一种空中多目标的定位及识别快速算法

针对空中多个目标进行精确定位及识别的特殊应用场合，该文采用了先进的改进不变矩的算法来进行目标的识别。由于不变矩识别对目标轮廓要求很高，该文提出了一种新的具有方向估计性质的形态学梯度边缘检测算子，弥补了现行的同类算子无法估计边缘方向的缺点；将模糊理论引入该算子，使其更好地抑制噪声，成功分离目标和背景；提出了一种新的断点连接及孤点消除的算法，使目标的轮廓更加清晰平滑，并用跟踪法找到其单像素轮廓线。实验证明，这种新算法极大地提高了对空中多个目标的识别率，并大大加强了系统的可靠性。

A Fast Arithmetic about Positioning and Identifying Multi-targets in Air

Aiming at the special situation of positioning and identifying multi-targets in air, this paper has the advanced invariant moment arithmetic to carry out. Because the moments request good quality of identifying, so this paper provides a new morphological gradient image edge detection operator which can estimate orientation, and proves them in theory and practice. There are four element structures with different direction. In these operators, the blurring processing is added, it makes that the operators not only depress the noise in image, but also enhance edge's definition. And then it changes the edge detection image into the binary image through the optimal threshold segmentation. It also shows a new simple arithmetic about jointing broken lines. It does not only make the targets' contour smooth, but also can find the single pixel boundary with tracking. Through kinds of tests, the identifying rate has been improved, and the dependability has been strengthened.