Study On High Dynamic Range Image Capture And Adaptive Method For AOI System

Automatic optical inspection(AOI) is a technology based on machine vision, which includes the knowledge of opitcs, electronic and automatic. As the development of the suface mount technology, the AOI system is widely used in PCB inspection. As the main source of AOI system, the image captured decides the accuracy of system. But during the PCB inspection, the local oversaturation, airing from the non-consistent reflectivity of the PCB surface, will lead to error result. To avoid the local oversaturation and improve the quality of image, a high dynamic range image captured system is built based on the Liquid Crystal on Silicon(LCoS). There are mainly two form images captured by the AOI system, so the corresponding image adptive methods are proposed using the high dynamic range image captured system built.The major content and innovations in this paper are shown as follow:(1) A high dynamic range image captured method is proposed for AOI system, and the high dynamic range image captured is built using the LCo S and the CCD camera.(2) The calibration method for system modulate curve is studied, and the relation between the pixel coordinate of LCoS control mask and the the pixel coordinate of image captured is calibrated using Space Mapping Theory. The simplified modulation curve model of the AOI high dynamic acquisition system is established, and a calibrated procedure is proposed based on the statistical analysis theory which is simple to realize. The Space Mapping model of the pixel coordinate of LCo S control mask and the the pixel coordinate of image captured is built, and the experiment result proved that the error was litter than 0.25 pixel.(3) A new high dynamic range image get method is proposed, which is based on the image fusion technology. The image reconstructing method is proposed according to the system modulation curve. To avoid the noise in the reconstructed image, the image fusion method is used to fuse the reconstructed data to one high dynamic range image, and the weight function is proposed based on statistical analysis. The experiment result proved that the fused high dynamic range image not only has high quality in lower brightness area, also has a high dynamic message oin saturated area.(4) The image adaptive method using the proportion integration differentiation(PID) controller and the laser light image adaptive method using mapping function is studied. The PID controller is designed, the parameter and the stability condition is determined according to the iterators and image quality. The saving time method is proposed whoes feedback procedure is simulated by the computer, and the LCo S mask at the stability transform to the LCoS chip to capture image. The experiment result proved that the image local oversaturation is avoided using this method, and the image entropy and image mean grads of the image local oversaturation area is reduced. The laser image adptive method is proposed which uses a mapping function generate an LCoS mask by the fusion high dynamic range image. The experiment result proved that the method proposed in this paper could accomplish the aim of getting real laser light from the scattering without affecting the system accuracy, and the modulated image’s cross section is more close to the g aussian distribution.