Abstract: A deep learning automatic classification tree species method based on remote sensing images addresses the limitations of existing tree species classification approaches, which tend to focus excessively on single-label classification tasks while overlooking hierarchical multi-label classification, and fail to achieve high-precision species identification through hierarchical labels. The proposed invention is implemented by constructing hierarchical tree species classification data sets at various scales, developing hierarchical tree species classification models, and evaluating the transferability of these models. This method enhances the accuracy of tree species classification, through the dynamic adjustment mechanism of the switcher within the Switcher-HNet and the collaborative functioning of the classifiers, and it reduces misclassification between needleleaf and broadleaf trees, thereby decreasing the overall classification error.

Abstract: The present disclosure provides a data search method, and relates to the field of computer technology. The method includes determining semantic-understanding information corresponding to a received search term by subjecting the search term to semantic understanding; analyzing the semantic-understanding information to obtain timeliness requirement information of the search term; determining an acquisition mode of resource result corresponding to the search term based on the timeliness requirement information of the search term; and obtaining the resource result corresponding to the search term by using the determined acquisition mode. The present disclosure further provides a data search apparatus, a system, a device, and a computer readable storage medium.

Abstract: The present disclosure provides a method of optimizing a search system, which relates to a field of data processing, and in particular to a field of data search. The method is implemented to include: determining a first hit rate of a cache unit of the search system for a plurality of user queries; for each element in a first set of elements, determining at least one key element by: generating a plurality of first queries that correspond to the plurality of user queries; determining a second hit rate of the offline cache unit for the plurality of first queries; and determining the element as one of at least one key element in response to determining that a difference between the second hit rate and the first hit rate is less than a difference threshold; and optimizing the search system based on the at least one key element.

Abstract: The present disclosure provides a data search method, and relates to the field of computer technology. The method includes determining semantic-understanding information corresponding to a received search term by subjecting the search term to semantic understanding; analyzing the semantic-understanding information to obtain timeliness requirement information of the search term; determining an acquisition mode of resource result corresponding to the search term based on the timeliness requirement information of the search term; and obtaining the resource result corresponding to the search term by using the determined acquisition mode. The present disclosure further provides a data search apparatus, a system, a device, and a computer readable storage medium.

Abstract: A photolithography method, a method of preparing a flexible substrate and a photoresist drying device are provided. The photolithography method includes: providing a base substrate on which a material layer to be etched is formed, in which the base substrate includes an intermediate region and a peripheral region surrounding the intermediate region; coating a layer of photoresist on the base substrate, in which the photoresist is coated in the intermediate region and the peripheral region, and is formed to cover the material layer to be etched; and drying the photoresist and simultaneously performing a first exposure process on the photoresist coated in the peripheral region.

Abstract: A photolithography method, a method of preparing a flexible substrate and a photoresist drying device are provided. The photolithography method includes: providing a base substrate on which a material layer to be etched is formed, in which the base substrate includes an intermediate region and a peripheral region surrounding the intermediate region; coating a layer of photoresist on the base substrate, in which the photoresist is coated in the intermediate region and the peripheral region, and is formed to cover the material layer to be etched; and drying the photoresist and simultaneously performing a first exposure process on the photoresist coated in the peripheral region.

Abstract: One embodiment of the invention is a multi layer GI POF where each layer comprises essentially one polymer and one or more dopants. Each layer may comprise the same or different polymers although it is preferred that each layer comprise the same homopolymer. Useful dopants include cyclic or acyclic organic compounds of less than about 20 carbons, alkyl metal oxides or rare earth alkyl oxides. Another embodiment of the invention is a method of making the GI POF according to the invention.

Abstract: A crosshead die for a plastic optical fiber fabrication system is provided. The crosshead die includes a body, an insert, mounting flanges on the upstream and downstream end for securing the crosshead die to adjacent equipment or another crosshead dies, an axial port for receiving a mixed molten material that makes up the core, and a radial port for receiving a mixed molten material that makes up the outer layering material to be co-extruded over the core. The die insert includes a material distribution channel system that provides substantially the same linear distance for the mixed molten material to travel prior to the co-extrusion. If multiple die inserts are utilized, the crosshead die is able to co-extrude multiple layers over the core. If the crosshead dies are serially assembled, additional concentric layers of material can be co-extruded atop the layered plastic optical fiber extruded from the upstream crosshead die.

Abstract: One embodiment of the invention is a multi layer GI POF where each layer comprises essentially one polymer and one or more dopants. Each layer may comprise the same or different polymers although it is preferred that each layer comprise the same homopolymer. Useful dopants include cyclic or acyclic organic compounds of less than about 20 carbons, alkyl metal oxides or rare earth alkyl oxides. Another embodiment of the invention is a method of making the GI POF according to the invention.

Abstract: One embodiment of the invention is a multi layer GI POF where each layer comprises essentially one polymer and one or more dopants. Each layer may comprise the same or different polymers although it is preferred that each layer comprise the same homopolymer. Useful dopants include cyclic or acyclic organic compounds of less than about 20 carbons, alkyl metal oxides or rare earth alkyl oxides. Another embodiment of the invention is a method of making the GI POF according to the invention.