KR100222545B1 - 3d distance estimation equipment and method using multi-image - Google Patents

Abstract

The present invention relates to an apparatus and method for estimating a three-dimensional distance using multiple images, the apparatus comprising: a first camera, a second camera, a third camera, a fourth camera for detecting an object by receiving light input from the outside and estimating the distance of the object; A camera unit comprising four cameras mounted at log intervals; A switch unit serving as a switch to selectively receive an image from the camera unit; An image capture unit which receives an image of a camera connected by the switch unit and converts an analog signal to digital; A memory unit for storing digital information of an image output from the capture unit; A control unit for outputting a control signal relating to a camera selection position of the switch unit and processing image data stored in a memory unit to measure the conversion of an image or a real distance from a subject, and to output information regarding an image and a distance to the outside; The three-dimensional distance estimation apparatus and method using multiple images with the effect of making accurate matching using four cameras and estimating the actual distance between the subject and the camera input to the image with less error can be precisely controlled during unmanned driving of the vehicle. to provide.

Description

3D distance estimation apparatus and method using multiple images

The present invention relates to an apparatus and method for estimating a 3D distance using multiple images. More specifically, since distance estimation using a dual image has a large real error, an image close to the image using a multiple image has an error in matching. Easy and far-field image relates to a three-dimensional distance estimation apparatus and method using multiple images to accurately estimate the actual distance of the object.

Manufacturing a vehicle requires technologies in various fields, and generally, it can be said to be a technique for integrating technologies in general science fields such as machinery, electronics, and chemistry.

In recent years, as a result of the development of automobiles, not only design and mechanical operations but also the technology of the internal electric system has emerged.

As the demand for automobiles gradually increases, the demands of consumers are also diversified, which is accelerating the development of devices other than the original functions of automobiles.

In addition, the development of safety devices and the like in driving a car is also important and many products are being produced.

For example, there is an air bag, a seat belt, etc. in consideration of the driver's safety in a car crash.

In addition, much research has been conducted on accessory devices for providing convenience to the driver such as an accessory device of a vehicle, an interior light, and an instrument panel.

While driving, the instrument panel was marked with the driver's speed, engine rotation speed, fuel, engine temperature and audio.

In addition, in the related art, night driving is provided to the driver so that the driver can check the contents of the instrument panel by turning on a light at the rear of the instrument panel.

In recent years, new maps, such as electronic map devices, have been provided for novice drivers, to provide drivers with information on the shortest roads to their destinations. The trend is spurring.

As shown in FIG. 1 (a) and FIG. 1 (b), a method of estimating a distance of an object of a stereo technique using a conventional dual image will be described.

In the first camera, an object located in front of the object is identified and detected in the image 1, and the second camera identifies an object in front of the object and detected in the image 2.

The focal length of the camera is λ, the distance between the focal points of the first and second cameras is B, the coordinates of an object detected in the image 1 are x1, y1, and the coordinates of the object detected in the image 2 are x2, y2. In this case, the distance Z between the object and the camera is a proportional expression of Equation 1).

In order to complete the above equation, it is necessary to confirm that the object detected in the two images is an object.

Referring to the drawings shown in FIG. 2, matching between objects detected by two cameras is as follows.

The size of the screen is 512x512, which detects objects in a block of 8x8 size, which is 8 spaces up and down from the first (0,0).

First, in order to check an object caught by the first camera, a block of 8 × 8 size is searched from left to right.

The image size of the object retrieved by the search is detected.

It checks the object size caught by the second camera like the first camera, detects the image size of the searched object and compares it with the size detected by the first camera. It is to let.

However, when the two cameras are far apart, the matching of objects is not easy.

2 (b) is as follows when the actual distances of the objects retrieved from the two cameras are calculated.

Therefore, the relationship between the object matching and the distance error rate for the distance between the two cameras is shown in Table 1).

However, the distance estimation using the dual image is difficult to match due to the fluidity of the object, and the distance is estimated when the image interval is narrowed for easy object matching. There is a problem that can be reduced but difficult to match.

Accordingly, an object of the present invention is to solve the above problems, and to facilitate matching between adjacent images using a multi-image method, and to enable a far-field image to be measured by reducing an error of a real object of a matched object. An object and method for estimating a three-dimensional distance using an image is provided.

As a means for achieving the above object, the configuration of the present invention,

A camera unit comprising four cameras mounted at logarithmic intervals of a first camera, a second camera, a third camera, and a fourth camera that detect an object by receiving light input from the outside and estimate the distance of the object; ;

A switch unit serving as a switch to selectively receive an image from the camera unit;

An image capture unit which receives an image of a camera connected by the switch unit and converts an analog signal to digital;

A memory unit for storing digital information of an image output from the capture unit;

The control unit outputs a control signal relating to a camera selection position of the switch unit, and processes image data stored in a memory unit to convert an image or measure a real distance with a subject to output information about an image and a distance to the outside.

As a means for achieving the above object, the characteristics of the present invention,

Detecting an object input to each camera;

Matching the object detected by the first camera with the object detected by the second camera at the closest distance;

Matching an object detected by the first and second cameras with an object detected by a third camera;

Estimating the position of the object detected by the fourth camera based on the position of the object detected by the first camera, the second camera, and the third camera;

Matching an object detected by the first camera and the fourth camera;

Computing the distance between the object and the camera using the conditions and the like in the first and fourth cameras.

