with automatic actuation responsive to movement or presence of persons or objects using photoelectric cells
- E—FIXED CONSTRUCTIONS
- E06—DOORS, WINDOWS, SHUTTERS, OR ROLLER BLINDS IN GENERAL; LADDERS
- E06B—FIXED OR MOVABLE CLOSURES FOR OPENINGS IN BUILDINGS, VEHICLES, FENCES OR LIKE ENCLOSURES IN GENERAL, e.g. DOORS, WINDOWS, BLINDS, GATES
- E06B9/00—Screening or protective devices for wall or similar openings, with or without operating or securing mechanisms; Closures of similar construction
- E06B9/02—Shutters, movable grilles, or other safety closing devices, e.g. against burglary
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V8/00—Prospecting or detecting by optical means
- G01V8/10—Detecting, e.g. by using light barriers
- G01V8/20—Detecting, e.g. by using light barriers using multiple transmitters or receivers
Definitions
- the present inventionrelates to a door opening/closing system and a sensor for the door opening/closing system.
- Patent Document 1a conventional automatic door system is known in which an image sensor such as a CCD camera or CMOS camera is installed around the automatic door, and the door is configured to automatically open when the sensor detects an object approaching the door.
- an image sensorsuch as a CCD camera or CMOS camera
- the inventors of the present applicationhave developed a system that installs a sensor on an industrial door, for example in a factory, and processes the image data obtained by the sensor, making it possible to identify whether an object approaching the door is, for example, a person or a moving object such as a forklift.
- the present inventionaims to solve all of the above problems at once, and its main objective is to enable accurate detection of objects approaching a door regardless of the surrounding brightness.
- the door opening and closing system of the present inventionis a door opening and closing system that detects an object with a sensor and opens the door, and is characterized in that the sensor is an event-based sensor that detects the luminance change of each pixel and outputs the position of the pixel where the luminance change was detected and the detection time.
- Such a door opening and closing systemuses an event-based sensor that detects changes in brightness, so unlike image sensors that can only distinguish brightness within a dynamic range, there is no whiteout or blackout, and it can accurately detect objects approaching the door.
- an object discrimination unitthat uses the data acquired by the event-based sensor to discriminate whether a detected object is a human or a moving object other than a human.
- the doorcan be opened only when a person approaches, or only when a moving object other than a person approaches, and different specifications can be used depending on the object approaching the door, thereby improving convenience and safety.
- the acquired data output from an image sensorincludes information on all pixels, whereas the event-based sensor can output acquired data including information on only pixels where a luminance change has occurred. Therefore, the amount of data acquired by the event-based sensor is smaller than that of data acquired by an image sensor, and the load on an information processing device required for image analysis, image processing, etc. is reduced.
- the object discrimination unitapplies a machine learning model to the acquired data to discriminate whether a detected object is a human or a non-human moving object.
- a machine-learned modelcan be applied that uses the vast amount of data obtained by the event-based sensor, making it possible to more accurately identify detected objects.
- a no-entry area for humans or non-human moving objectsbe set in the detection area of the event-based sensor, and that an alarm unit be provided that notifies when a human or non-human moving object identified by the object discrimination unit is entering the no-entry area. According to this configuration, since a notification is given when a person or a moving object enters a no-entry area, it is possible to prevent accidents involving contact between people and moving objects.
- An example of an embodiment in which the effect of preventing collision accidents is particularly pronouncedis when the door is a shutter and the moving body is an industrial vehicle.
- a position prediction unitthat predicts a future position, which is a position after the time the object is detected, based on the position and moving speed of the object detected by the event-based sensor, and an alarm unit that alerts the user that at least two objects will collide or are at risk of colliding based on the future position.
- the senor for a door opening and closing system of the present inventionis a sensor used in a door opening and closing system that detects an object with a sensor and opens a door, and is characterized in that it detects changes in luminance for each pixel and outputs the position of the pixel where the change in luminance was detected and the detection time.
- a sensor for a door opening/closing systemcan achieve the same effects as the door opening/closing system described above.
- the present inventionconfigured in this way, can accurately detect an object approaching the door regardless of the surrounding brightness.
- FIG. 2is a schematic diagram illustrating an example use of an event-based sensor in accordance with an embodiment of the present invention.
- FIG. 2is a functional block diagram showing functions of the information processing apparatus according to the embodiment.
- FIG. 2is a schematic diagram illustrating information output from the event-based sensor of the embodiment.
- FIG. 4is a schematic diagram illustrating a no-entry area in the embodiment.
- the door opening/closing system 100is used in industrial facilities such as factories, for example, and is for opening and closing an industrial door D (hereinafter, simply referred to as the door D) provided in the facilities.
- the industrial facilityis a facility in which not only people H such as workers walk, but also moving objects V other than people, such as forklifts, run, and the industrial door D can be a shutter or the like.
- the door opening and closing system 100includes a sensor 10 for detecting an object, and an information processing device 20 connected to the sensor 10 by wire or wirelessly.
- the sensor 10 and the information processing device 20may be integrated into a module.
- the sensors 10are arranged, for example, above the door D, and here, the sensors 10 are provided on both the indoor and outdoor sides of the door D.
- This sensor 10captures an image of a detection area set around the door D, and is specifically an event-based sensor 10 that detects an object by detecting changes in the brightness of each of multiple pixels.
- Event-based sensor 10is configured to detect luminance changes for each pixel and output data (hereinafter referred to as acquired data) including the position of the pixel where the luminance change was detected and the time when the luminance change was detected.
- acquired datadata including the position of the pixel where the luminance change was detected and the time when the luminance change was detected.
- event-based sensor 10is capable of detecting moving objects separately from stationary objects; in other words, it is configured to detect moving objects without detecting stationary objects.
- the event-based sensor 10converts the luminance of light incident on a pixel into a voltage, detects the difference between that voltage and a reference voltage, and detects this as a change in luminance if the difference exceeds a predetermined threshold.
