Optical microscopy is an important technique in petrological and biological. However, because of the limited
focus depth of a traditional optical microscope, 3D objects can't be observed with all its information in one
scene. The commonly used method to solve this problem is to use a digital microscope to collect a sequence
of multi-focus images with a constant change of the focal length and then get the final sharp image by fusing
the images. Wavelet pyramid algorithm is one of the commonly used fusion algorithms and has a good
performance. In this paper, a novel 3D reconstruction algorithm is presented based on the wavelet pyramid
image fusion algorithm. First, a sequence of multi-focus images is collected with a digital microscope and
then the images are fused with the wavelet pyramid fusion algorithm. After that, the deviation between the
fused image and each of the original multi-focus images are calculated pixel by pixel. For convenience, the
absolute value of the deviation is calculated. The smaller the deviation, the closer the corresponding pixel to
the focal plane and vice versa. Thus a 3D deviation matrix is obtained. For each pixel position (x, y), the z
index of the smallest deviation value in the third direction of the deviation matrix is considered the pixel's
height index. A complete height index map is obtained by selecting the indices of these smallest deviation
values. In order to eliminate the noise effect, a median filter is applied to the height index map. The height
index map times the height factor (the unit step length along the optical axis when collecting the multi-focus
images) is the object's actual height map. Finally, the surface of the 3D object is reconstructed with the
object's height map. In order to test the reconstruction algorithm, a dedicated high resolution Complementary
Metal Oxide Semiconductor (CMOS) digital microscope is designed and fabricated to collect the multi-focus
images. Experimental results show that the proposed algorithm produces a nice looking surface of a 3D
object.
Multi-core parallel computing is spreading in most industries and the imaging and machine vision industry is also taking
the advantage of this technology. The utilization of parallel computing will increase the throughputs and reduce response
times of the imaging system, especially for the high resolution CCD/CMOS based imaging system. Multi-core image
processing fully utilizes the ability of the CPU's parallel computing, for multiple cores share the processing task of an
imaging system. The parallel computing automatically detects the number of CPUs or the number of the CPU cores and
then automatically splits the image into the according number of logical blocks, which will be then passed on to the
processing threads separately. After all the processing threads finishes, the result will be synthesized. For high resolution
CCD/CMOS based digital microscope autofocus imaging system, the speed of measuring the sharpness of the current
collected image greatly affects the speed of the autofocus process. The real-time requirement of the system needs fewer
time cost for image sharpness evaluation and the multi-core parallel computing is applied in the algorithm to meet this
requirement. The proposed algorithm is as follows. First, the current collected image is divided into several logical
blocks; second, for each block, a worker thread will compute the sharpness of this block; finally, after all the worker
threads finishes, the sharpness will be summed for comparison with the next collected image. In order to test the
efficiency of the algorithm, a dedicated high resolution CCD/CMOS based digital microscope autofocus imaging system
is designed and implemented and several image sharpness evaluation algorithms are used, as well as the self-adaptive
mountain-climbing search (SAMCS) method for the searching method. The numeric simulation and the experimental
results show that the proposed algorithm greatly improves the speed of the autofocus process.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.