Apparatus and method for vision processing on network based intelligent service robot system and the system using the same

1. An apparatus for vision processing on a network based intelligent service robot comprising:

a stereo camera unit acquiring images from two cameras of left and right cameras incorporated therein;

an input image pre-processing unit for calibrating characteristic differences between the left and right cameras with respect to the left and right images inputted through the stereo camera unit;

a stereo matching unit for performing a stereo matching by computing a disparity map through the stereo matching over the left and right images inputted through the input image pre-processing unit;

an image post-processing unit for extracting a depth map for the image outputted from the stereo matching unit, and performing a segmentation for segmenting different objects from the extracted depth map; and

an image output selection unit for selectively outputting the output image of the respective units.

2. The apparatus of claim 1 , wherein the input image pre-processing unit uses camera 20 calibration, scale-down of the image, rectification of the image, brightness control of the image, or the like, in order that the stereo matching unit may perform the stereo matching over the image inputted from the stereo camera unit with ease.

3. The apparatus of claim 1 or 2 , wherein the image output selection unit, which selects the output image of the vision processing unit of the intelligent service robot, selects and outputs a predetermined image selected from the group consisting of the image acquired from the left and right camera, the resultant image of the input image pre-processing unit, the resultant image of the stereo matching unit, and the resultant image of the image post-processing unit, according to applications of the intelligent service robot terminal.

4. The apparatus of claim 3 , wherein the image post-processing unit segments a target to track and an obstacle on the basis of a three-dimensional image information computed through the stereo matching unit, and uses the horizontal, vertical lengths and the distance of the segmented object as information for the robot's traveling and moving.

5. The apparatus of claim 1 , wherein the image outputted from the stereo matching unit is represented such that distance information is represented as a bright color for a close object and as a light color for a distant object.

6. A method for vision processing on a network based intelligent service robot, the method comprising:

acquiring an input image of at least one of an object to track and an obstacle from two cameras comprising a left camera and a right camera;

performing an input image pre-processing to calibrate the left and right image signals acquired from the left and right cameras to have the at least one of the object to track and the obstacle, using at least one or more image processing methods;

performing a stereo matching, at a stereo matching unit, by computing a disparity map after finding out points corresponding to each other in the left and right images calibrated in an input image pre-processing unit;

extracting a depth map by calculating a depth based on the disparity map computed at the stereo matching unit, and performing an image post-processing for segmenting different objects by the extracted depth map; and

selectively extracting the horizontal length, the vertical length, and the distance of the segmented object from at least one of the acquired input image, the performed input image pre-processing, the performed stereo matching, and the extracted depth map by calculating the depth based on the disparity map computed at the stereo matching unit and the performed image post-processing for segmenting the different objects by the extracted depth map of a robot terminal, and

controlling the robot to travel and move by at least one of segmenting and recognizing the object to track and the obstacle based on the selected extraction result.

7. The method of claim 6 , wherein the controlling of the robot is performed by the robot terminal information provided by a robot server.

8. A system using an apparatus for vision processing on a network based intelligent service robot, comprising:

a robot terminal including a vision processing unit of a robot having the apparatus claimed in claim 1 for acquiring an external image, a robot control unit for controlling a whole robot terminal, a robot server communication unit for serving a role of communicating with a robot server, and a robot sensor/drive unit for sensing an external circumstance and driving the robot; and

the robot server interfaced with the robot terminal over network, for controlling the robot under predetermined circumstances.