- the reference voltagemay be set so that each time a change in luminance is detected, the actual voltage at that time (the voltage to be compared with the reference voltage) is set as a new reference voltage.
- the event-based sensor 10By using the event-based sensor 10 described above, when a person H passes through the detection area, as shown in FIG. 3, the event-based sensor 10 generates acquired data using only the information obtained from the pixels in which a luminance change has occurred, and outputs the acquired data, for example, in real time.
- the information processing device 20is physically a general-purpose or dedicated computer equipped with a CPU, memory, etc., and as shown in FIG. 2, the CPU and its peripheral devices work together in accordance with the program stored in the memory to perform the functions of at least an object detection unit 21, an object discrimination unit 22, and a door control unit 23.
- the object detection unit 21receives acquired data output from the event-based sensor 10 and detects objects based on the acquired data.
- the object detection unit 21detects the presence of an object in the detection area of the event-based sensor 10, for example, when the number of pixels in which a luminance change has occurred, in other words, the number of pixels contained in the received acquired data, exceeds a predetermined number of pixels.
- the object discrimination unit 22uses the acquired data obtained by the event-based sensor 10 to determine the type of object detected by the object detection unit 21, and specifically, determines whether the detected object is a person H or a non-human moving object V.
- non-human moving objects Vinclude industrial vehicles such as forklifts, turret trucks, and unmanned guided vehicles.
- the object discrimination unit 22When the object discrimination unit 22 receives a detection signal from the object detection unit 21 indicating that an object has been detected, it discriminates the object based on the acquired data corresponding to the detected object, and determines whether the object is a person H or a non-human moving body V based on, for example, the size, color, and moving speed of the detected object.
- the object discrimination unit 22 of this embodimentis configured to use a machine learning model to determine whether a detected object is a person H or a moving object V other than a human.
- This machine-learned modelis generated by machine learning using data acquired by the event-based sensor 10, and may be created outside the information processing device 20 of this embodiment, or the information processing device 20 may have a function as a model creation unit (not shown) that creates this model.
- a specific example of this machine-learned modelis one that has been machine-learned using the occurrence status of events contained in the acquired data, i.e., the occurrence status that indicates how many events (brightness changes) have occurred at which pixels over a certain period of time.
- the door control unit 23controls the operation of the door D based on the discrimination result of the object discrimination unit 22, and based on the discrimination result, outputs a control signal to the drive mechanism that drives the door D, switching the door D between an open state and a closed state.
- a specific embodiment of the door control unit 23is, for example, to switch the closed door D to an open state when the detected object is determined to be a person H, and to keep the door D in the closed state when the detected object is determined to be a moving object V other than a person.
- Another embodiment of the door control unit 23may be the opposite to the above-described operation, in that if the detected object is determined to be a moving object V other than a human, the door D may be switched from a closed state to an open state, and if the detected object is determined to be a person H, the door D may be maintained in a closed state.
- the embodiment of the door control unit 23is not limited to the one described above, and the state of the door D may be switched at various times based on the discrimination result by the object discrimination unit 22.
- the timing at which the door control unit 23 switches the door D from an open state to a closed statemay be set to various timings, for example, using the elapsed time after switching to the open state or the detection results from the object detection unit 21.
- no-entry area X for people H(hereinafter referred to as no-entry area X for people) and a no-entry area Y for moving bodies V other than people (hereinafter referred to as no-entry area Y for vehicles) are set.
- the information processing device 20 of this embodimentis further configured to function as a notification unit 24 and a position prediction unit 25, as shown in FIG. 2.
- the notification unit 24acquires the data obtained by the event-based sensor 10 and the discrimination result of the object discrimination unit 22 described above, and notifies the user when a person H enters the no-entry area X for people, and notifies the user when an object other than a person enters the no-entry area Y for vehicles.
- the notification destination by the notification unit 24may be the operator of the person H or non-human moving body V that has entered the no-entry area, or a manager inside or outside the facility.
- the position prediction unit 25predicts the future position of an object detected by the event-based sensor 10, based on the object's position and moving speed.
- the position prediction unit 25processes the data acquired by the event-based sensor 10 to calculate the moving speed and direction of the object based on the change in the object's position over time contained in the acquired data, and predicts the object's future position after a predetermined time has passed based on the moving speed and direction and the current position.
- this future positionis output to the above-mentioned notification unit 24, which notifies the user that two or more objects will collide or are at risk of colliding, based on the future positions of the objects predicted by the position prediction unit 25.
- the notification unit 24here obtains the future position of each object detected by the object detection unit 21, and if the future positions of multiple objects at the same time overlap or approach within a specified distance range, it issues a notification that the objects will collide or are at risk of colliding.
- the door opening and closing system 100configured in this manner uses an event-based sensor 10 that detects changes in the brightness of an object, and therefore, unlike an image sensor that can only distinguish brightness within a dynamic range, there is no whiteout or blackout, and it can accurately detect objects approaching the door D.
- the door control unit 23opens and closes the door D based on the results of the discrimination by the object discrimination unit 22, and the specifications can be changed depending on the approaching object, such as opening the door D only when a person H approaches, or opening the door D only when a moving object V other than a person approaches, thereby improving convenience and safety.
- the event-based sensor 10can output acquired data including information only about pixels in which a change in brightness has occurred, and the amount of data output is small, which reduces the load on the information processing device (CPU) required for image analysis, image processing, etc. Therefore, the object discrimination unit 22 uses a machine-learned model that uses the acquired data obtained by the event-based sensor 10 as training data to discriminate whether the object is a person H or a non-human moving object V, making it possible to more accurately discriminate the detected object while taking advantage of the advantages of the event-based sensor 10 described above.
- the notification unit 24notifies the user when a person H enters a no-entry area X for people, and also notifies the user when an object other than a person enters a no-entry area Y for vehicles, thereby preventing accidental contact between the person H and the moving object V.
