论文部分内容阅读
在风洞结冰模型3D冰形测量中,激光三角测量法相对传统测量方法检测速度快、精确度高,具有极高研究价值。然而由于冰体对激光透射严重,影响激光中心线提取精度从而影响测量精度。针对此问题,提出一种冰形表面激光光带中心线快速提取方法。该算法具体实现步骤为:首先,采用基于三维块匹配去噪算法对图像进行降噪,并对图像进行视觉显著性计算,分割出光带区域;其次,求取梯度图并转换至频域空间,根据图像频谱特征求取能量中心区域;最后,对区域采用灰度重心法求取中心线亚像素级位置。采用冰箱冻结的半径已知圆柱冰块评估该算法,测得算法处理时效为28.57FPS,使用该算法的冰形轮廓重建精度达到0.017mm。实验证明算法满足冰形在线测量技术要求,为开展结冰实验中结冰生长过程在线三维检测技术奠定技术基础。
In the 3D icing measurement of icing wind tunnel model, the laser triangulation method has the advantages of high detection speed and high accuracy compared with the traditional measurement methods, and has high research value. However, due to the ice body on the laser transmission is serious, affecting the laser center line extraction accuracy and thus affect the measurement accuracy. Aiming at this problem, a method of fast extraction of centerline of laser band on ice surface was proposed. The specific steps of the algorithm are as follows: Firstly, the image is denoised based on the three-dimensional block matching de-noising algorithm, and the image is calculated visually and the light zone is segmented. Secondly, the gradient map is obtained and transformed into the frequency domain space, According to the spectral features of the image, the central region of energy is obtained. Finally, the grayscale centroid method is used to find the sub-pixel position of the center line. This algorithm was evaluated by using a frozen block ice cube whose radius was known to be frozen. The processing time was 28.57 FPS, and the ice contour reconstruction accuracy was 0.017 mm. Experiments show that the algorithm meets the requirement of on-line ice measurement technology and lays the technical foundation for the on-line three-dimensional detection of icing growth process in icing experiment.