JPH09326096A - Human obstacle detector for vehicle - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a human obstacle detector improving detecting precision by detecting a human obstacle in a larger detecting size as it is nearer corresponding to a distance at each set detecting area. SOLUTION: An infrared camera 1 outputs an infrared video signal anterior to a vehicle to an obstacle detecting part 2. A navigation system 3 continuously reads traveling road information of the vehicle from a map memory to output to a detecting area setting part 4. The setting part 4 calculates the position of the road on which its own vehicle travels in an infrared video based on inputted road information. On the traveling road, three detecting areas are set corresponding to the distance and the size of each detecting area is set corresponding to a detecting size within it and the far and near of the distance. An obstacle detection part 2 binarizes video by setting the temperature of the exposed part of a human body to be a threshold value and executes graphic detection. The detected graphic is compared with the detecting size of the detecting area to detect whether it is a human obstacle. The human obstacle is position-displayed by a head up display through a display position setting part 5.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a structure of an apparatus for detecting a human obstacle by image processing of an infrared image in front of a vehicle, which is used for safe driving of a vehicle. The present invention relates to an improved detection accuracy.

[0002]

2. Description of the Related Art Conventionally, as an apparatus capable of detecting a human obstacle, there is an obstacle detecting apparatus for a vehicle as disclosed in, for example, Japanese Patent Laid-Open No. 6-230132. In this obstacle detection device, for example, an infrared camera having a sensitivity near a wavelength of 10 μm is used, a two-dimensional change in infrared intensity of the human body temperature is detected, and the surface temperature of the exposed portion of the human body is detected from the temperature image. A human obstacle is detected by selecting with a comparator having a temperature range of ± 1 ° C as its threshold value.
Further, based on the result of the selection of the comparator, in order to improve the detection reliability, the infrared source having a round shape such as the face of a human body is set as the detection target, and the infrared source having a rectangular shape including the vertical and horizontal directions is removed in advance. Image processing will be added.

[0003]

However, in the conventional one as described above, as the round infrared ray source, the reflection of sunlight by the building or the heat source other than the human body happens to be circular, and There are many things on the road that are close to the human body temperature, and detection without excluding them will include erroneous detection in the detection results. In order to avoid this, for example, it is easily conceivable to narrow the threshold value range, but a new problem occurs that the detection rate decreases. The present invention has been made in view of the above problems, and an object of the present invention is to provide a human obstacle detection device for a vehicle, the detection accuracy of which is further improved.

[0004]

Therefore, according to the present invention, there is provided an image pickup means for picking up a traveling direction of a vehicle to obtain an infrared image, and a human obstacle on the infrared image corresponding to a distance from the vehicle. A detection area setting means for setting a detection area and an obstacle detection means for detecting a human obstacle with a larger detection size as the distance from the set detection area is closer to the distance are provided.

It is desirable that the detection area setting means sets the detection area above and above the infrared image as the distance becomes shorter. Further, the detection area setting means includes a travel area detection means, the travel area detection means detects an area in which the host vehicle travels in the infrared image, and sets the detection area on the detected travel area. You can also The detection area setting means includes a navigation system having a built-in map memory that stores travel path information such as road width, and the travel area on the infrared image is based on the road width read by positioning of the navigation system. Can also be detected. The detection area setting means can also adjust the detection area in accordance with the traveling road curve and the undulation information included in the traveling road information.

[0006]

According to the present invention, the human detection area on the infrared image is set in correspondence with the distance from the vehicle, and the closer the distance is to each set detection area, the larger the detection size becomes, the larger the human obstacle is detected. Since detection is performed, the detection area can be set to a limited position and small, and the detection size can be changed based on the perspective relationship to keep false detection low while maintaining high detection rate. You can Also, by limiting the detection area on the infrared image to the vehicle travel area, only those who affect the travel of the vehicle are detected as obstacles, and it is easy for the driver to make decisions to avoid or take warnings. Becomes

[0007]

BEST MODE FOR CARRYING OUT THE INVENTION Next, embodiments of the present invention will be described with reference to examples. FIG. 1 shows the configuration of the first embodiment of the present invention. An infrared camera 1 is provided toward the front of the vehicle. The infrared camera 1 is composed of an optical system that collects infrared light and an infrared sensor that is arranged at the focal point of the optical system and that converts an infrared light image into an electrical signal and outputs it. The infrared sensor has a high resolution, for example, a temperature resolution of 0. ℃ 1
A pixel having the number of pixels of 80000 and a sensitivity of 8 to 14 μm in wavelength is used. The output of the infrared camera is the obstacle detection unit 2
Is sent to where the obstacles are detected.

