2results about How to "Positive technical effect" patented technology

The utility model relates to a product turnover carrying mechanism, including base, vertical drive component, bearing frame and two turnover subassembly, and turnover subassembly includes drive motor, transmission mechanism and adsorption component, and drive motor rotates adsorption component through transmission mechanism, bearing frame includes the mounting panel connected with the output end of vertical drive component, and first vertical plate and second vertical plate are connected respectively in the both sides of mounting panel, and one turnover subassembly is installed on first vertical plate or second vertical plate respectively, and the relief hole is set up on first vertical plate or second vertical plate, and one of two turnover subassembly, its adsorption component is worn in the relief hole, makes the adsorption component of two turnover subassembly to the same direction and presents high and low arrangement, the utility model discloses adopt the integrated design concept, through the clever relief structure of sharing bearing frame, makes two turnover subassembly in the limited space cooperation work, effectively solved the equipment layout limited problem of traditional mechanism because of the structure redundancy.

The application discloses a kind of based on unmanned aerial vehicle and deep learning's field crop seedling high-precision detection method.The method is first obtained by unmanned aerial vehicle shooting high-definition remote sensing image, then these images are spliced into DOM image with geographic coordinates by unmanned aerial vehicle image processing software, DOM image is cut, detected, spliced and handled step by step, to obtain the specific location and size information of field crop seedling.Combining data analysis and image processing technology, further obtain field crop seedling information.In the key link of detection and splicing, the application adopts IoA method, which is used to eliminate the errors produced in the detection process of deep learning model and the deviation brought by image cutting process.The results show that, when identifying field crop seedling, the seedling counting accuracy of the method reaches more than 95.3%, and the mAP index is not less than 87.1%.The method provides an effective technical approach for high-precision detection of field crop seedling.