tailieunhanh - Machine Vision - David Vernon Part 9

Tham khảo tài liệu 'machine vision - david vernon part 9', kỹ thuật - công nghệ, cơ khí - chế tạo máy phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | Robot programming and robot vision used. In order to facilitate simple image analysis we threshold the image and thin the resultant segmented binary image. The original image is acquired at the conventional 512 X 512 pixel resolution with eight-bit grey-scale resolution see Figure . To ensure fast processing and analysis we first reduce the resolution to 128 X 128 pixels see Figure resulting in a reduction of the complexity of subsequent operations by a factor of sixteen. This reduction is important as the computational complexity of thinning operations is significant. There are essentially two ways in which this reduced resolution image may be generated by sub-sampling the original image every fourth column and every fourth line or by evaluating the average of pixel values in a 4 X 4 window. As Figure 512 X 512 image. Figure 128 X 128 image. 190 Two vision algorithms for identifying ends of wires we saw in Chapter 4 it is desirable to reduce the image noise and since this can be accomplished by local averaging the reduced resolution image in this implementation is generated by evaluating the local average of a 4x4 nonoverlapping region in the 512 X 512 image. This also minimizes the degradation in image quality referred to as aliasing which would result from sub-sampling. The image is then segmented by thresholding in the manner discussed in Chapter 4 the threshold value is automatically selected using the approach associated with the Marr-Hildreth edge detector see Section . Figure shows the binary image generated by thresholding the original grey-scale image at the automatically determined threshold. Once the binary image has been generated the next step is to model the wires in some simple manner. The skeleton is a suitable representation of electrical wires objects which display obvious axial symmetry. In this instance we use the thinning technique described in Section Figure illustrates the application of this .