JP3366940B2 - Automatic identification device for people, cars, etc. on the road by infrared camera - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an identification device for automatically identifying vehicles, people and other objects on a road.

[0002]

2. Description of the Related Art Conventionally, vehicles, people and other objects on a road have been identified artificially by visually observing an image taken by a visible camera such as ITV. In such an identification method, there is a problem that the road information required by the vehicle automatic traveling system which is planned in the future cannot be obtained accurately at any time. Therefore, it has been proposed to observe the road with an infrared camera.In such a case, when the temperature fluctuates due to direct sunlight and shadows from the sun, an error occurs during image processing and the identification is normally performed. There is a problem of not being able to understand.

[0003]

SUMMARY OF THE INVENTION Therefore, the present invention solves such a problem and provides necessary road information at any time with accuracy.
Moreover, it is an object of the present invention to provide a device for identifying a person on the road, a vehicle, etc., which can be automatically detected.

[0004]

In order to achieve the above-mentioned object, the present invention provides a pedestrian detecting means, a road surface state detecting means, and road information in a device for identifying a person, a vehicle, etc. on a road. The storage means, a first difference processing means for performing a difference processing between an image from an infrared camera at a certain time and an image when there is nothing on the road , and information t.
Second difference processing means for performing a difference process between the image from the infrared camera after time and the image when there is nothing on the road;
The first and second extracted from the first and second difference processing means
Objects on the road based on the second and second difference information
And the object shape information or temperature
The database with information and the separation means
Shapes or temperatures of separated moving and stationary objects
And an identifying means for identifying the object by comparing with the degree.
It is characterized by having done.

The invention according to claim 2 is characterized in that the temperature information is a temperature peculiar to the object, and the invention according to claim 3 is that the temperature information is the road temperature and the object. It is a characteristic temperature that is a temperature difference from the temperature of.

According to a fourth aspect of the present invention, the frequency distribution of temperature is taken from the image from the infrared camera, and the data of the maximum frequency that can be regarded as the data of the road portion is reflected in the background image by this frequency distribution, and the temperature is reflected on the road With identification means for identifying each object
It is characterized by having done.

[0007]

BEST MODE FOR CARRYING OUT THE INVENTION In the first embodiment of the present invention shown in FIGS. 1 to 3, the temperature can be measured from the blackbody emissivity.
The infrared camera has a sensitivity in the wavelength band of ˜7 μm or 7˜14 μm, a recording unit 2 for recording the image, and an image processing unit for processing the image. Then, the image picked up by the infrared camera is stored by the FIFO (first in first out) method. Then, in the image processing unit, the road and the object on the road can be discriminated by taking a comparison (difference) between the images when there is nothing on the road immediately before the vehicles and people appear on the road. I can. Furthermore, the object can be specified based on the display position and the size of the separated object on the image, depending on whether the object moves on the road or is stationary.

FIG. 1 shows the system and flow of each member of the embodiment that operates as described above. In such an embodiment, the road condition is observed by the infrared camera 1. And the observation result is NTSC (National Tel
evision Standard Cord) converted to video signal.
This video signal is time stamped by a video timer and recorded in a VTR (Video Tape Recorder).

After that, the VTR is reproduced, and the reproduced image is subjected to color signal correction, VTR signal correction, etc. at the time base collector, A / D conversion is performed, and the image is taken into the image processing section. 30Hz in memory
It is recorded in the cycle. A screen in which an object does not exist on the road in this image is registered as a background screen in the recording unit in the processing unit. Then, difference processing is performed between the image input thereafter and the registered image. As a result, what emerges as a difference becomes an object on the road. By repeating such processing, difference images are also accumulated at a cycle of 30 Hz.

In order to further identify this object, a method as shown in FIGS. 2-A to 2-G is adopted by comparing these accumulated images, such as moving objects and obstacles. A stationary object is cut and the object is specified from the temperature information of the object. 2-A to FIG. 2-G, FIG. 2-A is a drawing taken by an infrared camera at a certain time, and FIG. 2-C is a drawing taken by an infrared camera after t hours. B and 2-D are drawings in a state where there is nothing on the road. First, Fig. 2-B and 2-D where there is nothing on the road, and the current time and the time after t hours. 2-D and 2-C are subjected to a difference process, and an object is recognized in each. 2A, 2-C, 2-E, 2-F
In, 2 is a vehicle and 3 is an obstacle or other object.

