Patent Application
20150348416
The present invention relates to an obstacle detection device which is used being mounted on an electric-powered vehicle such as an electric-powered wheelchair, and the electric-powered vehicle provided therewith.
In recent years, three-wheel or four-wheel one-sheet electric-powered vehicles which are made for elderly people have been in widespread use and known as electric-powered wheelchairs and mobility scooters.
The electric-powered vehicle has a small vehicle body compared to a car and are used by elderly people in many cases. Therefore, an obstacle such as a step is a risky place with a possibility of falling down for the electric-powered vehicle and it is desired to detect the obstacle in advance to notify a driver.
Thus, a technology by which an obstacle which is on a road is detected by a radar, ultrasonic waves, a camera image or the like has been proposed conventionally. Specifically, for example, for a step detection device and an electric-powered vehicle provided therewith described in PTL 1, an obstacle detection technology by using a camera image is disclosed, and for an obstacle detection device described in PTL 2, an obstacle detection technology by using ultrasonic waves is disclosed.
PTL 1: Japanese Unexamined Patent Application Publication No. 2011-177334 (Publication date: Sep. 15, 2011)
PTL 2: Japanese Unexamined Patent Application Publication No. 2011-133247 (Publication date: Jul. 7, 2011)
PTL 3: Japanese Unexamined Patent Application Publication No. 2005-106649 (Publication date: Apr. 21, 2005)
However, the aforementioned conventional obstacle detection device and electric-powered vehicle provided therewith have following problems.
First, the electric-powered vehicles such as electric-powered wheelchairs and mobility scooters are treated as pedestrians, and therefore also pass on a place where no car travels, such as a passage in a hospital or a shop in addition to an outside sidewalk in some cases, so that it is necessary to assume usage under a wide variety of situations.
Against this, the step detection device and the electric-powered vehicle provided therewith disclosed in PTL 1 propose a technology by which an obstacle is detected from camera images which are captured at different times. In a method using a camera, however, there may be a case where erroneous judgment is caused under a situation where a desired image is not able to be obtained because of an irradiation state of illumination light, reflection of a target object or the like. For example, in the case of weather like rainy weather or fog, an illumination condition like a dark place or backlight, or having reflection of a floor surface or reflection of a surrounding sight, probability of causing erroneous judgment increases.
Moreover, the obstacle detection device disclosed in PTL 2 proposes a technology by which ultrasonic waves are irradiated to detect an obstacle from a time delay of reflective waves from a target object. However, similarly, probability of causing erroneous judgment increases in the case of weather like rainy weather or fog, or for an obstacle of a shape from which reflective waves are difficult to be returned because of a curved surface or an inclined surface, for descending stairs or a step of a concave shape like a ditch from which reflection waves are not returned, or an obstacle which absorbs ultrasonic waves.
Accordingly, it is preferable in the obstacle detection device that reliability of a determination result is able to be grasped in consideration of risk factors by which probability of causing erroneous judgment increases.
On the other hand, in a car navigation field which is a field different from that of the invention of the present application, a technology regarding visibility of a navigation screen that allows driving without seeing a monitor screen and without moving a sight line by projecting a scenery of the navigation screen on a windshield or the like is disclosed in PTL 3.
As described that “Target objects include ones which are able to be identified visually, such as gas stations, convenience stores, restaurants, hotels, hot spring resorts, facilities including public or other buildings, or landforms including mountains, rivers or lakes.”, this PTL 3 targets a scenery and is not regarding risk factors. Moreover, as described that “Specific precision means accuracy that instruction display (such as an arrow) which is displayed being superposed with an actual scenery indicates a target object, and high (excellent) specific precision means that the instruction display (such as an arrow) indicates the target object in the actual scenery accurately,”, precision with respect to a positional relation with the scenery which is actually seen by a driver is shown. Accordingly, it is not judgment precision with respect to a risk degree.
Further, though it is described that “Precision detection means detects specific precision based on oscillation of a vehicle, behavior of the vehicle, posture change of an occupant of the vehicle, specific precision of a current position of the vehicle, an apparent interval of a plurality of target objects to be displayed, an apparent size of a target object to be displayed, or a distance to a target object to be displayed.”, this precision detection means is neither regarding detection of risk factors.
In this manner, it can be also said that the technology of PTL 3 is similar in terms of notifying the occupant of specific precision, but a technical field and an object thereof are different from those of the invention of this application as well as judgment precision of a risk degree and the precision detection means are also different from those of the invention of this application.
