海空目标多尺度波峰最佳阈值自动搜索策略

Automatic search strategy for sea-sky target by multi-scale peak optimal threshold algorithm

针对海上低对比度低信噪比的条件下,能够精确快速搜索空域目标,提出了一种快速精确的海空目标多尺度波峰最佳阈值自动搜索策略:采用金字塔波门搜索和多尺度波峰阈值法,逐级减小搜索区域并对搜索域内图像进行分割,然后根据目标特征进行概率统计分析,确定最优目标,实现海空目标的自动搜索。实验结果表明,该算法在序列图像中的搜索定位平均误差为0.413像素并全部识别成功,较OTSU算法的2.61像素和最大熵阈值算法的3.1像素的误差,精确度大大提升。整幅图像搜索时间优于21.34 ms,满足海空目标自动搜索的精度和实时性的要求。

To search spatial domain target precisely and quickly in the condition of low contrast and low SNR, a new fast and accurate automatic search strategy for sea-sky target by multi-scale peak optimal threshold algorithm was proposed. Using the method of pyramid wave gate search and multi-scale peak threshold, the search area was gradually reduced and image segmentation was accomplished in this search area. Then, the probability and statistics according to the characteristics of the target was analysed to determine the optimum target and achieve automatic search for sea-sky target. Experimental results show that the average error of searching and locating is 0.413 pixel and all identified with this algorithm in the image sequences, compared with 2.61 pixel of OTSU algorithm and 3.1 pixel of the maximum entropy threshold algorithm, the accuracy rises obviously. The whole image search time is less than 21.34 ms, which meets the requirements of the precision and real-time for automatically search of the sea-sky target.