1 is a distance measuring method using a conventional dual image,

(a) is a distance estimation position diagram of an object seen from a conventional side,

(b) is a distance estimation position diagram of an object seen in a conventional plane.

2 is a method of detecting an object of a general camera

(a) is a scanning diagram showing a method of detecting an object using a general group,

(b) is a diagram illustrating a distance detection method of an object in which two general cameras are detected.

3 is a view showing images seen by four cameras arranged at logarithmic intervals according to an embodiment of the present invention.

4 is a view showing an image displayed in FIG. 3 according to an embodiment of the present invention.

(a) is a view of vertically arranging the images shown in FIG. 3 according to an embodiment of the present invention,

(b) is a view showing an image of the center of the object shown in FIG. 3 according to an embodiment of the present invention.

5 is a control block diagram of a 3D distance estimation apparatus using multiple images according to an embodiment of the present invention.

6 is a control flowchart of a 3D distance estimation apparatus and method using multiple images according to an embodiment of the present invention.

Best Mode for Carrying Out the Invention The present invention will be described in detail with reference to the accompanying drawings.

The apparatus and method for estimating a 3D distance using multiple images according to the configuration of FIGS. 5 and 6 and the configuration described above are as follows.

Camera unit 100 consisting of four cameras mounted at a logarithmic interval with a first camera, a second camera, a third camera, and a fourth camera for detecting an object by receiving light input from the outside and estimating the distance of the object; ; A switch unit 200 serving as a switch to selectively receive an image from the camera unit 100; An image capture unit 300 which receives an image of a camera connected by the switch unit 200 and changes an analog signal to digital; A memory unit 400 for storing digital information of an image output from the capture unit 300; Outputs a control signal excessive to the camera selection position of the switch unit 200 and processes the image data stored in the memory unit 400 to convert the image or to measure the actual distance with the subject to output information about the image and distance to the outside It consists of a control unit 500.

The camera unit 100,

A first camera 10 and a second camera 20 for easily matching the detected object; A second camera 20 and a third camera 30 matching the object detected by the first camera 10 to estimate the position of the object detected by the fourth camera 40; The fourth camera 40 serves to reduce the error of estimating the distance by matching the object detected by the first camera 10.

The control unit outputs a selection signal for selecting a camera using a switch unit to receive and process an image output from the fourth camera from the first camera.

The switch connects the selected camera to the image capture unit by the selection signal of the controller.

The selected camera receives an external light source and outputs an electric analog video signal according to the light source, and the image capture unit receives the video signal and converts the digital signal into a data size of a manufactured circuit state.

The image signal stores information input from each camera in the memory unit, and the controller performs processing such as reconstructing the image information or classifying subjects to obtain necessary information by reading the image information stored in the memory unit. Information is output from the image.

In addition, the memory unit is composed of a static RAM (dynamic RAM) and a dynamic RAM (dynamic RAM) to quickly store the contents processed in the image capture unit and re-store the stored contents in a large capacity dynamic RAM for real-time processing Keep it close.

The method of obtaining the distance between a subject and a camera by the above structure,

The control unit outputs a selection signal to the switch unit, receives an external image from each camera, converts the image capture unit into a digital signal, and detects a subject from the image information (S20).

The controller first detects the same subject by matching the object detected by the first camera with the object detected by the second camera closest to the contents stored in the memory unit as shown in FIG. 2 (a) (S30). .

The subject is detected by matching the object detected by the first and second cameras with the object detected by the third camera as in step S30 (S40).

The controller estimates the position of the object detected by the fourth camera as shown in FIG. 4 (a) and FIG. 4 (b) based on the position of the subject matched by the first camera, the second camera and the third camera (S50). .

Finally, the controller matches the object detected by the first camera and the fourth camera using the image information stored in the memory (S60).

The controller finds the center coordinates of the subject input from the first camera and the subject input from the fourth camera, and uses a given stereo technique such as the conventional first (a) and first (b) diagrams using given conditions. Calculate the distance between the subject and the camera detected by the camera (S70).

In other words, the distance is obtained by using Equation 3).

As described above, in the embodiment of the present invention, multiple cameras have an effect of enabling accurate control during unmanned driving of a vehicle by performing accurate matching using four cameras and estimating the actual distance between the subject and the camera inputted in the image so that the error is small. Provided are an apparatus and method for estimating 3D distance using an image.

Claims (2)

A camera unit comprising four cameras mounted at logarithmic intervals of a first camera, a second camera, a third camera, and a fourth camera that detect an object by receiving light input from the outside and estimate the distance of the object; ; A switch unit serving as a switch to selectively receive an image from the camera unit; An image capture unit which receives an image of a camera connected by the switch unit and converts an analog signal to digital; A memory unit for storing digital information of an image output from the capture unit; And a control unit for outputting a control signal relating to a camera selection position of the switch unit and processing image data stored in a memory unit to measure the conversion of an image or a real distance from a subject, and to output information about an image and a distance to the outside. 3D distance estimation apparatus using multiple images. Detecting an object input to each camera; Matching the object detected by the first camera with the object detected by the second camera at the closest distance; Matching an object detected by the first and second cameras with an object detected by a third camera; Estimating the position of the object detected by the fourth camera based on the position of the object detected by the first camera, the second camera, and the third camera; Matching an object detected by the first camera and the fourth camera; Computing the distance between the object and the camera using a given condition in the first camera and the fourth camera 3D distance estimation method using multiple images.