Seed sizes of 74 woody species, including 40 trees and 34 shrubs, of middle subtropical evergreen broad-leaved forest in Mt. Jinyun were investigated and the likely mechanism of seed size variation was analysized. The...Seed sizes of 74 woody species, including 40 trees and 34 shrubs, of middle subtropical evergreen broad-leaved forest in Mt. Jinyun were investigated and the likely mechanism of seed size variation was analysized. The results showed that: 1) There was considerable variance in the seed length and great variance in the seed weight among the 74 species. The maximum seed length was 100 times greater than the minimum. The lengths of most seeds were in a range of 0.1～1 cm, accounted for 71.62% of the total species. The maximum seed weight was 10 000 times greater than the minimum. A 39.41% of species had their 1 000 seed weights in a range of 10～100 g. 2) The results of statistic analysis indicated the seed sizes were influenced by their habitats. The longest seeds were born in canopy layer, the middle in forest edge & understorey and the shortest in dankness habitat. Additionally, the seed sizes were also influenced by growth forms of species. The seed sizes gradually decreased with the growth forms sequence of large tree, middle tree, small tree, large shrub and small shrub. Moreover, the phylogenetic backgrounds of species had little influences on the seed sizes. Little correlations were found between seed sizes and their orders, families, taxonimical groups with the different weight ratio of endosperm and different fruit types.展开更多
在存在壁面反射的低照度火灾环境中,传统的火焰分割算法如颜色分割、运动检测等,在进行火焰分割时造成过分割现象,分割的效果不理想,影响后续的火灾正确识别。针对上述问题,提出了一种基于自动种子区域生长(Automatic Seeded Region Gro...在存在壁面反射的低照度火灾环境中,传统的火焰分割算法如颜色分割、运动检测等,在进行火焰分割时造成过分割现象,分割的效果不理想,影响后续的火灾正确识别。针对上述问题,提出了一种基于自动种子区域生长(Automatic Seeded Region Growing,ASRG)的火焰分割算法。首先将从火灾视频中获取的火灾图像从RGB颜色空间转换到YCbCr颜色空间,在Y通道中采用较大自适应阈值背景减法将火灾图像二值化,分别将可疑火焰像素点的横坐标和纵坐标按大小进行排序,取排序后的中间值作为种子点,再由原RGB火灾图像转换而成的灰度图像中,以该种子点进行区域生长,最后将区域生长后的火焰分割图像与采用较小自适应阈值背景减法得到的火焰分割图像进行交集处理,得到最终的火焰分割图像。实验表明ASRG算法在存在壁面反射的低照度火灾环境中,火焰分割效果好,有效解决了该环境下的火焰过分割问题,同时在其他火灾环境中也有较好的火焰分割效果。展开更多
文摘Seed sizes of 74 woody species, including 40 trees and 34 shrubs, of middle subtropical evergreen broad-leaved forest in Mt. Jinyun were investigated and the likely mechanism of seed size variation was analysized. The results showed that: 1) There was considerable variance in the seed length and great variance in the seed weight among the 74 species. The maximum seed length was 100 times greater than the minimum. The lengths of most seeds were in a range of 0.1～1 cm, accounted for 71.62% of the total species. The maximum seed weight was 10 000 times greater than the minimum. A 39.41% of species had their 1 000 seed weights in a range of 10～100 g. 2) The results of statistic analysis indicated the seed sizes were influenced by their habitats. The longest seeds were born in canopy layer, the middle in forest edge & understorey and the shortest in dankness habitat. Additionally, the seed sizes were also influenced by growth forms of species. The seed sizes gradually decreased with the growth forms sequence of large tree, middle tree, small tree, large shrub and small shrub. Moreover, the phylogenetic backgrounds of species had little influences on the seed sizes. Little correlations were found between seed sizes and their orders, families, taxonimical groups with the different weight ratio of endosperm and different fruit types.
文摘在存在壁面反射的低照度火灾环境中,传统的火焰分割算法如颜色分割、运动检测等,在进行火焰分割时造成过分割现象,分割的效果不理想,影响后续的火灾正确识别。针对上述问题,提出了一种基于自动种子区域生长(Automatic Seeded Region Growing,ASRG)的火焰分割算法。首先将从火灾视频中获取的火灾图像从RGB颜色空间转换到YCbCr颜色空间,在Y通道中采用较大自适应阈值背景减法将火灾图像二值化,分别将可疑火焰像素点的横坐标和纵坐标按大小进行排序,取排序后的中间值作为种子点,再由原RGB火灾图像转换而成的灰度图像中,以该种子点进行区域生长,最后将区域生长后的火焰分割图像与采用较小自适应阈值背景减法得到的火焰分割图像进行交集处理,得到最终的火焰分割图像。实验表明ASRG算法在存在壁面反射的低照度火灾环境中,火焰分割效果好,有效解决了该环境下的火焰过分割问题,同时在其他火灾环境中也有较好的火焰分割效果。
