Improvement of Lightweight Convolutional Neural Network Model Based on YOLO Algorithm and Its Research in Pavement Defect Detection.

To ensure the safe operation of highway traffic lines, given the imperfect feature extraction of existing road pit defect detection models and the practicability of detection equipment, this paper proposes a lightweight target detection algorithm with enhanced feature extraction based on the YOLO (You Only Look Once) algorithm. The BIFPN (Bidirectional Feature Pyramid Network) network structure is used for multi-scale feature fusion to enhance the feature extraction ability, and Varifocal Loss is used to optimize the sample imbalance problem, which improves the accuracy of road defect target detection. In the evaluation test of the model in the constructed PCD1 (Pavement Check Dataset) dataset, the mAP@.5 (mean Average Precision when IoU = 0.5) of the BV-YOLOv5S (BiFPN Varifocal Loss-YOLOv5S) model increased by 4.1%, 3%, and 0.9%, respectively, compared with the YOLOv3-tiny, YOLOv5S, and B-YOLOv5S (BiFPN-YOLOv5S; BV-YOLOv5S does not use the Improved Focal Loss function) models. Through the analysis and comparison of experimental results, it is proved that the proposed BV-YOLOv5S network model performs better and is more reliable in the detection of pavement defects and can meet the needs of road safety detection projects with high real-time and flexibility requirements.

EXPLORING THE ORIGINS OF DEUTERIUM ENRICHMENTS IN SOLAR NEBULAR ORGANICS.

Deuterium-to-hydrogen (D/H) enrichments in molecular species provide clues about their original formation environment. The organic materials in primitive solar system bodies generally have higher D/H ratios and show greater D/H variation when compared to D/H in solar system water. We propose this difference arises at least in part due to (1) the availability of additional chemical fractionation pathways for organics beyond that for water, and (2) the higher volatility of key carbon reservoirs compared to oxygen. We test this hypothesis using detailed disk models, including a sophisticated, new disk ionization treatment with a low cosmic-ray ionization rate, and find that disk chemistry leads to higher deuterium enrichment in organics compared to water, helped especially by fractionation via the precursors CH 2 D + / CH 3 + . We also find that the D/H ratio in individual species varies significantly depending on their particular formation pathways. For example, from ~20-40 au, CH4 can reach D/H ~ 2 × 10-3, while D/H in CH3OH remains locally unaltered. Finally, while the global organic D/H in our models can reproduce intermediately elevated D/H in the bulk hydrocarbon reservoir, our models are unable to reproduce the most deuterium-enriched organic materials in the solar system, and thus our model requires some inheritance from the cold interstellar medium from which the Sun formed.

The ancient heritage of water ice in the solar system.

Identifying the source of Earth's water is central to understanding the origins of life-fostering environments and to assessing the prevalence of such environments in space. Water throughout the solar system exhibits deuterium-to-hydrogen enrichments, a fossil relic of low-temperature, ion-derived chemistry within either (i) the parent molecular cloud or (ii) the solar nebula protoplanetary disk. Using a comprehensive treatment of disk ionization, we find that ion-driven deuterium pathways are inefficient, which curtails the disk's deuterated water formation and its viability as the sole source for the solar system's water. This finding implies that, if the solar system's formation was typical, abundant interstellar ices are available to all nascent planetary systems.