相关跟踪算法在水下高速运动体识别中的研究

Application of correlation tracking algorithm to underwater object in high-speed motion

针对水下高速运动体试验图像背景灰度分布不均、信噪比低、对比度下降明显以及图像效果差等缺点,采用一种基于特征相关系数作为相似性度量准则的相关跟踪方法,克服了传统算法易受图像中噪声、局部遮挡等因素影响的缺点,该算法相对于传统算法计算量降低,提高了系统的运算速度.帧间图像像素变化率快,因此仅仅用某种固定的模板进行相关跟踪容易导致失配甚至失去目标,进而又提出一种行之有效的模板更新方法,使得跟踪算法对环境的适应能力和稳定性得到较大提高.实验结果表明,该算法跟踪精度高、速度快、稳定性好,满足实时性系统的要求.

Because of shortcomings associated with images of underwater objects in high-speed motion, including background pixel non-uniform distribution of background pixels, low SNR, low contrast, and general low picture quality, a correlation tracking algorithm was adopted that is based on a correlation coefficient. The new algorithm overcomes the disadvantages of a traditional correlation disturbed frequently by noise and occlusion, and it also reduces the computational complexity and improves the operating speed of the system. Rapidly changing picture pixels can cause mismatch or even loss of an object with the apoptotic template. However, this tried and true method for template renewal can improve the adaptability of environment and the stability of the algorithm greatly. Experiments show the algorithm has great accuracy, high speed, and good stability, and can meet the requirements of a real-time system.