| 摘要 |
A semiautomatic method for digital feature separation uses a trained sample selected by an operator using a software "eye dropper" tool or a similar region-of-interest tool to sample features of interest on a stored digital image of, for example, an electronic component such as pads, bumps or leads of the component. Pixels from the sampled features are analyzed and plotted based on their color hue and saturation values or gray scale intensity values. The range of such values is chosen by a user. A second digital image is then compared to the sampled feature data gathered by the "eye dropper" tool. If the color and intensity values of the pixels from the second digital image fall within a user defined acceptable absolute value range, then the locations and values of those pixels in the image are saved. Background or surrounding elements in the second digital image which do not fall within the acceptable value range are deleted and replaced with a code that makes these background or surrounding element locations appear as a graphic transparency. This process leaves only the features of interest in their original locations saved in the graphics memory surrounded by pixels that have been made video transparent. The resulting saved image which contains only the selected features is then laid over a live image such as that of a target substrate with corresponding features of interest. The operator then manually aligns the saved features from the component image over the corresponding features of the live image of the substrate to achieve component to substrate registration. |
