Research on Natural Fiber Microstructure Detection Method Based on CA-DeepLabv3+ DOI Open Access
Shuaishuai Lv, Xiaoyuan Li, Hitoshi Takagi

и другие.

Materials, Год журнала: 2024, Номер 17(23), С. 5942 - 5942

Опубликована: Дек. 4, 2024

Natural fibers exhibit noticeable variations in their cross-sections, and measurements assuming a circular cross-section can lead to errors the values of properties. Providing more accurate geometric information fiber cross-sections is key challenge. Based on microscopic images natural structures, this paper proposes microstructure detection method based CA-DeepLabv3+ network model. The study investigates image segmentation algorithm that uses MobileNetV2 as feature extraction backbone network, optimizes Atrous Spatial Pyramid Pooling (ASPP) module through cascading, embeds an Efficient Multi-scale Attention (EMA) mechanism. results show proposed accurately segment microstructures multiple types fibers, achieving average pixel accuracy (mPA) 95.2% mean Intersection over Union (mIoU) 90.7%.

Язык: Английский

Effect of AO 4426 on damping properties of PVA/CPE-AO 2246 DOI

Jiang Sheng,

Yong Zhang

Journal of Polymer Engineering, Год журнала: 2025, Номер unknown

Опубликована: Фев. 3, 2025

Abstract To investigate the influence of AO 2246 and 4426 on damping properties polyvinyl alcohol/chlorinated polyethylene (PVA/CPE) composites, a series composites were prepared by adding into PVA/CPE-AO under constant mass ratio hindered phenol in composites. The dynamic mechanical microstructure materials investigated DMA, DSC, SEM, FT-IR. results showed that new peak appeared near 50 °C 4426, which indicated phase separation between matrix occurred. With increase peaks low-temperature section improved. value low-lying region double peaks, when coexisted was higher than containing only temperature domain effectively broadened. At melting microcrystalline microcrystalline, no obvious observed simultaneously, indicating phenols inhibited crystallization each other.

Язык: Английский

Процитировано

0

Enhancement of toughness and ductility to 1,2,3-triazole Click-cured elastomers via regulation of molecular network and microphase separation DOI

Shengda Zhang,

Ying He,

Shuiping Zhou

и другие.

European Polymer Journal, Год журнала: 2025, Номер unknown, С. 113835 - 113835

Опубликована: Фев. 1, 2025

Язык: Английский

Процитировано

0

Hierarchical Assembly of Cellulose Fibrils and Tannin in Biocomposite Foam: Scalable Production via Oven Drying and Customizable Metal Ions Release for Antimicrobial Activity DOI
Zonghong Lu, Hao Zhang, Qingbo Wang

и другие.

Small, Год журнала: 2025, Номер unknown

Опубликована: Апрель 14, 2025

Abstract Advanced cellulose‐based foams are urgently needed as sustainable packaging materials in an era of prioritizing environmental consciousness. However, transferring the mechanical properties cellulose fibers into porous structures is always limited by gas entrapment during foaming and irreversible structural collapse upon liquid evaporation. Herein, a hierarchical assembly strategy combines cationic nanofibrils (CCNF) with dynamic covalent tannin/borate (T/B) complex to fabricate 3D continuous distinct lamellar structure via oven drying proposed for scalable production. CCNF assembles T/B onto electrostatic attraction hydrogen bonding, while reversible bonds among impart shear‐thinning self‐healing properties, thereby ensuring foamability (exceeding 300%) stability. Moreover, foam offers versatile platform customization metal ions (Fe 3+ , Cu 2+ Ag + ), allowing tailoring physical properties. At optimized tannin addition 10%, 10T/5B‐Fe exhibits highest normalized strength above 410 Pa/density, maintaining ultralow density 9.2 mg cm − 3 . Additionally, pH‐responsiveness complexes enables release long‐term antimicrobial activity. This study demonstrates green functional production, offering new possibilities next‐generation materials.

Язык: Английский

Процитировано

0

Research on Natural Fiber Microstructure Detection Method Based on CA-DeepLabv3+ DOI Open Access
Shuaishuai Lv, Xiaoyuan Li, Hitoshi Takagi

и другие.

Materials, Год журнала: 2024, Номер 17(23), С. 5942 - 5942

Опубликована: Дек. 4, 2024

Natural fibers exhibit noticeable variations in their cross-sections, and measurements assuming a circular cross-section can lead to errors the values of properties. Providing more accurate geometric information fiber cross-sections is key challenge. Based on microscopic images natural structures, this paper proposes microstructure detection method based CA-DeepLabv3+ network model. The study investigates image segmentation algorithm that uses MobileNetV2 as feature extraction backbone network, optimizes Atrous Spatial Pyramid Pooling (ASPP) module through cascading, embeds an Efficient Multi-scale Attention (EMA) mechanism. results show proposed accurately segment microstructures multiple types fibers, achieving average pixel accuracy (mPA) 95.2% mean Intersection over Union (mIoU) 90.7%.

Язык: Английский

Процитировано

0