A navigation system 3 is attached to the vehicle, and the GPS receiver 31 is a G satellite from the satellite.
By receiving the PS radio wave signal, current position data of the vehicle indicating latitude and longitude is generated. The controller 32 reads out map information including the road width and the like from the map memory 33 based on the position data and outputs it to the detection area setting unit 4. On the other hand, the navigation system 3 is provided with a gyro 34 and a vehicle speed sensor 35 that autonomously measure the position of the vehicle, and the measured value and the position data from the GPS receiver 11 are used properly or used in combination to perform positioning. The accuracy of the current position is improved.

The detection area setting unit 4 calculates the position of the road on which the vehicle runs in the infrared image captured by the infrared camera 1 based on the road information from the navigation system 3. A detection area is set above the road according to the distance. The obstacle detection unit 2 detects a human obstacle from each detection area on the input infrared image with a detection size based on the perspective relationship. And the display position setting unit 5 to which the position data of the detected human obstacle is connected
And output to the head-up display unit 6.

The head-up display section 6 has a plurality of display units, and each display unit is arranged on the instrument panel at a predetermined distance in the lateral direction, for example.
Each is projected from the opening formed on the upper surface of the instrument panel toward the front windshield. The display position setting section 5 selects the display unit in the vicinity of the detected human obstacle based on the position data of the human obstacle and lights it to indicate the position of the human obstacle.

Next, the principle of detecting a human obstacle will be described. FIG. 2 schematically shows a state in front of the vehicle captured by the infrared camera. Person a, person b, and person c are standing apart from each other. A nearby person a has a face at the top of the screen and is large. As the distance from the vehicle increases, the faces of the persons b and c move toward the screen vanishing point and become smaller. Therefore, in the case of detecting only the head, set the detection range large at the top of the image,
As the distance increases, it approaches the vanishing point and can be set smaller. In the present embodiment, in consideration of simplification of processing, the road is divided into three areas of near, middle, and long distance, and the boundaries between the areas are set to 15 m and 35 m. Note that the detection area is not set because the area below 5 m and above 50 m is outside the field of view of the infrared camera.

On the other hand, since the human head is almost circular, the filling rate, which is the ratio of the figure to the circumscribed quadrangle, is relatively high and is 70% or more. Therefore, it can be determined that the detected figure with the filling rate of less than 70% is not a human. In addition, extremely vertical and horizontal figures are not human heads or faces. In this embodiment, a judgment condition of less than 1: 3 is used to judge that.

The temperature of the exposed surface of the human body is 32 ° C. ±
Within the range of 2 ° C, there are some differences between summer and winter, but individual differences are small. Although there are some differences in head and face sizes between adults and children, there is little individual difference between adults. Therefore, in detecting the figure, 32 ℃ ± 2 ℃
The infrared luminance range corresponding to is set as a threshold value, and those having a luminance within the range of 32 ° C. ± 2 ° C. are displayed on the infrared image as white, and the other ranges are displayed as black. If the lump displayed in white is detected and the filling rate for the area of the square that touches the outer circumference is 70% or more, and the size (area) is within the range set for each distance from the vehicle, You can judge. If the season is limited, the threshold value can be the range of infrared luminance corresponding to 32 ° C. ± 1 ° C.

3 and 4 are flow charts showing the above processing in detail. That is, in step 100, an infrared image in front of the vehicle from the infrared camera 1 is read into the obstacle detection unit 2. In step 101, the read image is binarized by applying a threshold having a range of 32 ° C. ± 2 ° C. As a result, a binarized image in which the inside of 32 ° C. ± 2 ° C. is displayed in white and the others are displayed in black is created.

