论文部分内容阅读
由于观察距离和角度的不同,待识别的目标图像相对模板图像会存在一定程度的角度畸变和比例畸变,大大限制了光学相关模式识别的发展。将最大平均相关高度(MACH)滤波器用于畸变目标识别,通过优化该滤波器的控制参数,并根据多次的计算机仿真实验和光学实验,使该改进型MACH滤波器具有畸变公差高、相关点明亮等特点。用改进后的MACH滤波器对角度畸变目标和比例畸变目标实施频域滤波,能有效增强相关峰强度,扩大畸变目标识别范围。作为实例,给出了复杂背景下识别汽车的计算机仿真实验和光学实验,有效验证了该算法的可行性。
Due to the different viewing distance and angle, there is a certain degree of angular distortion and proportional distortion of the target image to be recognized relative to the template image, greatly limiting the development of optical correlation pattern recognition. The maximum average correlation height (MACH) filter is used for distortion target identification. By optimizing the control parameters of the filter, and based on multiple computer simulation experiments and optical experiments, the improved MACH filter has high distortion tolerance and correlation point Bright and so on. The improved MACH filter is used to implement the frequency-domain filtering on the angle-distortion target and the proportional-distortion target, which can effectively enhance the correlation peak intensity and enlarge the range of distortion target identification. As an example, a computer simulation experiment and an optical experiment for recognizing a car with a complex background are given, and the feasibility of the algorithm is validated.