论文部分内容阅读
目的针对惯性约束核聚变实验中靶图像轮廓模糊、亮度不均匀等问题,从提高图像处理实时性角度出发,提出一种高可靠性和高精度的快速椭圆检测方法。方法首先利用椭圆边缘点在它与圆心相连方向上具有较大灰度变化率这一特点,以预估中心点为极点建立极坐标系,通过从极点出发的射线上灰度变化率极值点搜索实现椭圆边缘点检测,极值点搜索在图像局部范围进行保证边缘点检测的有效性和实时性;其次利用基于RANSAC的自适应椭圆参数提取算法得到最终椭圆参数,该方法利用椭圆参数空间聚类分析选取最优椭圆参数,从而实现了一致样本集的自适应选择,在保证了椭圆参数拟合精度的同时提高了算法的适应性和鲁棒性。结果采用本文算法检测一幅图像的平均时间约为110 ms,与常用椭圆检测方法相比检测速度有显著提高。结论对比实验结果表明,本文提出的椭圆检测方法与其他方法相比具有更高的精度、更快的实时性和更强的鲁棒性。
Aiming at the problems such as the obfuscation of the target image and the uneven brightness in the inertial confinement fusion experiment, this paper proposes a fast ellipse detection method with high reliability and precision from the perspective of improving the real-time performance of the image processing. Methods Firstly, using the feature that the ellipse edge has a larger rate of change of gray in the direction of the center of the ellipse, a polar coordinate system is established by using the estimated center as the pole. Based on the extremal gray value of the ray, Search for the realization of the ellipse edge point detection, extreme point search in the local area of the image to ensure the validity and real-time detection of edge detection; secondly, the use of RANSAC adaptive elliptic parameter extraction algorithm to obtain the final elliptic parameters, the method using elliptic parameter space poly The class analysis chooses the optimal ellipse parameters, so that the adaptive selection of consistent sample sets is achieved. The fitting accuracy of elliptic parameters is guaranteed and the adaptability and robustness of the algorithm are improved. Results The average time for detecting an image with this algorithm is about 110 ms, which shows a significant increase in detection speed compared with the commonly used ellipse detection method. Conclusion The experimental results show that the proposed ellipse detection method has higher accuracy, faster real-time performance and stronger robustness than other methods.