Fluid suspended matter classification system and method based on vortex flow filter and 3D U-Net network

Abstract

The invention provides a fluid suspended matter classification system and method based on a vortex flow filter and a 3D U-Net network. The fluid suspended matter classification system comprises a collection module, a filtering module, an analysis module and a screening module. The collection module comprises a screen filter, a trawl and a collection box which are sequentially arranged, and collecting holes are formed in the screen filter; the filtering module is the vortex flow filter connected with the collection box through a water pump I, and is separately connected with the analysis moduleand the screening module; the analysis module comprises a semitransparent guide pipe, an image acquisition device and an image analysis device, the semitransparent guide pipe is connected with an overflow port of the vortex flow filter through a water pump II, and the image acquisition device acquires image information of particles in the semitransparent guide pipe and then sends the image information to the image analysis device; and the image analysis device analyzes and calculates the image information by adopting the trained 3D U-Net network, and outputs attribute information of images; and the screening module comprises a plurality of sorting containers.

Description

technical field [0001] The invention relates to a system and method for classifying suspended matter in fluid, in particular to a system and method for classifying suspended matter in fluid based on a vortex filter and a 3D U-Net network. Background technique [0002] The collection and classification of suspended solids in the open system is generally completed in two stages. First, the collection stage. In this stage, the target objects are collected from the open system by manual sampling technology, classified, and then collected. Transported to a laboratory specializing in this type of work for assay analysis, the laboratory must perform extensive preprocessing on each object and filter to obtain the target object, which is then analyzed for each target object. [0003] Current solutions are error-prone and severely limited in terms of efficiency, coverage, and analysis. For example, determining the exact coordinates of thousands of individual objects in an area is ver...

AI Technical Summary

Problems solved by technology

[0003] Current solutions are error-prone and severely limited in terms of efficiency, coverage, and analysis

For example, determining the exact coordinates of thousands of individual objects in an area is very difficult if done manually

Images