In step 102, a white portion gathered in the binarized image is detected as a figure. Step 1
In 03, the detected figures are labeled and the total number thereof is calculated. In step 104, the label counter is set and the graphic input is started.

In step 105, it is determined whether or not the area of the input graphic is within a specified size in order to exclude an extremely large object or a small object such as noise. If it is within the specified size, the process proceeds to step 106. If the size is not the specified size, the process proceeds to step 123. As a result, the size to be detected is extracted and becomes a detection candidate. In step 106, a quadrangle that touches the outer circumference of the detection figure that is a detection candidate is generated. Step 1
At 07, it is determined whether the area filling rate of the detected figure is 70% or more for the generated quadrangle. In the above case, it is determined in step 108 whether the aspect ratio of the quadrangle is smaller than 1: 3. If it is smaller, the process proceeds to step 109. If either of the determinations in steps 107 and 108 is negative, the process proceeds to step 123.

In step 109, the navigation system 3 measures the position of the vehicle, and in step 110, the road information on the traveling vehicle is input to obtain road width information. In step 111, the detection area setting unit 4 calculates the road position on the infrared image based on the road width information, the mounting height of the infrared camera, and the focal length of the lens. Step 112
In FIG. 2, the binarized image corresponds to the distance of 5 m to 15 m, 15 m to 35 m, and 35 to 50 m as shown in FIG.
As described above, the area is divided into three areas of area 1, area 2, and area 3. Areas with a short distance are set above the image and areas far are set below the area corresponding to the road width on the image. The detection range of the figure detection size of each area is large at a short distance depending on the mounting height of the infrared camera 1 and the lens focal length.
The long distance is set small.

In step 113, it is determined whether the graphic is in the area 1. If it is within the region 1, the range of the size of the figure corresponding to the region 1 (detection range 1) is output in step 114. If it is not in the area 1, the process proceeds to step 115. At step 115, it is determined whether or not the figure is within the area 2. If it is within the region 2, the range of the size of the figure corresponding to the region 2 (detection range 2) is output in step 116. Area 2
If not, the graphic is in area 3
An internal decision is made. If it is within the region 3, the range of the size of the figure corresponding to the region 3 (detection range 3) is output in step 118. If it is not within the region 3, the process proceeds to step 123. In step 119, it is determined whether the area (size) of the graphic is within the input detection range. If it is within the specified range, step 120
The barycentric coordinates of the figure are calculated with. In step 121,
The display position setting section 5 selects the display unit closest to the position of the center of gravity of the graphic, and outputs a control signal to this display unit.

In step 122, the selected display unit of the head-up display section 6 lights the alarm mark. Then, in step 123, the label counter is incremented. Step 124
Then, after it is judged whether the label counter value is larger than the total number of labels, the process returns to step 105 and the judgment for the next figure is made. When it is determined that the label counter value is larger than the total number of labels, the flow ends.

The present embodiment is configured as described above, divides the detection area of the human obstacle into three, and sets each detection area and the detection size in the detection area according to the distance. Then, the infrared image is binarized with the exposed head as a detection target, and the circular shape, which is a feature of the head, is determined by the filling rate of the circumscribed quadrangle and the aspect ratio of the quadrangle to detect the candidate. Since the detected area is further judged by the detection area as to whether it is within the predetermined detection range, the outside of the traveling road that does not hinder the running is excluded from the detection target, and the human obstacle is surely detected on the traveling road. The effect that can be obtained is obtained.

Next, a second embodiment will be described. In the first embodiment, the detection area is set corresponding to the width of the traveling road in the detection area setting unit 4, but in the present embodiment, instead of the detection area setting unit 4, the detection area setting unit 4 ′ is used. The position of the detection area can be changed according to the inclination of the road in the traveling direction of the vehicle and the curve.

