Page 411 - Mechatronics with Experiments
P. 411

October 9, 2014 8:1
                                         Printer: Yet to Come
                        JWST499-Cetinkunt
            JWST499-c06
                                                                                        SENSORS  397  254mm×178mm
                                  are set to one fixed low value, and all pixel values above that value are set to a high
                                  value (i.e., black value). This turns the image into a binary image. Various detection
                                  algorithms can be run quickly on such an image, such as the edge detection algorith.
                                2. Edge detection of an object: an edge is detected when a sharp change occurs from
                                  one pixel to another in the gray scale value of the image. Once such a transition is
                                  detected between two pixels in a search direction on the image array, then all the
                                  pixels around the transition pixel are searched to determine the edge boundary.
                                3. Color or gray scale distribution (also called the histogram of image): this is a plot of
                                  the gray scale distribution of the image, that is how many pixels (y-axis) has a given
                                  gray scale (x-axis).
                                4. Connectivity of the object (detect discontinuities).
                                5. Image comparison with a reference image in memory (also called template matching),
                                  that is image system may store a set of “good part” images. A real-time image is
                                  compared to the stored images to determine whether it matches one of the template
                                  images or not.
                                6. Position and orientation of an object relative to another reference.
                                7. Dimensions (length, area) of an image: once the boundaries of a part or parts in an
                                  image are detemined, the dimensions and area information can be easily calculated.
                                8. Character recognition, that is recognizing letters and numerals.
                                9. Geometric image transformations, that is mathematical operations on the matrix data
                                  of the image to move, rotate, stretch the image.

                             Example     Consider a vision system with 1024 × 1024 pixel resolution. The camera
                             processes 60 frames per second. The resolution of the the A/D converter system is 8-bit.
                             What is the amount of data (bytes) processed per second? If the camera is focused into an
                             surface with 10 cm × 10 cm dimensions, what is the spatial resolution in measurement?
                                  Each pixel holds one byte of data since the ADC converter has 8-bit resolution. The
                             number of bytes processed per second is equal to the number of data bytes per frame times
                             the number of frames per second,

                                                        N = 1024 × 1024 × 60                   (6.190)
                                                          = 62 914 560 bytes∕s                 (6.191)
                                                          ≈ 63 MB∕s                            (6.192)

                             Clearly, the amount of data processed per second is very large in a high resolution camera.
                             The smallest distances in x and y directions the system can measure are
                                                      10 cm
                                                Δx =        = 0.00976 cm = 0.0976 mm           (6.193)
                                                   f
                                                       1024
                                                      10 cm
                                                Δy =        = 0.00976 cm = 0.0976 mm           (6.194)
                                                   f
                                                       1024
                             which indicates that the vision system can measure dimensions in x and y coordinates with
                             about 0.1 mm accuracy.


                      6.13 GPS: GLOBAL POSITIONING SYSTEM

                             The global positioning system (GPS) is a location measurement system on Earth, based
                             on signals received from a group of satellites (about 24 satellites) orbiting the Earth at an
   406   407   408   409   410   411   412   413   414   415   416