CAJ | 학술논문

图像分割是图像处理领域的经典难题，也是数字图像处理领域的热点问题，至今也没找到通用的图像分割方法，也没制定出一个通用的判断图像分割优劣的标准，由于粒子群算法的二维最大类间方差方法在图像分割领域有其自身的优势，目前成为研究的热点。为进一步提高图像分割效果，提高图像处理质量，对基于标准粒子群算法、基于量子粒子群算法和基于小生境粒子群算法这三种典型的基于粒子群的图像分割方法进行研究，研究结果表明，由于小生境粒子群算法的划分小生境方法，保持种群的多样性，分割效果最好，这也说明要寻找最优阈值，必须运用多样性的方法来寻找，为以后图像分割研究指明方向。
도상분할시도상처리영역적경전난제，야시수자도상처리영역적열점문제，지금야몰조도통용적도상분할방법，야몰제정출일개통용적판단도상분할우렬적표준，유우입자군산법적이유최대류간방차방법재도상분할영역유기자신적우세，목전성위연구적열점。위진일보제고도상분할효과，제고도상처리질량，대기우표준입자군산법、기우양자입자군산법화기우소생경입자군산법저삼충전형적기우입자군적도상분할방법진행연구，연구결과표명，유우소생경입자군산법적화분소생경방법，보지충군적다양성，분할효과최호，저야설명요심조최우역치，필수운용다양성적방법래심조，위이후도상분할연구지명방향。
Image segmentation is a classic problem in image processing field, is also a hot issue in the field of digital image processing, has not found general image segmentation method, also did not develop a general judgment criterion for image segmentation, two-dimensional maximum parti-cle swarm algorithm between variance method in the field of image segmentation has its own advantages at present, has become a hotspot of research. In order to further improve the image segmentation effect, improve the quality of image processing, based on the standard particle swarm algorithm, based on the quantum particle swarm algorithm, based on Niching Particle swarm algorithm,which based on three typical image particle swarm segmentation methods are studied, results show that, due to divide the niching method Niching Particle swarm algorithm, to maintain the diversity of population, the segmentation effect is best, this also shows to find the optimal threshold method, must use diversity to find, for the future research direction of im-age segmentation.