LED light design method for high contrast and uniform illumination imaging in machine vision.

In machine vision, illumination is very critical to determine the complexity of the inspection algorithms. Proper lights can obtain clear and sharp images with the highest contrast and low noise between the interested object and the background, which is conducive to the target being located, measured, or inspected. Contrary to the empirically based trial-and-error convention to select the off-the-shelf LED light in machine vision, an optimization algorithm for LED light design is proposed in this paper. It is composed of the contrast optimization modeling and the uniform illumination technology for non-normal incidence (UINI). The contrast optimization model is built based on the surface reflection characteristics, e.g., the roughness, the reflective index, and light direction, etc., to maximize the contrast between the features of interest and the background. The UINI can keep the uniformity of the optimized lighting by the contrast optimization model. The simulation and experimental results demonstrate that the optimization algorithm is effective and suitable to produce images with the highest contrast and uniformity, which is very inspirational to the design of LED illumination systems in machine vision.