The present invention has been made in view of the aforementioned conventional problem, and an object thereof is to provide an obstacle detection device capable of being used more safely and at ease by reducing erroneous judgment, and an electric-powered vehicle provided therewith.
An obstacle detection device in one aspect of the present invention is an obstacle detection device that has an obstacle detection sensor for detecting an obstacle, and notifies a user when the obstacle is detected by the obstacle detection sensor, the obstacle detection device including: a reliability calculation unit for calculating reliability of detection of the obstacle by the obstacle detection sensor, a first determination unit for comparing the reliability which is calculated with a threshold and determining that the reliability has decreased when the reliability is less than or equal to the threshold, and a first attention drawing unit for drawing attention of the user when the first determination unit determines that the reliability has decreased.
An electric-powered vehicle in one aspect of the present invention comprises the obstacle detection device described above.
According to one aspect of the present invention, an effect is exerted that an obstacle detection device capable of being used more safely and at ease by reducing erroneous judgment and an electric-powered vehicle provided therewith are provided.
a) and 2(b) are perspective views showing a configuration of an electric-powered vehicle provided with the aforementioned obstacle detection device.
a) is a view showing a first image captured by a left-side camera, 4(b) is a view showing a second image captured by a right-side camera, and 4(c) is a view in which the first image and the second image are superposed and only each border line between a sidewalk and a road surface is extracted.
a) is an explanatory view showing a method for calculating a distance, which shows the first image in a coordinate space, and 5(b) is a side view showing a focal plane of the cameras and position information of the cameras.
a) shows a method for calculating disparity, which is a side view showing the camera and 6(b) is a plan view showing the right and left cameras.
a) is a view showing the first image on a left side in which an origin P is seen at a coordinate of an origin (0, 0), and 7(b) is a view showing the second image on a right side in which the origin P is seen at a point of (−v1, 0).
a) is a side view showing a case where disparity v2 of an object is positive (larger than disparity v1 of a road surface), and 8(b) is a side view showing a case where the disparity v2 of the object is negative (smaller than the disparity v1 of the road surface).
a) is a front view showing a configuration when an LED display unit is lit in a notification unit of the aforementioned obstacle detection device, and 9(b) is a front view showing a configuration when only an upper-right LED display unit indicating a risky position and a risk level is lit in the notification unit of the aforementioned obstacle detection device.
Description will be given as follows for one embodiment of the present invention based on
An electric-powered vehicle provided with an obstacle detection device of the present embodiment is a vehicle which is movable by electric power, and represents, for example, a one-sheet small vehicle such as an electric-powered wheelchair or a mobility scooter, an electric car, or the like.
Moreover, an obstacle is an object which obstructs passing of the electric-powered vehicle, and represents, for example, a ditch, a step in a lower direction than a ground surface such as descending stairs, a pedestrian, a wall, a step in an upper direction than a ground surface such as ascending stairs, a slope having a large inclination angle, for example, 10° or more, or the like.
Description will be given for a configuration of the electric-powered vehicle provided with the obstacle detection device of the present embodiment based on
An electric-powered vehicle 1 provided with an obstacle detection device 4A of the present embodiment includes an obstacle detection sensor unit 10 which is fixed to a sensor fixation unit 2 provided in a forefront part in an advancing direction in the electric-powered vehicle 1 as shown in
The aforementioned obstacle detection sensor unit 10 detects an obstacle in the advancing direction, and has two cameras 12a and 12b as an obstacle detection sensor and a stereo camera in a sensor cover 11 as shown also in
That is, in the obstacle detection sensor unit 10 of the present embodiment, with a stereo camera method using the two cameras 12a and 12b, a distance to the obstacle and a height of the obstacle are measured from disparity of both of the cameras 12a and 12b.
Here, description will be given for a method for calculating a distance to and a height of an obstacle from disparity of the stereo camera based on
First, when a front side is shot by each of the cameras 12a and 12b, the first image captured by the left-side camera 12b shown in
In the present embodiment, by comparing such disparity v1 on the flat road surface with actual disparity v2 obtained by capturing an image of a detection region of a step, a height from the road surface of the detection region is detected.
Specifically, the disparity v1 when it is assumed that, with respect to the detection region with an arbitrary coordinate (X, Y) of the first image as a center, from a Y-coordinate thereof, position information of the camera 12b and the like, the detection region is on the road surface is obtained. Then, a comparison region with a coordinate (X−v1, Y) obtained by being shifted by the disparity v1 in the second image as a center is defined, so that the height from the road surface in the detection region is able to be obtained from the disparity v2 of the image of the detection region and the image of the comparison region.
