4results about How to "Capture more" patented technology

The application provides a forest fire smoke detection method based on semantic segmentation, and belongs to the technical field of fire warning, and comprises the following steps: feeding a real forest fire data set into a feature extraction network, inputting the extracted features into an improved DeepLabV3+ network for training, and outputting detection and recognition results. According to the complex and changeable forest fire scene, the non-rigid structure, the opacity and the variable mode of smoke and other characteristics, the application deeply studies a feature fusion method, adopts an improved ASPP structure to extract multi-scale spatial features, simultaneously introduces a bottom feature extraction module and a double-path convolution aggregation module to effectively extract smoke boundary information, and uses a lightweight backbone network to replace the original feature extraction network of DeepLabV3+, so that the detection speed is accelerated by using fewer parameters. The application can solve the problem that the false alarm and missed detection rates of the smoke detection method are high, thereby providing technical support for forest ecological protection.

The present application belongs to the technical field of green building, and particularly relates to a passive energy-saving building facade chord wave guide flow outer envelope system, which comprises an outer facade assembly, including chord wave guide flow mechanisms and an outer curtain wall arranged alternately from top to bottom, the chord wave guide flow mechanism is in a wave-shaped ridge structure and has a containing cavity inside; a wind turbine is installed in the containing cavity; wherein a gas collecting port is arranged at the upper top surface wave trough and / or the lower bottom surface wave peak of the chord wave guide flow mechanism, and a gas permeable hole is arranged on the outer surface. The present application can converge, guide and discharge natural wind field, improve the wind energy capture efficiency, can adapt to different direction air flow, hide the installation of the wind turbine, does not affect the building facade effect, can be arranged flexibly, can adapt to complex and changeable wind direction conditions, improves the stability and reliability of wind energy utilization, and can also continuously generate air flow power in windless weather, effectively makes up the problem of insufficient wind power, and ensures the continuity and stability of building energy supply.

The invention discloses outdoor glass cement with excellent weather resistance and a preparation method of the outdoor glass cement, and belongs to the technical field of binders. Comprising the following components in parts by weight: 100 parts of hydroxyl-terminated polydimethylsiloxane, 10 parts of dimethyl silicone oil, 80-120 parts of filler, 6-12 parts of a cross-linking agent, 0.05-0.2 part of an organic tin catalyst, 0.5-2 parts of an amino silane coupling agent, 5-8 parts of a cerium dioxide immobilized hindered phenol anti-aging agent and 10-12 parts of branched polyether modified silicone oil. According to the outdoor glass cement with excellent weather resistance and the preparation method thereof provided by the invention, the cerium dioxide immobilized hindered phenol anti-aging agent is added into the glass cement, and hindered phenol is immobilized by cerium dioxide to form a synergistic anti-aging effect; branched polyether modified silicone oil is also added into the glass cement to serve as a dispersing agent, so that the problem of agglomeration or poor dispersity caused by interface incompatibility between the filler and the matrix silicone rubber is solved.

This invention belongs to the field of text summarization extraction technology in NLP, and discloses a word-based short text summarization extraction method, including an extraction model and a word order model. The word order model steps are as follows: Step 1: Data labeling; S1.1: Select 500,000 short dialogue texts from a telemarketing scenario, and manually review and correct the word order of each sentence; S1.2: Use jieba word segmentation, with the original word order labeled as 1, then enumerate all word combinations for each sentence, labeled as 0, and finally manually review and correct all label data; Step 2: Data preprocessing. This invention uses a dilated convolutional neural network (DCNN) in the extraction model network structure, with the dilation coefficient set in the order of 1, 2, 4, 1, enabling the model to capture as much text information as possible, and adds a word order model, improving the effectiveness and continuity of the summary, thus perfectly solving the problem of incoherence in extraction-based summarization.