Sand dredger sand excavation monitoring method based on deep learning

Abstract

The invention discloses a sand dredger sand excavation monitoring method based on deep learning, and the method comprises the following steps: obtaining a sand excavation monitoring video of a tipping bucket type sand dredger, and processing a to-be-recognized video to obtain an image; constructing a sand excavation detection model and a sand excavation tracking model of the sand dredger based on a Pytorch framework; training the sand excavation detection model; transplanting the trained model to an edge calculation module; an edge calculation module is used for controlling a camera to obtain a sand excavation real-time monitoring video of the tipping bucket type sand dredger, target detection and recognition of a sand excavation hopper are carried out on the sand excavation real-time monitoring video, and then target tracking is carried out on the recognized sand hopper; the sand excavation monitoring device can detect the number of sand excavation hoppers of the sand excavation ship within a period of time, identify the filling condition of each sand excavation hopper, judge whether the sand excavation hoppers are full or empty, perform target tracking on the identified sand hoppers and count the number of the sand excavation hoppers within a period of time. The sand excavation condition and quantity are monitored and counted, so that sand excavation management becomes more scientific, efficient and normative.

Description

technical field [0001] The invention relates to the technical field of deep learning target detection and tracking, in particular to a sand mining monitoring method for a sand mining ship based on deep learning. Background technique [0002] River sand exists in rivers and is an important material basis for building river ecosystems, which is conducive to maintaining river stability and sand balance. And river sandstone is an important natural resource, which can be used in infrastructure construction such as building construction, road bridge, embankment construction, etc., and has very important value in facility construction. Therefore, the development and utilization of river sandstone is very important in engineering construction, and the mining of river sandstone is an indispensable link. With the rapid development of society, the growth rate of engineering construction and the demand for river sand is getting faster and faster, the contradiction between supply and de...

AI Technical Summary

Problems solved by technology

[0003] At present, the existing dynamic monitoring system for sand mining in rivers in China only installs an ordinary video system to monitor the sand mining situation on the tipping bucket sand mining ship, and uses the vibration frequency of the sand mining engine sent back by 4G / GPRS to judge the current state of the sand mining ship. Sand mining status, but cannot accurately measure sand mining volume

Images