Description will be given for a specific method for calculating a distance to and a height of an obstacle from disparity of the stereo camera based on
First, as shown in
In order to obtain the disparity v1, first, as shown in
Subsequently, an arbitrary coordinate point (X, Y) of the detection region in which the step is detected is selected in the coordinate space CP. The coordinate space CP corresponds to the focal plane A1 which is a plane vertical to optical axes of the cameras 12a and 12b shown in
Next, the distance d1 from the camera 12b to the origin P of the focal plane A1 when it is assumed that the coordinate point (X, Y) exists on the road surface is obtained. Here, calculation of the distance d1 with the coordinate point (X, Y) as a base point brings complication. Thus, by utilizing that all the coordinate points existing on the focal plane A1 have same disparity, the distance d1 to the origin P of the focal plane A1 is calculated with a coordinate point Q (0, Y) which is on the same focal plane A1 with the coordinate point (X, Y) as the base point.
Next, as shown in
θy=arctan(tan θ2×Y/h) (formula 1).
Moreover, as shown in
d1′=hc/sin(θ3+θy) (formula 2).
Accordingly, the distance d1 from the camera 12b to the origin P of the focal plane A1 is able to be obtained with:
d1=d1′×cos(δy) (formula 3).
Next, disparity v1 of the first image and the second image when it is set that the road surface is on the focal plane A1 is obtained. In the case of the lens having no distortion, since all points on the focal plane A1 are considered to have same disparity, the disparity v1 may be obtained by using the distance d1 obtained above (formula 3). Note that, due to distortion of the lens or the like, another value is used or correction is required for the distance d1 in some cases.
As shown in
θx=arctan(g/d1) (formula 4)
At this time, the origin P is seen at the coordinate of the origin (0, 0) in the first image on the left side as shown in
v1=w×tan θx/tan θ1 (formula 5)
Since v1 which is the pixel number of disparity is the same at the point P and the point (X, Y), it is to be seen at a position of (Xr, Y) in the right image. Here, it is set that:
Xr=X−v1 (formula 6).
Next, whether an object appearing at a coordinate of a detection region (X, Y) of the first image on the left side is at the same height with the road surface is determined. At this time, it may be confirmed whether an object same as the object appearing at the coordinate of (X, Y) of the first image on the left side appears at a position of a comparison region (Xr, Y) of the second image on the right side.
Though methods for confirming whether to be the same object includes various methods, for example, comparison may be made by extracting luminance for some pixels surrounding target points of the right and left images. When both are matched within a range of error factors such as noise of the cameras 12a and 12b, that point is able to be determined as being at the same height with the road surface. If both are not matched and are determined to be shifted to either leftward or rightward, it is possible to determine as being at a higher position or a lower position than the road surface according to disparity thereof.
For example, in a case where it is determined that an object which is at a position of (X, Y) of the first image on the left side is at a coordinate of (Xr−v2, Y) in the second image on the right side, when an actual distance to the object is set as d2, it is set that with (formula 4) and (formula 5):
v1=(w×g)/(d1×tan θ1)
v1+v2=(w×g)/(d2×tan θ1).
When v1 is eliminated, it is set that:
d2=(d×w×g)/(w×g+v2×tan θ1) (formula 7).
Here, as shown in
hs=hc×(d1−d2)/d1 (formula 8),
which shows the height hs of a step on the road surface. That is, when the disparity v2 of the object is positive (larger than the disparity v1 of the road surface), as shown in
The aforementioned procedure is repeated also at other coordinate points with an appropriate interval, and when detection of a height from the road surface is completed in a required range in the image, it is finished.
Note that, a method for calculating a distance to and a height of an obstacle from disparity of the stereo camera is not necessarily limited thereto and other well-known methods are usable.
Next, when detecting an obstacle by the obstacle detection sensor unit 10, the notification unit 20 shown in
As the notification unit 20, for example, one shown in
Meanwhile, detection of an obstacle by the stereo camera method is generally used because of having a simple configuration and being capable of measurement with relatively high precision. However, because of using camera images, an image of a target obstacle becomes difficult to be captured correctly depending on an illumination condition like a dark place or backlight or a weather condition including rainy weather or fog, so that erroneous judgment is caused in some cases. For example, it is a case where it is judged that there is no obstacle even though there is one or vice versa.
