Настройки

Укажите год
-

Небесная энциклопедия

Космические корабли и станции, автоматические КА и методы их проектирования, бортовые комплексы управления, системы и средства жизнеобеспечения, особенности технологии производства ракетно-космических систем

Подробнее
-

Мониторинг СМИ

Мониторинг СМИ и социальных сетей. Сканирование интернета, новостных сайтов, специализированных контентных площадок на базе мессенджеров. Гибкие настройки фильтров и первоначальных источников.

Подробнее

Форма поиска

Поддерживает ввод нескольких поисковых фраз (по одной на строку). При поиске обеспечивает поддержку морфологии русского и английского языка
Ведите корректный номера.
Ведите корректный номера.
Ведите корректный номера.
Ведите корректный номера.
Укажите год
Укажите год
Применить Всего найдено 7. Отображено 7.
29-08-2023 дата публикации

Large-scale grid generation algorithm for parallel block extraction

Номер: CN116663364A
Автор: ZHAO WANQING, LI CHUNNAN, PENG JINYE, LI JIAMING, YANG WENJING, PENG SHENGLIN, WANG JUN
Принадлежит:

The invention belongs to the technical field of large-scale grid generation, and particularly discloses a large-scale grid generation algorithm for parallel block extraction. Comprising the following steps of: reading a model, generating coarse-grained body grids, partitioning global point elements, merging global coarse-grained surface grid information, extracting sub-region coarse-grained surface grid information, encrypting sub-region surface grids, generating sub-region fine-grained body grids, performing communication to obtain adjacent surface body grids, merging to generate large-scale fine-grained body grids, and outputting the large-scale fine-grained body grids. According to the method, the fine-grained sub-region body grids with the body grid element information on the adjacent surfaces are generated, topological relation reconstruction can be simplified when large-scale grids are combined, the time for combining the large-scale grids is greatly shortened, and the memory occupied ...

Подробнее
Номер записи: 1
07-04-2023 дата публикации

Gradient-based Gabor and channel attention-based mural flaking disease marking method

Номер: CN115937582A
Автор: WANG JUN, LI YUHENG, YU KAI, LIANG JINGYI, PENG SHENGLIN, PENG JINYE, WANG LIN, ZHAO WANQING
Принадлежит:

The invention discloses a gradient trainable Gabor and channel attention-based mural flaking disease marking method, and belongs to the technical field of image processing, and the method comprises the following steps: S1, obtaining an ancient mural image containing a flaking disease, and obtaining a training set, S2, inputting the training set into a feature extraction network, and obtaining a feature extraction result; respectively obtaining a feature map Fg and a feature map Fc through gradient training Gabor and classic convolution; s3, the feature map Fg and the feature map Fc are superposed together through an encoder, and Fgc is obtained to serve as input of a subsequent network for encoder convolution; and S4, the decoder adds a channel attention mechanism to the Fgc to enhance the relation of channel information among the feature maps, and a segmentation prediction map Rs is obtained. The invention provides an end-to-end mural flaking disease marking network, so that automatic ...

Подробнее
Номер записи: 2
04-04-2023 дата публикации

Style migration method for converting landscape photo into Chinese landscape painting

Номер: CN115908602A
Автор: HU QIYAO, XIE PENGLIN, PENG XIANLIN, ZHAO WANQING, YUAN JIANHUI, PENG SHENGLIN, PENG JINYE, FAN JIANPING
Принадлежит:

The invention provides a style migration method for converting a landscape photo into a Chinese landscape painting, which adopts an image style migration framework based on CycleGAN, and adopts an RRDB module with a reflection filling layer for the conditions of artifacts and light spots existing in the migration process; a spectrum normalization U-Net discriminator is adopted to improve the capability of the discriminator, and the training dynamic state is stabilized; and high-frequency information such as edges and details is reserved by using an LMS-SSIM loss function, and the brightness and color of the generated image are optimized. A loss function is adopted to enable the model to restrain the image after style migration from the aspects of brightness, contrast and structure, so that the image generated in style conversion can better conform to the visual perception of people.

