47results about How to "Reduce the impact of extraction" patented technology

The invention discloses a complete roasting process for energy-saving regenerated lithium iron phosphate, which belongs to the technical field of new energy. A complete roasting process for energy-saving regenerated lithium iron phosphate comprises the following steps: dismantling waste lithium batteries, ball milling positive electrode materials, Raw material powder mixing, mixing and dehydration, roasting and melt impregnation, dry grinding is expected, the obtained lithium iron phosphate is expected to be put into the mill, and dry grinding is carried out for a certain period of time to obtain lithium iron phosphate material, which can realize the utilization of carbon-based Glucose coats the lithium iron phosphate material with carbon, increases the diffusion coefficient of lithium ions between the two phases, increases the discharge capacity of the lithium iron phosphate cathode material, thereby improving the quality of the recovered lithium iron phosphate material, and at the same time through its own special The structure collects the protective gas during the roasting process, so that the protective gas during the roasting process of lithium iron phosphate can be recycled, and the cost consumption of regeneration and recovery of lithium iron phosphate is reduced.

The present invention relates to a high-resolution remote sensing image-based multi-index fusion landslide detection method, which comprises the steps of (1) acquiring a high-resolution remote sensing image and a stereoscopic image pair; (2) generating a digital elevation model and calculating terrain feature indexes; (3) preprocessing the remote sensing image; (4) conducting the multi-scale image segmentation for the remote sensing image; (5) selecting appropriate terrain feature indexes for different scale layers and obtaining the values of the terrain feature indexes; (6) conducting the fusion processing for terrain feature indexes and ground object feature indexes, and comparing each feature index with a preset rule set to realize the landslide detection. Compared with the prior art, the noise influence caused by the scattered distribution of landslide regions can be avoided, so that the landslide recognition and extraction problem caused by earthquakes or other natural damages can be solved. Therefore, the method provides support for the post-disaster reconstruction and restoration. The above feature indexes can reflect the features of a ground object within a selected region, and a threshold corresponding to the feature indexes is unique. Based on the combination of features and the determination of the threshold, the landslide region can be successfully extracted and studied.

The invention discloses a method for extracting urban built-up areas by fusing NPP-VIIRS night light data and Landsat8OLI images, which may include the following steps: 1. Obtain the city's NPP-VIIRS night light data and Landsat8OLI image spectrum data, and analyze NPP-VIIRS VIIRS night light data and Landsat8OLI data are preprocessed. 2. Extract NDVI and NDBI from Landsat8OLI data, where NDVI is the normalized difference vegetation index and NDBI is the normalized difference building index. 3. Use VBANUI (improved night light city index method) to extract urban built-up areas. The present invention can reduce the saturation phenomenon of luminous data in the urban core area by introducing the two indexes of NDVI and NDBI, and can remove the light overflow phenomenon of the night light brightness in the vegetation area of ​​the water body and the urban core area, and enhance the expression of building information. Improved the accuracy of urban built-up area extraction.

The present invention proposes a business card recognition method and device, wherein the method includes: acquiring a business card image to be recognized; inputting the business card image into a text detection model to obtain each text line area; wherein, the text detection model has learned image features The corresponding relationship with each text line area; each text line area is input into the character recognition model to obtain the business card information corresponding to each text line area. This method can implement a text detection model based on deep learning to identify each text line area in the business card image, which is robust and can reduce the impact of low-quality and noise data on text extraction, thereby improving the versatility and performance of the method. application space. In addition, the text recognition model based on deep learning performs end-to-end recognition of each text line area without the need for word segmentation, which has a higher accuracy rate, and also has a stronger recognition ability for various complex changes. The generality and recognition effect of the method.