- the door opening/closing system 100 of this embodimentis used at sites where the door is a shutter and the moving object is an industrial vehicle, making it particularly useful at sites where collision accidents are frequent.
- the notification unit 24uses the future positions of the objects predicted by the position prediction unit 25 to notify that two or more objects will collide or are at risk of colliding, making it possible to notify in advance that there is a risk of objects coming into contact with each other.
- the event-based sensor 10does not detect stationary objects, there is no alert that a stationary object is a danger. This makes it possible to eliminate unnecessary signals and reduce false alerts when alerting the risk of a collision between moving objects such as a person H and a moving body V.
- event-based sensor 10outputs position information and time information of a pixel where a luminance change is detected, but it may also output information regarding the amount of change in luminance of incident light.
- the shades of white and black contained in the acquired datacan be changed according to the amount of change in luminance, making it possible to create acquired data with higher resolution.
- the acquired data obtained by the event-based sensor 10does not have to be created based on only one luminance change that occurs at a certain time, but may be created based on multiple luminance changes that occur within a specified period of time.
- the acquired datamay be created by coloring pixels that have experienced a luminance change within, for example, 30 seconds in white and pixels that have not experienced a luminance change in black.
- the door control unit 23 in the above embodimentcontrols the state of the door D based on the discrimination result of the object discrimination unit 22, but the door control unit 23 may control the state of the door D based on the detection result of the object detection unit 21. That is, when the object detection unit 21 detects an object, the door control unit 23 may switch the door D from a closed state to an open state. In this embodiment, the information processing device 20 does not need to have the function of the object discrimination unit 22 .
- the machine-learned modeldoes not necessarily have to be created using the acquired data.
- image data of an image of the detection areamay be generated based on the acquired data, and the machine-learned model may be created using this image data.
- One example of a method for generating image datais to output the coordinates of pixels that have experienced a luminance change at a certain time and the coordinates of pixels that have not experienced a luminance change at the same time in a manner that makes them identifiable, for example by binarization.
- the image datamay be output by further combining information on the polarity of the luminance change, i.e., whether the luminance has changed in the direction of increasing or decreasing. Specifically, according to the polarity of the luminance change, pixels whose luminance has changed in the direction of increasing may be displayed in white, and pixels whose luminance has changed in the direction of decreasing may be displayed in black, the same as pixels whose luminance has not changed. With this configuration, the contours of the object are clearer and image data with higher resolution can be created.
- the object discrimination unit 22does not necessarily need to use a machine-learned model, and may use various types of image analysis to discriminate whether a detected object is a person H or a moving object V other than a human.
- the functions of the object detection unit 21, object discrimination unit 22, control unit, notification unit 24, and position prediction unit 25are provided in the information processing device 20, but some or all of these functions may be provided in the event-based sensor 10 or another computer, etc.
- the event-based sensor 10creates the acquired data, but the information processing device 20 or another computer may create the acquired data.
- the present inventionmakes it possible to accurately detect an object approaching a door regardless of the surrounding brightness.
- Door opening and closing system DDoor H: Person V: Moving object 10: Sensor (event-based sensor) 20: Information processing device 21: Object detection unit 22: Object discrimination unit 23: Door control unit 24: Notification unit 25: Position prediction unit X: Person entry prohibited area Y: Vehicle entry prohibited area
Landscapes
- Engineering & Computer Science (AREA)
- Structural Engineering (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Geophysics (AREA)
- Architecture (AREA)
- Civil Engineering (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
Description
Translated fromJapanese本発明は、扉開閉システム及び扉開閉システム用センサに関するものである。The present invention relates to a door opening/closing system and a sensor for the door opening/closing system.
従来の自動扉システムとしては、特許文献１に示すように、自動扉周辺にＣＣＤカメラやＣＭＯＳカメラ等のイメージセンサが設けられており、このセンサが、扉に接近する物体を検知した場合に、扉を自動で開くように構成されたものが知られている。As shown in
このような中で、本願発明者は、例えば工場などの産業用扉にセンサを設けるとともに、そのセンサにより得られる画像データを画像処理することで、扉に接近する物体が例えば人であるかフォークリフト等の移動体であるかを識別可能にするシステムを開発している。In light of this, the inventors of the present application have developed a system that installs a sensor on an industrial door, for example in a factory, and processes the image data obtained by the sensor, making it possible to identify whether an object approaching the door is, for example, a person or a moving object such as a forklift.
しかしながら、これまでのセンサは、上述した通りイメージセンサを用いているため、日差しや照明が強くダイナミックレンジを越えて明るい場合には、画像データに白飛びが生じてしまい、扉に接近する物体を正確に検出できなくなるし、逆にダイナミックレンジを越えて暗い場合には、画像データに黒つぶれが生じてしまい、やはり扉に接近する物体を正確に検出できなくなる。However, as mentioned above, previous sensors used image sensors, so when sunlight or lighting is strong and the image is brighter than the dynamic range, the image data will be overexposed, making it impossible to accurately detect an object approaching the door. Conversely, when the image is dark and the dynamic range is exceeded, the image data will be underexposed, making it impossible to accurately detect an object approaching the door.
そこで、本発明は上述した問題を一挙に解決するものであり、扉に接近する物体を周囲の明るさによらず精度良く検出できるようにすることをその主たる課題とするものである。The present invention aims to solve all of the above problems at once, and its main objective is to enable accurate detection of objects approaching a door regardless of the surrounding brightness.
すなわち本発明に係る扉開閉システムは、物体をセンサで検知して扉を開く扉開閉システムであって、前記センサが、画素毎の輝度変化を検出するとともに、輝度変化が検出された画素の位置及び検出時刻を出力するイベントベースセンサであることを特徴とするものである。In other words, the door opening and closing system of the present invention is a door opening and closing system that detects an object with a sensor and opens the door, and is characterized in that the sensor is an event-based sensor that detects the luminance change of each pixel and outputs the position of the pixel where the luminance change was detected and the detection time.