Moreover, when an obstacle on a road surface is detected, it is necessary to capture an image including the road surface, but when the road surface is in a mirror state or when being wet, a surrounding sight or illumination is reflected to be erroneously judged as an obstacle in some cases.
Thus, the obstacle detection device 4A of the present embodiment has the reliability calculation 14 and the determination unit 15 in order to judge that a situation is such that erroneous judgment is likely to be caused with the stereo camera method using the cameras 12a and 12b, that is, a situation is such that reliability is low as shown in
In the aforementioned obstacle detection device 4A, the reliability calculation unit 14 calculates reliability based on signals of the cameras 12a and 12b. As the signals, luminance, a correlation degree, contrast, an exposure time or a combination thereof is preferably used. Description will be given below for calculation of reliability of each of them.
Reliability is calculated from luminance values of images captured by the cameras 12a and 12b. For example, when outputting luminance at a grayscale of 255, a pixel number with a luminance threshold of 230 or more is counted, and when the counted number is a fixed number or more, it is judged that the reliability is low. Thereby, it is possible to prevent that erroneous judgment is easily caused, for example, when there is much reflective light on a floor surface or a glossy surface like metal as well as to draw attention of a user to that obstacle detection is not performed successfully.
Moreover, it is more preferable to reduce erroneous judgment by ignoring pixels with the luminance threshold or more and detecting an obstacle with the stereo camera method. The luminance threshold and the counted number are able to be set arbitrarily. In addition, in a case where comparison of images of the cameras 12a and 12b is performed for each fixed block region in addition to count with the pixel number, when the pixel number with the luminance threshold or more is more than or equal to a fixed value in the block, the block may be ignored. For example, a method in which a range of around 64×32 pixels is set as one block and disparity at feature points of right and left images in the block is calculated is generally used in the stereo camera method. In this case, reliability of judgment in the block is discriminated from the pixel number with the luminance threshold or more included in the block, and when the number of blocks having low reliability is more than or equal to a fixed number, a user may be notified.
In the stereo camera method, a shifted amount of disparity of feature points in fixed regions of images of the right and left cameras 12a and 12b is calculated and converted to a distance. At this time, a correlation degree which is a coincidence degree of the feature points is calculated. As the correlation degree is high, reliability of a result is considered to be high, and when the correlation degree is less than or equal to an arbitrary threshold, it is judged that the reliability of the result for the region is low. Further, when there is a fixed number or more of regions with low reliability, the reliability of the detection result itself is judged to be low and notified to the user.
For example, when a surrounding sight is reflected on a floor surface, it is erroneously judged that there is an obstacle even though there is none in some cases, but the reflected sight is generally not clear and an image becomes blurred. In this case, it becomes possible that the user is notified and warned to travel carefully when a situation where erroneous judgment is easily caused due to the reflection because the correlation degree has also decreased comes.
When disparity of the right and left images is calculated in a unit of a block as described above with the stereo camera method, a contrast difference (distribution) in the block may be calculated, a block whose contrast is lower than a fixed value may be judged as having low reliability, and the user may be notified when the number of blocks having low reliability is more than or equal to a fixed number. Moreover, erroneous detection of an obstacle may be reduced by ignoring a block having low reliability at the time of obstacle judgment.
By calculating reliability with contrast, it is possible to draw attention of the user by notifying that a situation is such that erroneous judgment is made easily in the case of a dark place like night-time or when an image is too bright due to backlight or the like.
Image capturing at the cameras 12a and 12b may be performed with automatic exposure to calculate reliability from an exposure time. When the exposure time is more than or equal to a fixed time or less than or equal to a fixed time, it is possible to determine that an image-capturing environment is too dark or too bright, and judge that detection precision for an obstacle has decreased.
Thereby, it is possible to draw attention of the user by notifying that a situation is such that an obstacle is erroneously judged easily in a situation of a dark place, backlight or the like.
By using signals of the cameras themselves, it is possible to discriminate brightness of an image-capturing environment without using an illuminance sensor or the like separately, and a configuration at low cost becomes possible.
In this manner, an obstacle is detected by the stereo camera in the obstacle detection device 4A of the present embodiment. In addition, signals of the stereo camera are used for judgment of reliability. Moreover, the reliability is calculated from signals other than disparity information of the stereo camera. Examples thereof include luminance, a correlation degree, contrast and an exposure time.
Description will be given for a judgment flow of reliability in the obstacle detection device 4A with the aforementioned configuration based on
As shown in
In this manner, the obstacle detection device 4A of the present embodiment has the cameras 12a and 12b as the obstacle detection sensor for detecting an obstacle, and notifies the user when the obstacle is detected by the cameras 12a and 12b.
