CAJ | 학술논문

现有的图像分割算法存在着耗时量大，分割效果不佳等问题，不适用与红外系统领域的应用。针对上述问题，根据灰度级－梯度二维直方图的目标分割优势，通过与蚁群算法相结合，提出了一种结合蚁群算法与二维直方图的红外图像分割算法。通过在传统的灰度－梯度二维直方图进行引入边缘与噪声区域的相关量；通过将图像窗口化，并根据最佳分割阈值对蚁群的启发函数以及信息素更新进行重新定义，来实现红外目标的快速提取。实验结果表明，该算法分割后的红外目标边缘清晰，抗干扰能力较强，且运算速度也得到了有效提高。
현유적도상분할산법존재착모시량대，분할효과불가등문제，불괄용여홍외계통영역적응용。침대상술문제，근거회도급－제도이유직방도적목표분할우세，통과여의군산법상결합，제출료일충결합의군산법여이유직방도적홍외도상분할산법。통과재전통적회도－제도이유직방도진행인입변연여조성구역적상관량；통과장도상창구화，병근거최가분할역치대의군적계발함수이급신식소경신진행중신정의，래실현홍외목표적쾌속제취。실험결과표명，해산법분할후적홍외목표변연청석，항간우능력교강，차운산속도야득도료유효제고。
Existing image segmentation algorithm has the problems of large calculation amount and poor segmentation effect.Aiming at these problems,according to target segmentation characteristics of grayscale-gradient two-dimension-al histogram,an infrared image segmentation algorithm based on ant colony algorithm and two-dimensional histogram is proposed.The edges and noise correlation values are introduced into the traditional gray-gradient two-dimensional his-togram.In order to extract infrared target rapidly,the image is divided into small windows,and the heuristic function and pheromone update function are redefined according to the optimal segmentation threshold.Experimental results show that the edge of infrared target is clear.The algorithm has a stronger anti-jamming capability and faster compu-ting speed.