このような扉開閉システムであれば、輝度の変化を検出するイベントベースセンサを用いているので、ダイナミックレンジ内の輝度しか識別できないイメージセンサとは異なり、白飛び又は黒つぶれすることがなく、扉に接近する物体を精度よく検出することができる。Such a door opening and closing system uses an event-based sensor that detects changes in brightness, so unlike image sensors that can only distinguish brightness within a dynamic range, there is no whiteout or blackout, and it can accurately detect objects approaching the door.
前記イベントベースセンサにより得られた取得データを用いて、検知された物体が人であるか、人以外の移動体であるかを判別する物体判別部を備えることがより望ましい。
このような構成によれば、例えば人が接近した場合にのみ扉を開いたり、人以外の移動体が接近した場合にのみ扉を開いたり、扉に接近する物に応じて仕様を使い分けることができ、利便性及び安全性の向上を図れる。It is more preferable to include an object discrimination unit that uses the data acquired by the event-based sensor to discriminate whether a detected object is a human or a moving object other than a human.
With such a configuration, for example, the door can be opened only when a person approaches, or only when a moving object other than a person approaches, and different specifications can be used depending on the object approaching the door, thereby improving convenience and safety.
イメージセンサから出力される取得データには全画素の情報が含まれている一方、イベントベースセンサからは輝度変化が生じた画素のみの情報を含む取得データを出力させることができる。
従って、イベントベースセンサにより得られる取得データは、イメージセンサにより得られる取得データに比べると、データ量が少なく、例えば画像解析や画像処理などに必要な情報処理装置の負荷が少なくて済む。The acquired data output from an image sensor includes information on all pixels, whereas the event-based sensor can output acquired data including information on only pixels where a luminance change has occurred.
Therefore, the amount of data acquired by the event-based sensor is smaller than that of data acquired by an image sensor, and the load on an information processing device required for image analysis, image processing, etc. is reduced.
こうしたイベントベースセンサの利点を活かすべく、前記物体判別部が、前記取得データに機械学習済みモデルを適用して、検知された物体が人であるか、人以外の移動体であるかを判別することが好ましい。
このような構成であれば、機械学習済みモデルとして、イベントベースセンサにより得られる膨大な取得データを用いたものを適用することができ、検知された物体をより精度良く判別することが可能となる。In order to take advantage of the advantages of such event-based sensors, it is preferable that the object discrimination unit applies a machine learning model to the acquired data to discriminate whether a detected object is a human or a non-human moving object.
With this configuration, a machine-learned model can be applied that uses the vast amount of data obtained by the event-based sensor, making it possible to more accurately identify detected objects.
前記イベントベースセンサの検知領域に人又は人以外の移動体の進入禁止エリアが設定されており、前記物体判別部により判別された人又は人以外の移動体が前記進入禁止エリアに進入していることを報知する報知部を備えることがより望ましい。
このような構成によれば、人又は移動体が進入禁止エリアに進入したことが報知されるので、人と移動体との接触事故を未然に防ぐことができる。It is more desirable that a no-entry area for humans or non-human moving objects be set in the detection area of the event-based sensor, and that an alarm unit be provided that notifies when a human or non-human moving object identified by the object discrimination unit is entering the no-entry area.
According to this configuration, since a notification is given when a person or a moving object enters a no-entry area, it is possible to prevent accidents involving contact between people and moving objects.
接触事故を防ぐといった作用効果が顕著に発揮される実施態様としては、前記扉がシャッターであり、前記移動体が産業用車両である態様を挙げることができる。An example of an embodiment in which the effect of preventing collision accidents is particularly pronounced is when the door is a shutter and the moving body is an industrial vehicle.
接触事故を未然に防ぐための別の態様としては、前記イベントベースセンサにより検知された物体の位置及び移動速度から、当該物体の検知された時刻よりも後の位置である将来位置を予測する位置予測部と、前記将来位置に基づいて、少なくとも２つの物体が衝突する又は衝突する恐れのあることを報知する報知部とを備えることが好ましい。
このような構成によれば、物体同士が接触する恐れのあることを未然に知らせることができる。As another aspect for preventing collision accidents, it is preferable to provide a position prediction unit that predicts a future position, which is a position after the time the object is detected, based on the position and moving speed of the object detected by the event-based sensor, and an alarm unit that alerts the user that at least two objects will collide or are at risk of colliding based on the future position.
With this configuration, it is possible to notify in advance that there is a risk of objects coming into contact with each other.
また、本発明に係る扉開閉システム用センサは、物体をセンサで検知して扉を開く扉開閉システムに用いられるセンサであって、画素毎の輝度変化を検出するとともに、輝度変化が検出された画素の位置及び検出時刻を出力することを特徴とするものである。
このような扉開閉システム用センサであれば、上述した扉開閉システムと同様の作用効果を発揮させることができる。In addition, the sensor for a door opening and closing system of the present invention is a sensor used in a door opening and closing system that detects an object with a sensor and opens a door, and is characterized in that it detects changes in luminance for each pixel and outputs the position of the pixel where the change in luminance was detected and the detection time.
Such a sensor for a door opening/closing system can achieve the same effects as the door opening/closing system described above.
このように構成した本発明によれば、扉に接近する物体を周囲の明るさによらず精度良く検出することができる。The present invention, configured in this way, can accurately detect an object approaching the door regardless of the surrounding brightness.
以下、本発明に係る扉開閉システムの一実施形態について、図面を参照しながら説明する。Below, one embodiment of the door opening and closing system according to the present invention will be described with reference to the drawings.
