| CPC G05D 1/667 (2024.01) [G01N 1/08 (2013.01); G05D 1/243 (2024.01); G05D 2105/80 (2024.01); G05D 2107/27 (2024.01); G05D 2111/10 (2024.01)] | 6 Claims |
|
1. A sampling method for surface sediments in an intertidal zone based on an unmanned aerial vehicle platform, comprising:
moving an unmanned aerial vehicle to an appropriate height above a sampling position of the surface sediments in the intertidal zone, wherein the appropriate height above is preset;
equipping the unmanned aerial vehicle platform with a releaser, a winch and a surface sampler;
releasing a sampler, and inserting the sampler into seabed by gravity to obtain a sample;
taking the sample off an intertidal zone ground by a lift of the unmanned aerial vehicle, and recovering the sampler with the winch;
after recovery is completed, moving the unmanned aerial vehicle to an initial position, and after the sample is disposed, moving the unmanned aerial vehicle to a next sampling position to carry out a next sampling of the surface sediments in the intertidal zone;
before releasing the sampler, comprising acquiring an image of a sampling position, and judging the image of the sampling position; and if a judgment result shows an exposure of the surface sediments, manually operating to release the sampler; if the image shows coverage of the sampling position with seawater, manually controlling to wait, move again or return;
judging by the image of the sampling position returned by the unmanned aerial vehicle; and
after moving to the sampling position, judging whether the seawater in the intertidal zone recedes and whether the surface sediments in the intertidal zone are exposed, and correcting a height of the unmanned aerial vehicle according to a surface height; after moving to a current position, collecting an image at the current position through an image sensor installed on the unmanned aerial vehicle, and processing the image to judge whether the surface sediments are exposed; after a judgment is completed, if a judgment result shows a presence of the surface sediments, returning a release signal and a collected image, and using the sampler for sampling through a manual operation control; if the surface sediments are not present, returning the collected image to relevant personnel for review to judge whether to wait or return to the initial position or directly sample; in a process of judging the image of the surface sediments, performing detection and classification of the image by a deep learning model, wherein the classification is whether the surface is exposed, and the deep learning model selects a convolution neural network, and the convolution neural network selects a structure with a plurality of convolutional modules and fully connected layers connected; and burning a program of the deep learning model by a microprocessor mounted on the unmanned aerial vehicle for identification; if the identification shows an exposure of the surface sediments, controlling the sampler to sample; if the identification shows no exposure of the surface sediments, returning a currently judged image to the relevant personnel for review; and in this case, replacing a control model of the unmanned aerial vehicle from automatic control to manual control, and using remote manual control instructions to enable the unmanned aerial vehicle to perform behaviors of waiting, returning to the initial position, or directly sampling; wherein a sampling process needs to be at a low tide time, but a surface of the intertidal zone under the sampling position at the low tide time may not be completely exposed, and an existence of water on the surface has a certain impact on the sampling process and final sampling results; therefore, there is a need to avoid sampling when there is water in the intertidal zone as much as possible; meanwhile, in order to improve an automatic control mode of the unmanned aerial vehicle, when a judgement shows no exposure of the surface, the unmanned aerial vehicle moves to a central position of a plurality of adjacent grids for an image judgment again, and records central position information of the adjacent grids; after judging the central position of the plurality of adjacent grids, if there is a phenomenon of surface exposure, the sampler is used to sample directly; if there is still no content exposed on the surface, a control mode is handed over to remote manual control or returning directly is selected, and collected image data is transmitted to the relevant personnel for review.
|