论文部分内容阅读
根据海面运动目标及成像传感器的特点,本文提出了一种红外目标图像序列跟踪点自动识别的智能跟踪算法.以平均梯度为特征进行目标图像分割;并根据目标最亮区与吃水线的空间关系选择目标最亮区中心点的垂线与吃水线的交点作为跟踪点;利用D-S证据理论融合帧间关联特征的跟踪点的自动识别以及可信度确认.实验证明本方法合理而且十分有效.
According to the characteristics of sea surface moving targets and imaging sensors, an intelligent tracking algorithm for automatic recognition of infrared target image sequence tracking points is proposed in this paper.The target image segmentation is characterized by average gradient, and according to the spatial relationship between the target brightest area and the waterline The intersection of the vertical line and the waterline of the center of the brightest area of the target is selected as the tracking point, the automatic identification of the tracking points and the confirmation of the credibility of the inter-frame correlation feature are fused by DS evidence theory, and the experiment proves that the method is reasonable and very effective.