摘要
针对机器视觉领域日益增长的对嵌入式、实时、健壮图像特征算法的需求,以TI DM642数字图像处理器为平台移植了OpenSURF库,并针对TI DM642处理器的特点对其进行了优化。实验表明,在DM642平台上实现的SURF算法能够满足实时处理的要求。同时,可以方便地向具有类似性能的器件移植。
In order to satisfy the increasing need for embedded,real time and robust image feature extraction algorithms in the field of machine vision,OpenSURF is ported to TI DM642 digital signal processor,and optimized based on the characteristics of the device.Test results have confirmed that the ported version of OpenSURF has the capability of processing video in realtime.Meanwhile,this implementation is highly portable.
出处
《电子科技》
2012年第5期133-136,共4页
Electronic Science and Technology
