Study On Digital Observation Of Collimator Based On CCD

As a optical instrument which can produce level, collimator is widely used inindustrial measurement and scientific research. But subjective factors will influencethe accuracy and human eyes will be tired because of long-time small fieldobservation. Image measurement technique, based on CCD, is widely used in manydetection areas. This thesis introduces a digital observation method based on CCD.This method can improve the accuracy and ease labor intensity.First, the digital observation is introduced in detail. Some problems about CCDpicking, lens picking, light source picking and mechanical connections are studied.The system debugging is also studied.Second, there are an introduce and a contrast about three sub-pixel divisionalgorithm. Curve fitting based on least squares is selected as the core algorithm in thisthesis. According to ideal edge model, cubic curve fitting is used to fit the reticle edge.The implement of cubic curve fitting is introduced in detail. The contrast between thedivision and the no-division is carried on. A edge point excluding tactic which isbased on the shape feature of gray gradient is proposed to improve the fittingaccuracy.Third, image filter, image section, rough edge extraction, fine edge extraction,distance measure and target sighting are explained in detail. A study on thresholdselecting, section location selecting, section number selecting, excluding gross errorand reducing random error is carried on. According to the above study, there is asoftware system which is completed on VC++6.0platform.Finally, From the CCD acquisition, software and axis consistency, system error isanalyzed. Although the analysis is carried on idea condition, it is worth reference.The result shows digital observation is more precise than traditional human eyeobservation in the distance measure and target sighting.