＜装置構成＞
本実施形態に係る扉開閉システム１００は、図１に示すように、例えば工場などの産業用設備に用いられるものであり、その設備に設けられている産業用扉Ｄ（以下、単に扉Ｄともいう）を開閉するためのものである。なお、産業用設備は、作業者などの人Ｈが歩行しているのみならず、フォークリフトなどの人以外の移動体Ｖも走行している設備であり、産業用扉Ｄとしてはシャッターなどを挙げることができる。<Device Configuration>
1, the door opening/
この扉開閉システム１００は、図２に示すように、物体を検知するためのセンサ１０と、このセンサ１０に有線又は無線で接続された情報処理装置２０とを備える。なお、センサ１０と情報処理装置２０とは、一体的にモジュール化されていても良い。As shown in FIG. 2, the door opening and
センサ１０は、図１に示すように、例えば扉Ｄの上方に配置されており、ここでは扉Ｄの屋内側と屋外側とのそれぞれに設けられている。As shown in FIG. 1, the
このセンサ１０は、扉Ｄの周囲に設定された検知領域を撮影するものであり、具体的には、複数の画素それぞれの輝度変化を検出することで物体を検知するイベントベースセンサ１０である。This
イベントベースセンサ１０は、画素毎の輝度変化を検出するとともに、輝度変化が検出された画素の位置と、その輝度変化が検出された時刻とを含むデータ（以下、取得データという）を出力するように構成されたものである。すなわち、イベントベースセンサ１０は、動いているものを静止しているものから切り分けて検出可能なものであり、言い換えれば、静止物体を検知することなく、移動物体を検知できるように構成されたものである。Event-based
より具体的に説明すると、イベントベースセンサ１０は、画素に入射する光の輝度の大きさを電圧に変換し、その電圧と基準電圧との差分を検出して、その差分が所定の閾値を越えた場合に、そのことを輝度変化として検出する。なお、基準電圧については、輝度変化を検出する度に、その際の実電圧（基準電圧の比較対象となる電圧）が新たな基準電圧として設定されるようにしても良い。More specifically, the event-based
上述したイベントベースセンサ１０を用いることにより、検知領域を人Ｈが通過すると、図３に示すように、イベントベースセンサ１０からは、輝度変化が生じた画素により得られる情報のみを用いて取得データが生成されて、その取得データが例えばリアルタイムで出力される。By using the event-based
続いて、情報処理装置２０について説明する。Next, we will explain the
情報処理装置２０は、物理的には、ＣＰＵ、メモリなどを備えた汎用乃至専用のコンピュータであり、図２に示すように、前記メモリに格納されているプログラムに従ってＣＰＵやその周辺機器が協働することにより、少なくとも物体検知部２１、物体判別部２２、及び扉制御部２３としての機能を発揮するものである。The
物体検知部２１は、イベントベースセンサ１０から出力される取得データを受け付けて、その取得データに基づいて物体を検知するものである。The
具体的に物体検知部２１は、例えば、輝度変化が生じた画素数が、言い換えれば、受け付けた取得データに含まれる画素数が、所定の画素数を超えた場合に、イベントベースセンサ１０の検知領域に物体が存在していることを検知する。Specifically, the
物体判別部２２は、イベントベースセンサ１０により得られた取得データを用いて、物体検知部２１により検知された物体の種別を判定するものであり、具体的には、検知された物体が、人Ｈであるか、人以外の移動体Ｖであるかを判別する。なお、人以外の移動体Ｖとしては、フォークリフト、ターレットトラック、無人搬送車などの産業用車両を挙げることができる。The
この物体判別部２２は、物体検知部２１から物体を検知したことを示す検知信号を受け付けると、その検知された物体に対応する取得データに基づいて、その物体を判別するものであり、例えば検知された物体の大きさ、色、移動速度などから、その物体が人Ｈであるか、人以外の移動体Ｖであるかを判別する。When the
本実施形態の物体判別部２２は、機械学習済みモデルを用いて、検知された物体が人Ｈであるか、人以外の移動体Ｖであるかを判断するように構成されている。The
この機械学習済みモデルは、イベントベースセンサ１０により取得された取得データを用いて機械学習させてなるものであり、本実施形態の情報処理装置２０の外部で作成されたものであっても良いし、情報処理装置２０が、このモデルを作成するモデル作成部（不図示）としての機能を備えていても良い。This machine-learned model is generated by machine learning using data acquired by the event-based
この機械学習済みモデルの具体的な一例としては、取得データに含まれるイベントの発生状況、すなわち、一定時間の間にどこの画素にどれだけイベント（輝度変化）が発生しているかを示す発生状況を用いて機械学習させたものを挙げることができる。A specific example of this machine-learned model is one that has been machine-learned using the occurrence status of events contained in the acquired data, i.e., the occurrence status that indicates how many events (brightness changes) have occurred at which pixels over a certain period of time.
扉制御部２３は、物体判別部２２の判別結果に基づいて扉Ｄの動作を制御するものであり、その判別結果に基づいて、扉Ｄを駆動させる駆動機構に制御信号を出力して、その扉Ｄを開状態又は閉状態に切り替えるものである。The
扉制御部２３の具体的な実施態様としては、例えば検知された物体が人Ｈであると判別された場合は、閉状態にある扉Ｄを開状態に切り替え、検知された物体が人以外の移動体Ｖであると判別された場合は、扉Ｄを閉状態のまま維持する態様を挙げることができる。A specific embodiment of the
また、扉制御部２３の別の実施態様としては、上述した動作とは逆に、検知された物体が人以外の移動体Ｖであると判別された場合は、閉状態にある扉Ｄを開状態に切り替え、検知された物体が人Ｈであると判別された場合は、扉Ｄを閉状態のまま維持する態様であっても良い。Another embodiment of the
なお、扉制御部２３の実施態様は、上述したものに限られるものではなく、物体判別部２２による判別結果に基づいて種々のタイミングで扉Ｄの状態を切り替えるようにして構わない。The embodiment of the
また、扉制御部２３が開状態にある扉Ｄを閉状態に切り替えるタイミングとしては、例えば開状態に切り替わってからの経過時間や、物体検知部２１からの検知結果などを用いて、種々のタイミングに設定して構わない。The timing at which the
ところで、産業用設備の屋内及び屋外の一方又は両方において、人Ｈと人以外の移動物体の通路が分かれている場合がある。Incidentally, in either or both of the indoor and outdoor areas of industrial facilities, there are cases where the passageways for people H and non-human moving objects are separate.
