半导体封装图像处理中,环境光照的变化对视频设备截取的图像影响较大。为了减少光照的影响,提出了一种基于投影法和空域同态滤波的多模版算法,该算法首先利用投影法粗略确定芯片的位置并截取包含芯片的图像,接着应用空域变换对截取图像进行同态滤波,削弱了光照变化对图像产生的影响。最后利用模板库中的模板进行传统的模板匹配并更新模板库。实验证明该算法对光照具有很强的自适应性,具有良好的应用前景。
Because of light changing from environment in semiconductor packaging, the images video equipment exists significant light differences. In order to reduce the affect from light differences, a multi- template algorithm based on homomorphic filtering and air space was proposed. First a rough projection method was used to determine the location of the chip, then homomorphic filtering was applied on intercepting images with space transform, thus it weakened the impact of the light change on images. Finally a template from template library was used for traditional template matching and updating the template library with interception image. Experimental results show that the algorithm of illumination has a strong self-adaptability and good application prospects.