A package tracking method based on vision and laser point cloud AI algorithm

Abstract

The invention discloses a package tracking method based on a vision and laser point cloud AI algorithm, which is characterized in that the data of wrapping point cloud is collected by a lidar, a clustering point cloud algorithm is used to classify the collected wrapping point cloud data, the system automatically picks up the similarities and differences of packages and annotates, the system automatically learns the massive annotated pictures, utilizes the depth convolution algorithm to establish an abstract package model, the model is optimized by the actual package, and then is auto-tuned, the step 4 is repeated until the success rate reaches 99.9 per cent. The system automatically stores the optimal package model to the client server. The invention utilizes the laser radar to collect thewrapping point cloud data, classifies the wrapping data through the clustering point cloud algorithm, automatically learns the massive labeled pictures, utilizes the depth convolution algorithm to establish the abstract wrapping model, realizes the tracking of the accuracy of the wrapping and saves the cost.

Description

technical field [0001] The invention relates to a package tracking method, in particular to a package tracking method based on vision and laser point cloud AI algorithm. Background technique [0002] Most of the current package tracking methods are realized by the cooperation of multiple sensors, and a large number of sensors are required. In the cooperation of a large number of sensors, deviations are prone to occur, so that each package cannot be accurately tracked, and a large amount of material is consumed. . Contents of the invention [0003] The purpose of the present invention is to provide a package tracking method based on vision and laser point cloud AI algorithm in order to overcome the deficiencies in the prior art. [0004] In order to achieve the above object, the present invention adopts the following scheme: [0005] A package tracking method based on vision and laser point cloud AI algorithm, characterized in that it includes: [0006] Step 1: Collect p...

AI Technical Summary

Problems solved by technology

[0002] Most of the current package tracking methods are realized by the cooperation of multiple sensors, and a large number of sensors are required. In the cooperation of a large number of sensors, deviations are prone to occur, so that each package cannot be accurately tracked, and a large amount of material is consumed.

Images