论文部分内容阅读
提出了一种粗搜索结合精匹配定位的两级图像相关匹配算法 ,该法利用目标中心信息的稳定性 ,并结合图像序列间的相关不变性和目标的自相关性给出了一种新的置信度评估方法 ,在存在复杂背景、类似目标、较强噪声和一定旋转及尺寸变化的情况下仍有较高的匹配精度和跟踪稳定性 ,匹配速度也很快 .还描述了采用 CPL D芯片与 DSP处理器共同实现算法的方法 .
A two-level image correlation matching algorithm based on rough search and fine matching positioning is proposed. This method uses the stability of the target center information and gives a new method based on the correlation invariance between the image sequences and the target autocorrelation Confidence evaluation method, in the presence of complex background, similar targets, strong noise and a certain rotation and size changes still have higher matching accuracy and tracking stability, matching speed is also described using CPL D chip Method of implementing algorithm together with DSP processor.