このような場合を想定して、本実施形態におけるイベントベースセンサ１０の検知領域には、図４に示すように、例えば、人Ｈの進入禁止エリアＸ（以下、人進入禁止エリアＸという）及び人以外の移動体Ｖの進入禁止エリアＹ（以下、車両進入禁止エリアＹという）が設定されている。In anticipation of such a case, in the detection area of the event-based
そして、本実施形態の情報処理装置２０は、図２に示すように、報知部２４及び位置予測部２５としての機能をさらに発揮するように構成されている。The
報知部２４は、イベントベースセンサ１０により得られる取得データと、上述した物体判別部２２の判別結果とを取得するとともに、人進入禁止エリアＸに人Ｈが進入している場合に、そのことを報知するとともに、車両進入禁止エリアＹに人以外の物体が進入している場合に、そのことを報知するものである。The
なお、報知部２４による報知先は、進入禁止エリアに進入した人Ｈ又は人以外の移動体Ｖの操縦者であっても良いし、設備内又は設備外の管理者などであっても良い。The notification destination by the
位置予測部２５は、イベントベースセンサ１０により検知された物体の位置及び移動速度から、当該物体の検知された時刻よりも後の位置である将来位置を予測するものである。The
具体的にこの位置予測部２５は、イベントベースセンサ１０により得られる取得データをデータ処理することにより、その取得データに含まれる物体の位置の時間変化に基づいて、その物体の移動速度及び移動方向を算出し、それらの移動速度及び移動方向と現在の位置とから、所定時間経過後の将来位置を予測する。Specifically, the
そして、この将来位置は、上述した報知部２４に出力されて、報知部２４は、位置予測部２５が予測した物体の将来位置に基づいて、２つ又はそれ以上の物体が衝突する又は衝突する恐れのあることを報知する。Then, this future position is output to the above-mentioned
具体的にここでの報知部２４は、物体検知部２１により検知された物体それぞれの将来位置を取得して、同時刻における複数の物体の将来位置が重なる、或いは、所定距離範囲内に近づく場合に、それらの物体が衝突する又は衝突する恐れのあることを報知する。Specifically, the
＜本実施形態の効果＞
このように構成された扉開閉システム１００によれば、物体の輝度の変化を検出するイベントベースセンサ１０を用いているので、ダイナミックレンジ内の輝度しか識別できないイメージセンサとは異なり、白飛び又は黒つぶれすることがなく、扉Ｄに接近する物体を精度よく検出することができる。<Effects of this embodiment>
The door opening and
また、物体判別部２２による判別結果に基づいて扉制御部２３が扉Ｄを開閉しており、人Ｈが接近した場合のみ扉Ｄを開いたり、あるいは人以外の移動体Ｖが接近した時のみ扉Ｄを開いたりと、接近する物体に応じて仕様を変えることができるため、利便性及び安全性の向上を図れる。In addition, the
イベントベースセンサ１０からは輝度変化が生じた画素のみの情報を含む取得データを出力させることができ、出力されるデータ量が少なく、例えば画像解析や画像処理などに必要な情報処理装置（ＣＰＵ）の負荷が少なくて済む。
そこで、物体判別部２２が、イベントベースセンサ１０により得られる取得データを教師データとした機械学習済みモデルを用いて、物体が人Ｈであるか人以外の移動体Ｖであるかを判別するので、上述したイベントベースセンサ１０の利点を活かしながら、検知された物体をより精度良く判別することが可能となる。The event-based
Therefore, the
報知部２４が、人進入禁止エリアＸに人Ｈが進入している場合に、そのことを報知するとともに、車両進入禁止エリアＹに人以外の物体が進入している場合に、そのことを報知するので、人Ｈと移動体Ｖとの接触事故を未然に防ぐことができる。The
また、本実施形態の扉開閉システム１００は、扉がシャッターであり、移動体が産業用車両である現場に用いられているので、特に接触事故の多い現場に資するものとなる。In addition, the door opening/
また、報知部２４が、位置予測部２５が予測した物体の将来位置を用いて、２つ又はそれ以上の物体が衝突する又は衝突する恐れのあることを報知するので、物体同士が接触する恐れのあることを未然に知らせることができる。In addition, the
そのうえ、イベントベースセンサ１０は、静止物体を検知しないので、静止物体を危険対象とする報知がなくなる。これにより、人Ｈや移動体Ｖなどの移動物体同士が衝突する恐れを報知する上で、余分な信号を排除して、誤報知を軽減することができる。Furthermore, because the event-based
＜その他の実施形態＞
なお、本発明は前記実施形態に限られるものではない。<Other embodiments>
The present invention is not limited to the above-described embodiment.
例えば、前記実施形態では、イベントベースセンサ１０は、輝度変化を検出した画素の位置情報と時間情報とを出力していたが、さらに入射光の輝度の変化量に関する情報を出力しても良い。
このような構成によれば、例えば、取得データに含まれる白色や黒色の濃淡を輝度の変化量に応じて変化させることができ、より解像度の高い取得データを作成することができる。For example, in the above embodiment, event-based
According to such a configuration, for example, the shades of white and black contained in the acquired data can be changed according to the amount of change in luminance, making it possible to create acquired data with higher resolution.
