摘要
当前图像识别采取边缘算子切割技术,图像目标边缘会出现一些灰度急剧变化的情况。采用阈值法,图像中存在的阴影,照度不均匀,各处的对比度不同的情况,所以不能采用全局阈值法,而采用局部阈值法,即将图像划分为若干个小图像,先对各子图像用阈值法进行分割,再将分割后的小区域合并在一起,得到整幅图像的完整分割结果,这样,不同的区域由于不同的情况,就可以选取相应的阈值,达到最好得分割效果。
Taking cutting edge operators current image recognition technology,some dramatic changes in gray image target Edge occurs.Used threshold value law,image in the exists of shadow,according to degrees does not uniform,throughout of comparison degrees different of situation,so cannot used global threshold value law,and used local threshold value law,will image Division for several a small image,first on the child image with threshold value law for segmentation,again will segmentation Hou of small regional merged in together,are whole site image of full segmentation results,such,different of regional due to different of situation,on can select corresponding of threshold value,reached best of segmentation effect.
出处
《科技通报》
北大核心
2013年第2期76-78,共3页
Bulletin of Science and Technology
关键词
自动语义
标注
图像识别
automatic semantic
marking
image recognition