Next, FIG. 2G shows a moving object in which an object on the screen moves like a vehicle or another stationary object by performing a difference processing between the present drawing of the object and the drawing after t time. The figure shows a state in which the object 3 is cut into pieces, and a stationary object 3 is shown in FIG. As a method of this division, if there is an object that moves using difference processing, the current and t
The object after the time remains, and if there is a stationary object, there is no locus on the drawing. The moving object may be displayed by displaying the difference result as it is, and the locus of the stationary object may be displayed again. Further, the temperature of each of the divided objects is compared with the temperature distribution from a temperature database that is stored in advance to identify what the moving object or the stationary object is.

Since it is known in advance whether the object is a moving object or a stationary object, the efficiency of identifying the object using the temperature database is improved. Further, although it changes depending on the road such as asphalt, which is the center of the background screen, weather, nighttime, temperature change, etc., this change can be absorbed by updating the background screen for continuous observation at intervals as short as possible. Also, it becomes possible to improve the identification rate of objects by creating a database of the temperature difference between the road and the discovered object. Also, by using this temperature difference database, it occurs when the absolute temperature database is used. It is also possible to prevent erroneous recognition when the road temperature and the vehicle or person temperature are close to each other. Further, in order to specify the object, instead of using the temperature database, a database in which the shape of the vehicle is registered may be used.

The second embodiment of the present invention shown in FIG. 4 has a configuration similar to that of the first embodiment in many parts, and as in the first embodiment, an image obtained by observing with an infrared camera is used. Signal is captured by the image processing unit after A / D conversion and stored in the memory. Then, in the image processing unit, the road and the object on the road can be identified by comparing (difference) the images when there is nothing on the road immediately before vehicles and people appear on the road. it can. At the same time, the frequency distribution (histogram) of the temperature of the screen is taken, and the data of the temperature range showing the maximum frequency is reflected in the identification process, whereby the temperature change can be dealt with. The image stored in the memory is an image proportional to temperature, and the temperature frequency distribution is taken from this image. If the range of monitoring is limited to the road portion, the maximum frequency can be regarded as the road. Therefore, by comparing this temperature with other temperatures, the road and other objects can be distinguished. Since this method can be performed for each screen, it is possible to cope with temperature changes and the like.

[0014]

As described above, the present invention compares the extracted images from the infrared camera at a certain time and t seconds later,
Since the object is specified after the object is divided into the moving object and the stationary object, there is an effect that the object specifying ability can be improved.

[Brief description of drawings]

FIG. 1 is a diagram showing a function and a signal flow of each unit of a first embodiment of the present invention.

FIG. 2 is a view showing an obstacle detection concept of the above.

FIG. 3 is a circuit configuration diagram of the above.

FIG. 4 is a diagram showing a function and a signal flow of each unit of the second embodiment of the present invention.

[Explanation of symbols]

1 infrared camera 2 vehicle 3 objects

─────────────────────────────────────────────────── ─── Continuation of the front page (56) Reference JP-A-10-49799 (JP, A) JP-A-7-210795 (JP, A) JP-A-8-317373 (JP, A) JP-A-10- 162287 (JP, A) JP 2002-157583 (JP, A) JP 2002-163644 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) G06T 1/00 330 G08G 1 / 04

Claims (4)

(57) [Claims] 1. An image from an infrared camera at a time,
The first difference processing of the image when there is nothing on the road
The difference processing means, the image from the infrared camera after the time t , and the image when there is nothing on the road are subjected to difference processing.
Second difference processing means and the first and second difference processing hands
On the road from the first and second difference information extracted from the step
A means for separating objects into moving and stationary objects
And a database with object shape information or temperature information
A moving object and a still object that are separated by the separating means
By comparing the shape or temperature of each object,
An automatic identification device for identifying a person, a car, etc. on a road by an infrared camera, which is provided with a specifying means for specifying . 2. The automatic identification device for a person, a car, etc. on a road by the infrared camera according to claim 1, wherein the temperature information is a temperature peculiar to the object. 3. The infrared information of a person, a vehicle, etc. on the road by the infrared camera according to claim 1, wherein the temperature information is an inherent temperature that is a temperature difference between a road temperature and an object temperature. Automatic identification device. 4. A discriminating means for discriminating an object on a road by taking a frequency distribution of temperature from an image from an infrared camera and reflecting the data of the maximum frequency which can be regarded as the data of the road portion in the background image in the frequency distribution. A contract characterized by having
An automatic identification device for a person, a car, etc. on a road by the infrared camera according to claim 1 .