また、イベントベースセンサ１０により得られる取得データとしては、ある時刻に生じた１回の輝度変化のみに基づいて作成される必要はなく、所定時間内に生じた複数回の輝度変化に基づいて作成されても良い。具体的には、例えば３０秒間に輝度変化が生じた画素を白色、輝度変化が生じなかった画素を黒色にすることで取得データを作成しても良い。Furthermore, the acquired data obtained by the event-based
さらに、前記実施形態の扉制御部２３は、物体判別部２２の判別結果に基づいて扉Ｄの状態を制御していたが、物体検知部２１の検知結果に基づいて扉Ｄの状態を制御しても良い。すなわち、物体検知部２１が物体を検知した場合に、扉制御部２３が扉Ｄを閉状態から開状態に切り替えても良い。
この実施態様においては、情報処理装置２０としては物体判別部２２としての機能を備えていなくても良い。Furthermore, the
In this embodiment, the
さらに加えて、機械学習済みモデルとしては、必ずしも取得データを用いて作成されたものに限らず、例えば、取得データに基づいて検知領域を撮像した画像データを生成し、この画像データを用いて機械学習済みモデルを作成しても良い。In addition, the machine-learned model does not necessarily have to be created using the acquired data. For example, image data of an image of the detection area may be generated based on the acquired data, and the machine-learned model may be created using this image data.
画像データの生成方法の一例としては、ある時刻に輝度変化が生じた画素の座標と、同時刻に輝度変化が生じなかった画素の座標とを二値化等により識別可能に出力する方法を挙げることができる。One example of a method for generating image data is to output the coordinates of pixels that have experienced a luminance change at a certain time and the coordinates of pixels that have not experienced a luminance change at the same time in a manner that makes them identifiable, for example by binarization.
また、画像データの別の生成方法としては、輝度変化の極性、すなわち輝度が大きくなる方向に変化したかあるいは小さくなる方向に変化したかに関する情報をさらに組み合わせて、画像データを出力しても良い。具体的には、輝度変化の極性に応じて、輝度が大きくなる方向に変化した画素を白色に表示するとともに、輝度変化が小さくなる方向に変化した画素を、輝度変化していない画素と同じく黒色に表示するようにしても良い。
このような構成であれば、物体の輪郭がはっきりし、より解像度の高い画像データを作成することができる。As another method of generating image data, the image data may be output by further combining information on the polarity of the luminance change, i.e., whether the luminance has changed in the direction of increasing or decreasing. Specifically, according to the polarity of the luminance change, pixels whose luminance has changed in the direction of increasing may be displayed in white, and pixels whose luminance has changed in the direction of decreasing may be displayed in black, the same as pixels whose luminance has not changed.
With this configuration, the contours of the object are clearer and image data with higher resolution can be created.
さらに、物体判別部２２としては、必ずしも機械学習済みモデルを用いる必要はなく、種々の画像解析などを用いて検知された物体が人Ｈであるか、人以外の移動体Ｖであるかを判別しても良い。Furthermore, the
また、物体検知部２１、物体判別部２２、制御部、報知部２４、及び位置予測部２５としての機能は、前記実施形態では情報処理装置２０に備えさせていたが、これらの機能の一部又は全部をイベントベースセンサ１０や別のコンピュータなどに備えさせても良い。In addition, in the above embodiment, the functions of the
加えて、前記実施形態ではイベントベースセンサ１０が取得データを作成していたが、情報処理装置２０や別のコンピュータが取得データを作成しても良い。In addition, in the above embodiment, the event-based
その他、本発明は前記実施形態に限られず、その趣旨を逸脱しない範囲で種々の変形が可能であるのは言うまでもない。It goes without saying that the present invention is not limited to the above-described embodiment, and various modifications are possible without departing from the spirit of the invention.
本発明によれば、扉に接近する物体を周囲の明るさによらず精度良く検出することができる。The present invention makes it possible to accurately detect an object approaching a door regardless of the surrounding brightness.
１００・・・扉開閉システム
Ｄ ・・・扉
Ｈ ・・・人
Ｖ ・・・移動体
１０ ・・・センサ（イベントベースセンサ）
２０ ・・・情報処理装置
２１ ・・・物体検知部
２２ ・・・物体判別部
２３ ・・・扉制御部
２４ ・・・報知部
２５ ・・・位置予測部
Ｘ ・・・人進入禁止エリア
Ｙ ・・・車両進入禁止エリア
100: Door opening and closing system D: Door H: Person V: Moving object 10: Sensor (event-based sensor)
20: Information processing device 21: Object detection unit 22: Object discrimination unit 23: Door control unit 24: Notification unit 25: Position prediction unit X: Person entry prohibited area Y: Vehicle entry prohibited area
Claims (7)
Translated fromJapanese前記センサが、画素毎の輝度変化を検出するとともに、輝度変化が検出された画素の位置及び検出時刻を出力するイベントベースセンサであることを特徴とする開閉システム。A door opening and closing system that detects an object with a sensor and opens a door,
An opening and closing system, wherein the sensor is an event-based sensor that detects a luminance change for each pixel and outputs the position of the pixel where the luminance change was detected and the detection time.
前記物体判別部により判別された人又は人以外の移動体が前記進入禁止エリアに進入していることを報知する報知部を備えることを特徴とする請求項２記載の扉開閉システム。a no-entry area for humans or moving objects other than humans is set in the detection area of the event-based sensor,
3. The door opening and closing system according to claim 2, further comprising an alarm unit that notifies the user that a human or a moving object other than a human identified by the object identification unit is entering the no-entry area.
前記移動体が産業用車両であることを特徴とする請求項２記載の扉開閉システム。The door is a shutter,
3. The door opening and closing system according to claim 2, wherein the moving body is an industrial vehicle.
前記将来位置に基づいて、少なくとも２つの物体が衝突する又は衝突する恐れのあることを報知する報知部とを備えることを特徴とする請求項１記載の扉開閉システム。a position prediction unit that predicts a future position of an object detected by the event-based sensor based on the position and moving speed of the object;
2. The door opening and closing system according to claim 1, further comprising an alarm unit that notifies the user that at least two objects will collide or may collide with each other based on the future positions.