Here, for example, when the reliability of the cameras 12a and 12b has decreased because of fog or the like, etc., measurement is not able to be performed successfully, resulting that erroneous judgment is caused in some cases. In this case, in a case where, even when the reliability of the cameras 12a and 12b has decreased, the user believes the judgment as it is, an accident may be led, for example, when it is judged that there is no step even though there is a step.
Thus, in the present embodiment, the reliability calculation unit 14 for calculating reliability of detection of an obstacle by the cameras 12a and 12b, the determination unit 15 as a first determination unit for comparing the reliability which is calculated with a threshold and determining that the reliability has decreased when the reliability is less than or equal to the threshold, and the notification unit 20 as a first attention drawing unit for drawing attention of the user when the determination unit 15 determines that the reliability has decreased are provided.
Thereby, by drawing attention of the user in a situation where the reliability of judgment is low and erroneous determination is easily caused, the user is able to use the obstacle detection device 4A more safely and at ease. In a case where a situation where measurement is not able to be performed successfully, for example, when the reliability of the cameras 12a and 12b has decreased due to fog or the like, etc., comes, attention of the user is drawn to that effect. Thereby, the user is to perform operation more carefully without taking a result of the obstacle detection device 4A on faith, thus making it possible to prevent an accident.
Accordingly, it is possible to provide the obstacle detection device 4A capable of being used more safely and at ease by reducing erroneous judgment.
Moreover, in the obstacle detection device 4A of the present embodiment, the obstacle detection sensor is composed of the cameras 12a and 12b as the stereo camera for detecting an obstacle from disparity of a plurality of cameras. Thereby, it is possible to make the obstacle detection device easily with an inexpensive system by using the stereo camera which is conventionally and generally used as the obstacle detection sensor.
Moreover, in the obstacle detection device 4A of the present embodiment, the reliability calculation unit 14 calculates reliability from images captured by the cameras 12a and 12b.
Thereby, it is possible that a separate sensor for measuring reliability is not required by using images captured by the stereo camera as a method for calculating the reliability, thus making it possible to detect the reliability at low cost.
Here, as the method for calculating reliability from images captured by the stereo camera, for example, a luminance value, contrast, a correlation degree, or an exposure time is usable. This makes it possible to judge the reliability easily, and to easily judge that the reliability has decreased when disparity judgment precision has reduced due to lacking of a light quantity at night-time, local light incidence at the time of backlight, or low contrast.
Moreover, when the aforementioned correlation degree is used, a correlation degree indicating a matching degree of feature points of images of right and left cameras of the stereo camera is used as the method for calculating reliability. This makes it possible to judge the reliability, and, for example, by utilizing that the correlation degree decreases at the time of rainy weather or when blur of an image is generated due to reflecting on a road surface or the like, the decrease in the reliability is able to be detected to draw attention of the user.
Moreover, in the obstacle detection device 4A of the present embodiment, the reliability calculation unit 14 is able to calculate reliability based on luminance of images captured by the cameras 12a and 12b.
This makes it possible to judge the reliability easily by using a luminance value and to easily judge that the reliability has decreased when disparity judgment precision has reduced due to lacking of a light quantity at night-time, or local light incidence at the time of backlight. Moreover, both of a luminance value and a correlation degree are usable together. Thereby, it is possible to draw attention of the user to that the reliability has decreased in the case of weather like rainy weather or fog, an illumination condition like a dark place or backlight, or generation of reflecting of a floor surface or reflection of a surrounding sight. In this case, the user is able to perform operation more at ease by paying attention.
Moreover, in the obstacle detection device 4A of the present embodiment, the notification unit 20 as the first attention drawing unit is able to draw attention of the user to that reliability has decreased with at least one of sound, voice and display. This makes it possible to surely notify the user that the reliability has decreased aurally and visually.
In addition, the electric-powered vehicle 1 of the present embodiment is provided with the obstacle detection device 4A of the present embodiment. With the aforementioned configuration, it is possible to provide the electric-powered vehicle 1 capable of driving safely and at ease.
Moreover, it is possible to set that the electric-powered vehicle 1 of the present embodiment is provided with a driving control unit for controlling driving operation, and the driving control unit controls the driving operation based on a detection result of the aforementioned obstacle detection device.
This makes it possible to perform driving more safely and to prevent control mistake by erroneous judgment by not performing driving control of the electric-powered vehicle 1 when reliability is low.
