Abstract:
An image data processing device capable of thinning video data is provided to shorten the thinning processing time and to drive physical movement by analyzing thinned image data. A camera unit(120) generates original image data, and an image processor(130) creates thinned image data by thinning the original image data according to predetermined method. A controller(110) creates at least one kind of path information by analyzing the thinned image data, and outputs a driving initiation command by selecting one kind of path information among the plural kinds of the path information. A driving unit(150) initiates an operation by responding to the driving initiation command, and a physical operation is performed according to the operation of the driving unit.
Abstract:
An apparatus for detecting connected components in real time and a method therefor are provided to input and output images independently without using a universal processor. An apparatus for detecting connected components comprises the followings: a binarization portion which produces a binary image by binarizing an inputted original image according to a preset threshold value; the first detector(120) which receives the binary image and generates an index for a connected component; a representative value calculating unit(130) which produces a representative value list of representative values corresponding to each index; the second detector(150) which receives the index and representative value list and designates representative values of each connected component; and a re-arrangement unit(160) changing representative values of each connected component into successive values.
Abstract:
PURPOSE: A sensor data real-time morphology processing device and a method and a system of drawing a real-time environment map for an intelligent mobile robot using the same are provided to draw an environmental map for an intelligent robot by adopting morphology technique while parallel data processing. CONSTITUTION: A sensor data real-time morphology processing device and a method and a system of drawing a real-time environment map for an intelligent mobile robot using the same is comprised of the steps: determining the validity of data measured from a sensor adhered to an intelligent mobile robot(S31); Storing sensed data which is determined as valid into a memory unit(S32); performing binarization by using a threshold value of the stored sensor data(S33); performing an dilation calculation and erosion calculation of the sensed data which is processed as binarization(S34); and strong the result of the dilation calculation and erosion calculation into the memory unit(S35).
Abstract:
PURPOSE: An image processing apparatus and method for detecting motion information in real time are provided to be used in an intelligent camera, an industrial field easily by being embedded through simple hardware structure. CONSTITUTION: An image processing apparatus(100) for detecting motion information includes a sub-sampling unit(120), a sensor's conversion unit(140), a correlation calculation unit(150) and a motion detecting unit(160). The sub-sampling unit performs a sub-sampling for each image by receiving plural previous images and current images, and a sensor's conversion unit obtains a sensor's vector by performing conversion for each image sub-sampled. The correlation calculation unit calculates and compares the correlation values between the current and previous images, and detects a region that has high correlation. The motion detection unit detects the motion information within an image by detecting the location of a pixel corresponding to the region that has the high correlation.
Abstract:
A stereo image matching error removal apparatus and a removal method using the same are provided to generate a window in a stereoscopic image and determine whether a corresponding center pixel is outputted according to a difference of gray levels which are pixel values between a center pixel and an adjacent pixel of the window without changing the center pixel. An image processing unit(10) generates a window from a stereoscopic image including matching error. A difference value calculation unit(20) calculates a difference value between a pixel value of a center pixel and a pixel value of an adjacent pixel in the window. A comparison unit(30) compares a comparison reference value with the difference value. The comparison unit divides an adjacent pixel into the first adjacent pixel and the second adjacent pixel based on the comparison result. A summation unit(40) counts the number of the first adjacent pixels in the window.
Abstract:
A real-time object tracking method and an apparatus thereof using a color histogram capable of continuously tracing the change of an object which is chosen on a track station are provided to modify a reference histogram based on the present histogram of the window which is chosen on the track station. A histogram generator(30) produces a plurality of windows from a successively inputted image. A similarity calculator(40) produces the similarity with the set standards histogram by each present histogram. A tracking target selector(60) extracts a standards histogram and the most high present histogram among the current histograms. The window corresponding to the extracted present histogram is selected as a tracking window. A modifier(70) modifies the standard histogram by reflecting the standard histogram.
Abstract:
An image processing device is provided to reduce a thinning process time. Binarization data are divided into predetermined image blocks and stored in an input storing unit(120). A thinning unit(140) parallel thins the divided image blocks. An input switch unit(130) connects the input storing unit and the thinning unit. An output storing unit(160) stored the thinned image blocks. An output switch unit(150) connects the thinning unit and the output storing unit. An image data reconfiguration unit(170) combines the thinned image blocks stored in the output storing unit and reconfigures the blocks as thinned image data.
Abstract:
A chain code generating apparatus and method are provided to generate the chain code for an interest decoding target from an input image in real-time only with a small hardware device, thereby reducing the overload of a total image processing system and maximizing the efficiency of the total image processing system. A chain code generating apparatus includes an image input unit(110), an image storage unit(120), a code generation unit(130), a code storage unit(140), and a transmission unit(150). The image input unit receives image signals from a camera(200), converts the received image signals into digital image signals, and divides and outputs sync signals from the image signals. The image storage unit stores image data corresponding to an effective image section by frame out of the digital image signals on the basis of the sync signals. The code generation unit reads the image data stored in the image storage unit to perform the outline searching of an interest decoding area, and generates the chain code in correspondence to location correlation between adjacent pixels forming the outline of the interest decoding area according to a searching result. The image storage unit includes a first buffer memory and a second buffer memory, and stores the image data corresponding to the effective image section alternately in the first buffer memory and the second buffer memory whenever a frame is changed. When the image data is stored in one of the first buffer memory and the second buffer memory, the code generation unit reads image data from the other of the first buffer memory and the second buffer memory to generate the chain code.
Abstract:
PURPOSE: An image processing device and a method for comparing an object shape in real-time are provided to reduce the error in shape comparison depending on the size and arrangement angle of an object by judging the shape similarity. CONSTITUTION: A binary image generator(120) generates binary image data, and a subject connection component detector(130) analyzes the binary image data. The subject connection component detector displays the pixel value of an image in arrangement. A subject property decision unit(140) detects the characteristic information depending on the size and angle. A subject shape acquisition unit converts the component value arrangement. The component value arrange is mapped to an object window. The subject shape comparison unit(180) compares the component value arrangement with the reference component arrangement.