The specific operation will be described with reference to FIG.
5A shows a straight road, and FIGS. 5B and 5C show a curved road. For example, if there is a right curve 100 m ahead in the vehicle traveling direction, this is recognized from the vehicle position information and the map information of the navigation system. Since the maximum detection distance of the human obstacle is 50 m, when the vehicle reaches 50 m, the road with the right curve intersects the right end of the detection range 3 as shown in (b). Therefore, from this time, the detection area 3 calculated corresponding to the traveling road width is shifted to the right of the screen, and the position is changed so that the detection area includes the straight traveling path and the right curve traveling path. 35m on the curve
The detection area 2 corresponding to 15 m to 35 m from the front side is shifted to the right, and the detection area 2 is repositioned so as to include the straight running path and the right curve running path.

When the curve is less than 15 m, the detection area 1 corresponding to 5 m to 15 m is shifted to the right as shown in FIG. 5 (c) to include the straight running path and the right curve running path. Change the position to. Also when moving out of a curve and returning to a straight road, the figure detection area on the screen is moved left and right so that all the detection areas include the straight travel path and the right curve travel path. In addition, in FIG. 5, the dotted line shows the position before the detection area is changed, and the solid line shows the state after the change. In the above, the right curve is taken as an example. Just shift the region in the opposite direction.

Since the present embodiment is configured as described above and the position of the detection area is changed according to the curve condition of the road, it is possible to cope with the change of the road image without expanding the detection area. The effect that the range of application is expanded is obtained. In addition,
In the present embodiment, the curve has been described, but in addition to this, it is possible to deal with, for example, ups and downs of the traveling road, in the same manner as described above. Incidentally. In the above description, the map information possessed by the navigation is used as a means for recognizing the road environment in front of the vehicle, but the present invention is not limited to this, and the white line on the road is detected from an infrared image or another visible camera, and this curve Therefore, it is possible to recognize the road environment such as curves and undulations and determine the detection area based on the recognition result. It is also possible to divide the detection area into a larger number or to change it steplessly according to the distance. .

[0025]

As described above, according to the present invention, the human detection area on the infrared image is set corresponding to the distance from the vehicle, and the distance is short according to each set detection area. Since human obstacles are detected with a relatively large detection size, the detection area can be set to a limited position and small, and a high detection rate can be achieved by changing the detection size based on the perspective relationship. While maintaining, false detection can be suppressed low. Further, when the detection area on the infrared image is limited to the traveling area of the vehicle, only the person who influences the traveling of the vehicle is detected, and it becomes easy for the driver to make a decision such as avoidance or warning.

[Brief description of drawings]

FIG. 1 is a block diagram illustrating a configuration of an embodiment.

FIG. 2 is a diagram for explaining setting of a detection range.

FIG. 3 is a flowchart showing a detection operation of a human obstacle.

FIG. 4 is a flowchart showing an operation of detecting a human obstacle.

FIG. 5 is an explanatory diagram of a second embodiment.

[Explanation of symbols]

 1 infrared camera 2 obstacle detection unit 3 navigation system 4 (4 ') detection area setting unit 5 display position setting unit 6 head-up display unit

Claims (5)

[Claims] 1. An image pickup means for picking up an infrared image by picking up a traveling direction of a vehicle, and a detection area setting means for setting a detection area of a human obstacle on the infrared image corresponding to a distance from the vehicle. An obstacle detecting device for a vehicle, comprising: an obstacle detecting unit that detects a human obstacle with a detection size that is larger for each of the set detection areas corresponding to the distance. 2. The human obstacle detection device for a vehicle according to claim 1, wherein the detection area setting unit sets the detection area above and above the infrared image as the distance becomes shorter. 3. The detection area setting means includes a travel area detection means, the travel area detection means detects a travel area of a vehicle in the infrared image, and the detection area is set on the detected travel area. 3. The setting is performed according to claim 2.
The human obstacle detection device for a vehicle described. 4. The detection area setting means has a navigation system having a built-in map memory in which road information such as road width is stored, and the detection area setting means is based on the road width read by positioning of the navigation system. The human obstacle detection device for a vehicle according to claim 3, wherein a traveling area on the infrared image is detected. 5. The human obstacle for a vehicle according to claim 4, wherein the detection area setting means adjusts the detection area in accordance with a traveling road curve and undulation information included in the traveling road information. Detection device.