Description will be given as follows for another embodiment of the present invention based on
An obstacle detection sensor unit 30 in an obstacle detection device 4B of the present embodiment is different in terms of having a reliability detection sensor 31 as shown in
That is, the obstacle detection sensor unit 30 in the obstacle detection device 4B of the present embodiment detects an obstacle by the cameras 12a and 12b serving as the stereo camera. Moreover, it is set that, for example, an illuminance sensor or a temperature and humidity sensor is used other than the stereo camera for judgment of reliability.
Description will be given for a configuration of the obstacle detection sensor unit 30 of the present embodiment based on
The obstacle detection sensor unit 30 of the present embodiment includes the cameras 12a and 12b and a reliability detection sensor 31 in a sensor cover 11 as shown in
As the aforementioned reliability detection sensor 31, for example, an illuminance sensor, a temperature and humidity sensor or the like is usable. When the illuminance sensor is used as the reliability detection sensor 31, illuminance surrounding the obstacle detection sensor unit 30 is detected, and when the illuminance is more than or equal to a fixed value or less than or equal to a fixed value, it is judged that the reliability has decreased. That is, when the illuminance is high, detection reliability at the cameras 12a and 12b serving as the stereo camera is likely to have decreased due to effects of backlight or reflective light. Moreover, when the illuminance is low, the detection reliability at the cameras 12a and 12b is likely to have decreased at night-time or in a dark place.
In addition, by using the temperature and humidity sensor as the reliability detection sensor 31, it is possible to predict a situation where rainy weather, fog, or condensation is easily caused, etc., and calculate the reliability from a result thereof.
In this manner, like the obstacle detection sensor unit 30 of the present embodiment, by using one other than the stereo camera as data for calculating reliability, it is possible to calculate reliability which is difficult to be calculated only with the stereo camera and to calculate the reliability more surely.
Description will be given for an operation flow of the obstacle detection device 4B provided with the obstacle detection sensor unit 30 with the aforementioned configuration based on
As shown in
In this manner, in the obstacle detection device 4B of the present embodiment, the reliability calculation unit 14 is provided with the reliability detection sensor 31 as the illuminance sensor for detecting surrounding brightness, and calculates reliability based on an output value of the reliability detection sensor 31.
Thereby, it becomes possible to judge reliability easily by detecting lightness and darkness of a measurement environment by the reliability detection sensor 31 as the illuminance sensor and determining that the reliability has decreased in the case of being too dark or in the case of being too bright. In addition, it becomes possible to perform judgment even in a situation where reliability is difficult to be judged with camera images.
Moreover, in the obstacle detection device 4B of the present embodiment, the notification unit 20 as the first attention drawing unit is able to draw attention of a user to that the reliability has decreased with at least one of sound, voice and display. This makes it possible to surely notify the user that the reliability has decreased aurally and visually.
In addition, the electric-powered vehicle 1 of the present embodiment is provided with the obstacle detection device 4B of the present embodiment. With the aforementioned configuration, it is possible to provide the electric-powered vehicle 1 capable of driving safely and at ease.
Moreover, it is possible to set that the electric-powered vehicle 1 in the present embodiment is provided with a driving control unit for controlling driving operation, and the driving control unit controls the driving operation based on a detection result of the aforementioned obstacle detection device. This makes it possible to perform driving more safely and to prevent control mistake by erroneous judgment by not performing driving control of the electric-powered vehicle when reliability is low.
Description will be given as follows for still another embodiment of the present invention based on
An obstacle detection sensor unit 40 in an obstacle detection device 4C of the present embodiment is different in that a plurality of obstacle detection sensors are used together as shown in
Description will be given for a configuration of the obstacle detection sensor unit 40 in the obstacle detection device 4C of the present embodiment based on
The obstacle detection sensor unit 40 of the present embodiment is provided with the first detection sensor 41 and the second detection sensor 42 as shown in
For example, a stereo camera method is used as the first detection sensor 41 as well as the ultrasonic sensor is used as the second detection sensor 42. In this manner, sensors having different methods, that is, different detection principles are preferably combined for the first detection sensor 41 and the second detection sensor 42.
