Simulation of the Laser Stripe Reflection in Sub-Pixel Estimation with Autoconvolution Algorithm

The accuracy of the profile scanners depends, among other factors, mostly on algorithms for image processing of the reflected laser stripe and its sub-pixel position. Different signal processing techniques can be easily tested and analyzed with correctly simulated physical effects of the system. This study explains a method that simulates laser stripe reflection in a dynamic environment, especially in the case when the scanner images are saturated with the flattened top of the reflection. The expected shape of the laser stripe reflection follows the Gaussian profile which can easily be distorted by ambient light, improper camera exposure time, or irradiated laser power. In this case, some of the pre-processing techniques, such as autoconvolution, can be used for mitigating those effects. Implemented as a multiplierless filter, this paper analyzes the usage of autoconvolution for filtering techniques on Field Programmable Gate Array (FPGA) through quantization of the pixel word length.