Based on multiple fields of view (FOV) point spread function (PSF) estimation, we propose a novel gradient-constrained image restoration method to solve optical degradation in microscopic imaging. The whole FOV is segmented into several parts. The modulation transfer function (MTF) is measured to obtain the corresponding PSF for each part. L0 gradient constraint is treated as a regularization term, a fast image restoration method is designed to deblur degraded images of each field of view. Finally, gradual weight approach is used to stitch the multiple field of view (m-FOV) restoration images. Several microscopic images are tested and evaluated. Comparing with other methods, the results indicate that our method performs better, and runs fastest of all.
In this paper we propose a projective LED low-beam headlamp of motorbikes. An ellipsoidal reflector is used to collect
light emitting from the LED source. A baffle plate is located in the focal point of the ellipsoidal reflector to form the
cut-off line.Then the light is redistributed by an optical lens. We set the measuring screen as an semi-circular zone and
divide it into many small lattices, divide the spatial angle of the LED source into many parts and make relationships
between them. According to the conservation of energy and Snell law, the lens is generated by freeform optics design
method. Then the optical system is simulated by Monte Carlo method using ASAP software. Light pattern of simulation
could meet the national standard. The low-beam headlamp is finally fabricated and assembled into a physical object.
Experiment results can fully comply with regulations GB5948-1998.