Description will be given for an operation flow of the obstacle detection device 4C provided with the obstacle detection sensor unit 40 of the present embodiment based on
As shown in
In this manner, in the obstacle detection device 4C in the present embodiment, the obstacle detection sensor is configured by the first detection sensor 41 and the second detection sensor 42 serving as the plurality of obstacle detection sensors having different detection methods from each other. In addition, a determination unit 15 as a second determination unit for determining that detection reliability has decreased when judgment results are different between the plurality of the first detection sensor 41 and the second detection sensor 42, and a notification unit 20 as a second attention drawing unit for drawing attention of a user when the determination unit 15 determines that the reliability has decreased are provided.
This makes it possible to determine that that the reliability is high in the case of same judgment when using a plurality of sensors having different characteristics, like the stereo camera and the ultrasonic sensor, and that the reliability is low in the case of being different. As a result thereof, it is possible to perform judgment of the reliability more surely. That is, an obstacle is detected at the first detection sensor 41 and the second detection sensor 42 serving as the plurality of obstacle detection sensors having different methods, and when results of both of them are not coincident, the reliability is judged to be low to be notified to a user. This makes it possible to notify the user that there is an obstacle only when detection has been performed more surely, and obtain the obstacle detection device 4C having very high reliability.
Moreover, in the obstacle detection device 4C of the present embodiment, the notification unit 20 as the second attention drawing unit is able to draw attention of the user to that the reliability has decreased with at least one of sound, voice and display. This makes it possible to surely notify the user that the reliability has decreased aurally and visually.
In addition, the electric-powered vehicle 1 in the present embodiment is provided with the obstacle detection device 4C of the present embodiment. With the aforementioned configuration, it is possible to provide the electric-powered vehicle 1 capable of driving safely and at ease.
Moreover, the electric-powered vehicle 1 in the present embodiment is provided with a driving control unit for controlling driving operation, and the driving control unit controls the driving operation based on a detection result of the obstacle detection device 4C. This makes it possible to perform driving more safely and to prevent control mistake by erroneous judgment by not performing driving control of the electric-powered vehicle 1 when the reliability is low.
An obstacle detection device 4A in one aspect of the present invention is the obstacle detection device 4A having an obstacle detection sensor (cameras 12a, 12b) for detecting an obstacle and notifying a user when the obstacle is detected by the obstacle detection sensor (cameras 12a, 12b), the obstacle detection device including: a reliability calculation unit 14 for calculating reliability of detection of the obstacle by the obstacle detection sensor (cameras 12a, 12b), a first determination unit (determination unit 15) for comparing the reliability which is calculated with a threshold and determining that the reliability has decreased when the reliability is less than or equal to the threshold, and a first attention drawing unit (notification unit 20) for drawing attention of the user when the first determination unit (determination unit 15) determines that the reliability has decreased are provided.
With the aforementioned configuration, the reliability calculation unit for calculating reliability of detection of the obstacle by the obstacle detection sensor, the first determination unit for comparing the reliability which is calculated with the threshold and determining that the reliability has decreased when the reliability is less than or equal to the threshold, and the first attention drawing unit for drawing attention of the user when the first determination unit determines that the reliability has decreased are provided.
Thereby, by drawing attention of the user in a situation where reliability of judgment is low and erroneous determination is easily caused, the user is able to use the obstacle detection device more safely and at ease. In a case where a situation where measurement is not able to be performed successfully, for example, when reliability of the obstacle detection sensor has decreased due to fog or the like, etc., attention of the user is drawn to that effect, and thereby, the user is to perform operation more carefully without taking a result of the obstacle detection device on faith, thus making it possible to prevent an accident.
Accordingly, it is possible to provide the obstacle detection device capable of reducing erroneous judgment and being used more safely and at ease.
In the obstacle detection device 4A in an aspect 2 of the present invention, the obstacle detection sensor may comprise a stereo camera (cameras 12a, 12b) for detecting the obstacle from disparity of a plurality of cameras in the obstacle detection device 4A of the aspect 1.
Thereby, it is possible to make the obstacle detection device easily with an inexpensive system by using the stereo camera which is conventionally and generally used as the obstacle detection sensor.
In the obstacle detection device 4A in an aspect 3 of the present invention, it is preferable that the reliability calculation unit 14 calculates the reliability from images captured by the stereo camera (cameras 12a, 12b) in the obstacle detection device 4A of the aspect 2.
Thereby, it is possible that a separate sensor for measuring reliability is not required by using images captured by the stereo camera as a method for calculating reliability, thus making it possible to detect the reliability at low cost.
Here, as the method for calculating reliability from images captured by the stereo camera, for example, a luminance value, contrast, a correlation degree, or an exposure time is usable. This makes it possible to judge the reliability easily, and to easily judge that reliability has decreased when disparity judgment precision has reduced due to lacking of a light quantity at night-time, local light incidence at the time of backlight, or low contrast.
