Liping Liu

To identify superior eucalyptus genotypes adapted to the cold-prone northern margin of Guangdong Province and to elucidate the genetic basis underlying growth, cold tolerance, and heterosis, this study evaluated 51 eucalyptus genetic materials with diverse genetic backgrounds, including Eucalyptus urophylla families and E. urophylla-derived interspecific hybrid combinations. Growth and cold tolerance traits were measured at 4.0 and 6.5 years after planting to systematically assess genetic variation, combining ability, and selection potential among different genetic materials. Significant genetic variation was observed among the tested materials for both growth and cold tolerance traits. Several families and hybrid combinations exhibited superior performance in stem volume production and cold tolerance, indicating substantial potential for genetic improvement. BLUP analysis revealed a genetic trade-off between growth and cold tolerance in 31 families; however, several elite genotypes exhibited a favorable combination of rapid growth and enhanced cold tolerance, making them promising candidates for future breeding programs. Combining ability analysis demonstrated that both additive and non-additive genetic effects contributed to growth performance, and superior parental combinations with high general combining ability and specific combining ability effects were identified. Furthermore, interspecific hybrid evaluation showed that E. urophylla × Eucalyptus grandis hybrids generally exhibited greater growth potential and more stable productivity, whereas E. urophylla × Eucalyptus camaldulensis hybrids displayed stronger adaptation to low-temperature stress. These findings provide valuable genetic resources and theoretical support for developing fast-growing, cold-tolerant eucalyptus cultivars and improving the stability and productivity of subtropical plantation forests.

In recent years, tragedies caused by forest fires have been frequently reported. Forest fires not only result in significant economic losses but also cause environmental damage. The utilization of computer vision techniques and unmanned aerial vehicles (UAVs) for forest fire monitoring has become a primary approach to accurately locate and extinguish fires during their early stages. However, traditional computer-based methods for UAV forest fire image segmentation require a large amount of pixel-level labeled data to train the networks, which can be time-consuming and costly to acquire. To address this challenge, we propose a novel weakly supervised approach for semantic segmentation of fire images in this study. Our method utilizes self-supervised attention foreground-aware pooling (SAP) and context-aware loss (CAL) to generate high-quality pseudo-labels, serving as substitutes for manual annotation. SAP collaborates with bounding box and class activation mapping (CAM) to generate a background attention map, which aids in the generation of accurate pseudo-labels. CAL further improves the quality of the pseudo-labels by incorporating contextual information related to the target objects, effectively reducing environmental noise. We conducted experiments on two publicly available UAV forest fire datasets: the Corsican dataset and the Flame dataset. Our proposed method achieved impressive results, with IoU values of 81.23% and 76.43% for the Corsican dataset and the Flame dataset, respectively. These results significantly outperform the latest weakly supervised semantic segmentation (WSSS) networks on forest fire datasets.