Подробнее
Номер записи: 3
28-07-2023 дата публикации

Safety analysis method for rammed earth brick-wrapped urban wall ruins

Номер: CN116502500A
Автор: PENG JINYE, ZHANG GUOQI, YANG WENJING, ZHONG SHENG, ZHAO WANQING, WANG JUN, PENG SHENGLIN, WEI JIAQI, WANG LIN
Принадлежит:

The invention belongs to the technical field of ancient site protection, and relates to a safety analysis method for a rammed earth brick urban wall site, and a two-dimensional finite element simulation method comprises the following steps: 1, obtaining simplified three-dimensional data of the rammed earth brick urban wall site; 2, establishing a three-dimensional model of the rammed-earth brick-wrapped urban wall ruins according to the three-dimensional data; step 3, simplifying the three-dimensional model of the rammed earth brick-wrapped city wall ruins to construct a two-dimensional model; and 4, performing finite element analysis according to the two-dimensional model, and determining the safety coefficient of the rammed earth brick urban wall ruins. According to the method, the two-dimensional model of the rammed earth brick wall ruins is established, finite element simulation is carried out through the two-dimensional model, the stability of the rammed earth brick wall ruins is analyzed ...

Подробнее
Номер записи: 4
07-04-2023 дата публикации

Weak supervision whole heart segmentation method based on 3D convolution

Номер: CN115937513A
Автор: WANG JUN, SIMA MINGHUI, PENG SHENGLIN, WANG LIN, PENG JINYE, ZHAO GUOYING, WEI JIAQI, YUE MINGNAN
Принадлежит:

The invention discloses a weak supervision whole heart segmentation method based on 3D convolution. The method comprises the following steps: firstly, generating a pseudo tag with relatively high accuracy for a whole heart data set by using a deformable image registration technology; then cutting preprocessing is carried out by using GT to cut out a sub-block with a high foreground ratio so as to relieve the problem of class imbalance, and a cut pseudo tag and original data are fused and put into network training so as to provide rich feature information for distinguishing the foreground from the background; then, an attention mechanism is introduced into jump connection of the network, so that the network learns more useful features, and finally, an improved weighted cross entropy loss function is used to pay attention to boundary voxels which are difficult to segment. According to the method, the advantages of the pseudo labels are effectively utilized, accurate segmentation of the whole ...

Подробнее
Номер записи: 5
27-06-2023 дата публикации

Hyperspectral image fusion method based on dual-resolution residual mutual guidance network

Номер: CN116342994A
Автор: WANG JUN, MA KAI, LIU LU, PENG JINYE, PENG SHENGLIN, YU KAI, WANG LIN, ZHAO WANQING
Принадлежит:

The invention provides a hyperspectral image fusion method based on a dual-resolution residual mutual guidance network. The hyperspectral image fusion method comprises the following steps: step 1, establishing a training data set of a hyperspectral image and a visible light image; down-sampling all visible light images in the training data set, and forming a low-resolution data pair with the hyperspectral image in the training data set; all hyperspectral images in the training data set are subjected to up-sampling, and the hyperspectral images and visible light images in the training data set form high-resolution data pairs; step 2, obtaining a fusion result of a hyperspectral image and a visible light image in a data set of the high-resolution data pair and the low-resolution data pair through a dual-resolution residual mutual guidance network model; and training the dual-resolution residual mutual guidance network model by using the training data set to obtain a fusion model. Finally, ...

Подробнее
Номер записи: 6
11-04-2023 дата публикации

Traditional Chinese medicine formula for improving male idiopathic oligoasthenospermia, decoction and tablet

Номер: CN115944687A
Автор: CHENG YIMIN, MEI GUANGQUAN, ZOU YIPING, PENG SHAOJUN, WU ZHENGPING, PENG SHENGLIN, PENG ZHEN
Принадлежит:

The invention relates to a traditional Chinese medicine formula for improving male idiopathic oligospermia, a decoction and a tablet, and belongs to the technical field of traditional Chinese medicines. The traditional Chinese medicine formula comprises the following components in parts by mass: 10-20 parts of astragalus membranaceus, 15-20 parts of fructus lycii, 8-12 parts of cistanche, 8-12 parts of medicated leaven, 10-14 parts of rhizoma polygonati, 6-10 parts of rhizoma typhonii, 6-10 parts of rhizoma acori graminei, 4-8 parts of walnut meat, 4-8 parts of polygonum multiflorum, 2-6 parts of ginseng, 2-4 parts of angelica sinensis, 2-4 parts of radix rehmanniae preparata and 1-3 parts of liquorice. After the components are prepared into tablets or decoction for treating idiopathic oligospermia and asthenospermia, the sperm density, the forward movement sperm percentage and the sperm motility can be remarkably improved after 60 days of treatment.

Подробнее
Номер записи: 7