Moreover, when the aforementioned correlation degree is used, a correlation degree indicating a matching degree of feature points of images of right and left cameras of the stereo camera is used as the method for calculating reliability. This makes it possible to judge reliability, and, for example, by utilizing that the correlation degree decreases at the time of rainy weather or when blur of an image is generated due to reflecting on a road surface or the like, the decrease in the reliability is able to be detected to draw attention of the user.
In the obstacle detection device 4A in an aspect 4 of the present invention, the reliability calculation unit 14 may calculates the reliability based on luminance of the images captured by the stereo camera (cameras 12a, 12b) in the obstacle detection device 4A of the aspect 3.
This makes it possible to judge reliability easily by using a luminance value and to easily judge that the reliability has decreased when disparity judgment precision has reduced due to lacking of a light quantity at night-time, or local light incidence at the time of backlight. Moreover, by using both of a luminance value and a correlation degree together, it is possible to draw attention of the user to that the reliability has decreased in the case of weather like rainy weather or fog, an illumination condition like a dark place or backlight, or generation of reflection of a floor surface or reflection of a surrounding sight. In this case, the user is able to perform operation more at ease by paying attention.
In an obstacle detection device 4B in an aspect 5 of the present invention, the reliability calculation unit 14 may be provided with an illuminance sensor (reliability detection sensor 31) for detecting surrounding brightness, and may calculate the reliability based on an output value of the illuminance sensor (reliability detection sensor 31) in the obstacle detection device according to any one of the aspects 1 to 4.
Thereby, it becomes possible to judge reliability easily by detecting lightness and darkness of a measurement environment by the illuminance sensor and determining that the reliability has decreased in the case of being too dark or in the case of being too bright. In addition, it becomes possible to perform judgment even in a situation where reliability is difficult to be judged with camera images.
In the obstacle detection device 4A, 4B in an aspect 6 of the present invention, the first attention drawing unit (notification unit 20) may draw attention of the user to that the reliability has decreased with at least one of sound, voice and display in the obstacle detection device 4A, 4B according to any one of the aspects 1 to 5.
This makes it possible to surely notify the user that the reliability has decreased aurally and visually.
In an obstacle detection device 4C in an aspect 7 of the present invention, the obstacle detection sensor may be configured by a plurality of obstacle detection sensors (first detection sensor 41 and second detection sensor 42) having different detection methods from each other, and the obstacle detection device may comprise a second determination unit (determination unit 15) for determining that detection reliability has decreased when judgment results are different between the plurality of obstacle detection sensors (first detection sensor 41 and second detection sensor 42), and a second attention drawing unit (notification unit 20) for drawing attention of the user when the second determination unit (determination unit 15) determines that the reliability has decreased in the obstacle detection device according to any one of the aspects 1 to 6.
This makes it possible to determine that the reliability is high in the case of same judgment when using the plurality of sensors having different characteristics, like the stereo camera and the ultrasonic sensor, and that the reliability is low in the case of being different. As a result thereof, it is possible to perform judgment of reliability more surely.
An electric-powered vehicle 1 in one aspect of the present invention comprises the obstacle detection device 4A, 4B or 4C according to any one of aspects 1 to 7.
With the aforementioned configuration, it is possible to provide the electric-powered vehicle capable of driving safely and at ease.
In the electric-powered vehicle 1 in an aspect 9 of the present invention, a driving control unit for controlling driving operation may be provided, and the driving control unit may control the driving operation based on a detection result of the obstacle detection device in the electric-powered vehicle 1 according to the aspect 8.
This makes it possible to perform driving more safely and to prevent control mistake by erroneous judgment by not performing driving control of the electric-powered vehicle when reliability is low.
Note that, the present invention is not limited to each of the embodiments described above, various modifications are possible within the scope indicated in Claims, and an embodiment acquired by combining appropriately technical means each disclosed in a different embodiment is also included in the technical scope of the present invention.
The present invention relates to an obstacle detection device which is used being mounted on an electric-powered vehicle such as an electric-powered wheelchair, and the electric-powered vehicle provided therewith, and is applicable to an electric-powered vehicle such as a three-wheel or four-wheel one-sheet electric-powered vehicle or mobility scooter which is made for elderly people, and an obstacle detection device mounted thereon.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2013-064962 | Mar 2013 | JP | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/JP2014/050583 | 1/15/2014 | WO | 00 |