画素毎の輝度変化を検出するとともに、輝度変化が検出された画素の位置及び検出時刻を出力することを特徴とする扉開閉システム用センサ。
A sensor used in a door opening and closing system that detects an object and opens a door,
A sensor for a door opening/closing system, which detects a change in luminance for each pixel, and outputs the position of the pixel where the change in luminance was detected and the detection time.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/JP2022/047145WO2024134803A1 (en) | 2022-12-21 | 2022-12-21 | Door opening/closing system and sensor for door opening/closing system |
| JP2024565476AJPWO2024134803A1 (en) | 2022-12-21 | 2022-12-21 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/JP2022/047145WO2024134803A1 (en) | 2022-12-21 | 2022-12-21 | Door opening/closing system and sensor for door opening/closing system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2024134803A1true WO2024134803A1 (en) | 2024-06-27 |
Family
ID=91588189
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2022/047145PendingWO2024134803A1 (en) | 2022-12-21 | 2022-12-21 | Door opening/closing system and sensor for door opening/closing system |
Country Status (2)
| Country | Link |
|---|---|
| JP (1) | JPWO2024134803A1 (en) |
| WO (1) | WO2024134803A1 (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09245271A (en)* | 1996-03-06 | 1997-09-19 | Atsumi Electron Corp Ltd | Object detection system |
| JP2005129003A (en)* | 2003-06-09 | 2005-05-19 | Hitachi Kokusai Electric Inc | Change detection device |
| JP2011034458A (en)* | 2009-08-04 | 2011-02-17 | Sumitomo Electric Ind Ltd | Monitoring system, monitoring device, and monitoring method |
| JP2017048575A (en)* | 2015-08-31 | 2017-03-09 | 三和シヤッター工業株式会社 | Automatic door opening and closing control system |
| WO2020008685A1 (en)* | 2018-07-06 | 2020-01-09 | コニカミノルタ株式会社 | Information notification device, program for information notification device, and information notification system |
| JP2021067704A (en)* | 2019-10-17 | 2021-04-30 | 株式会社デンソーウェーブ | Imaging device |
| JP2021188352A (en)* | 2020-05-29 | 2021-12-13 | ヒースト株式会社 | Door open / close control device |
| JP2022163931A (en)* | 2021-04-15 | 2022-10-27 | ソニーグループ株式会社 | Information processing system, information processing device, and information processing method |
- 2022
- 2022-12-21JPJP2024565476Apatent/JPWO2024134803A1/jaactivePending
- 2022-12-21WOPCT/JP2022/047145patent/WO2024134803A1/enactivePending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09245271A (en)* | 1996-03-06 | 1997-09-19 | Atsumi Electron Corp Ltd | Object detection system |
| JP2005129003A (en)* | 2003-06-09 | 2005-05-19 | Hitachi Kokusai Electric Inc | Change detection device |
| JP2011034458A (en)* | 2009-08-04 | 2011-02-17 | Sumitomo Electric Ind Ltd | Monitoring system, monitoring device, and monitoring method |
| JP2017048575A (en)* | 2015-08-31 | 2017-03-09 | 三和シヤッター工業株式会社 | Automatic door opening and closing control system |
| WO2020008685A1 (en)* | 2018-07-06 | 2020-01-09 | コニカミノルタ株式会社 | Information notification device, program for information notification device, and information notification system |
| JP2021067704A (en)* | 2019-10-17 | 2021-04-30 | 株式会社デンソーウェーブ | Imaging device |
| JP2021188352A (en)* | 2020-05-29 | 2021-12-13 | ヒースト株式会社 | Door open / close control device |
| JP2022163931A (en)* | 2021-04-15 | 2022-10-27 | ソニーグループ株式会社 | Information processing system, information processing device, and information processing method |
Also Published As
| Publication number | Publication date |
|---|---|
| JPWO2024134803A1 (en) | 2024-06-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3192763B1 (en) | Elevator system | |
| EP3867889B1 (en) | Perimeter breach warning system | |
| EP3409629B2 (en) | Image analytics for elevator maintenance | |
| US7768549B2 (en) | Machine safety system with mutual exclusion zone | |
| AU2003209905B2 (en) | Monitoring a lift area by means of a 3D sensor | |
| US10479653B2 (en) | Monitoring system of a passenger conveyor, a passenger conveyor, and a monitoring method thereof | |
| CN113903058A (en) | Intelligent control system based on regional personnel identification | |
| KR20150121055A (en) | Device and method for monitoring moving entity | |
| JP2018123653A (en) | Automatic door system and control method of automatic door system | |
| CN112433261A (en) | Scanning imaging monitoring system and method and electric vehicle elevator entering scanning imaging monitoring system | |
| WO2024134803A1 (en) | Door opening/closing system and sensor for door opening/closing system | |
| JP2009242045A (en) | Door device | |
| JP3812666B2 (en) | Monitoring device | |
| KR102113489B1 (en) | Action Detection System and Method Using Deep Learning with Camera Input Image | |
| JPS63306894A (en) | Safety monitor method | |
| KR20150075508A (en) | Monitoring and controlling system of door having camera module | |
| TWI776741B (en) | Method for controlling in-vehicle camera and in-vehicle device | |
| JP2020144738A (en) | Security device and security method | |
| JP2000298777A (en) | Intrusion detecting device and method therefor | |
| CN110002305B (en) | Elevator door suture line detection method and system based on computer vision technology | |
| KR101858625B1 (en) | Apparatus and method for detecting window opening of vehicle using room camera | |
| KR102836440B1 (en) | Elevator safety accident prevention device and method using TOF camera | |
| EP2805581B1 (en) | Method and algorithm for self-learning/auto-commissioning by multiple sensor elements for outdoor lighting application | |
| JP2020114777A (en) | Elevator system | |
| KR102819264B1 (en) | Automatic door open-shut system using walking line prediction system through body recognition and regression analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application | Ref document number:22969190 Country of ref document:EP Kind code of ref document:A1 | |
| WWE | Wipo information: entry into national phase | Ref document number:2024565476 Country of ref document:JP | |
| NENP | Non-entry into the national phase | Ref country code:DE |