Mechanical assembly image segmentation method and equipment based on deep learning network

Abstract

The invention relates to a mechanical assembly image segmentation method based on a deep learning network. The method comprises the following steps: constructing a mechanical assembly lightweight semantic segmentation model composed of an encoder network and a decoder network through the deep learning network; establishing a mechanical assembly data set; inputting a depth image in the mechanical assembly data set into an encoder network for feature extraction to obtain a feature map with high information quality; inputting the feature map with high information quality into a decoder network, recovering the size of the image and carrying out feature fusion to obtain a multi-dimensional segmentation map; updating parameters in the encoder network according to an error between the multi-dimensional segmentation image and the corresponding label image; iteratively executing the above steps by using the depth images in the training set until a preset number of training times is reached; outputting a lightweight semantic segmentation model of the mechanical assembly after testing; and carrying out image segmentation by utilizing the trained lightweight semantic segmentation model of themechanical assembly body to segment each part of the mechanical assembly body.

Description

technical field [0001] The invention relates to a method and equipment for image segmentation of a mechanical assembly based on a deep learning network, and belongs to the technical fields of computer image recognition and intelligent manufacturing. Background technique [0002] The current manufacturing industry is ushering in the production stage of mass customization, and the continuous change of product types will increase the difficulty of manual product assembly. In the assembly process of complex mechanical products, once the errors in the assembly process (such as wrong assembly sequence, missed assembly, wrong assembly, etc.) cannot be detected in time, the quality and assembly efficiency of mechanical products will be directly affected. Mechanical assembly image segmentation can accurately segment the outline and area of ​​the assembled parts of the mechanical assembly from the mechanical assembly image, obtain relevant information of the mechanical assembly, and f...

AI Technical Summary

Problems solved by technology

[0004] There are few researches on the application of semantic segmentation in the field of machinery, especially in the field of mechanical assembly. The main reasons are as follows: (1) The structure of mechanical products is complex, and the occlusion between parts is serious, which will lead to inaccurate segmentation of mechanical assemblies

(2) The color and texture information of mechanical parts is single, which increases the difficulty of segmentation of mechanical assemblies

(3) In the mechanical assembly segmentation task, there is a lack of dedicated labeled datasets, and the quality of the labeled datasets often determines the quality of the segmentation model

Images