Now robotic automated welding technology has become an important way to improve production efficiency, and has changed the problem of inefficient manual welding in complex environments, and the problem of inefficiency in incomprehensible environments, such as inaccurate detection of welding position and weld size, requiring continuous manual intervention and correction, and inability to meet the quality of welds.According to customers' feedback, some problems they faced in the actual welding process in a large welding process:

(1) Complex environmental structure: some environments have small space locations and complex structures, and it is difficult to observe the shape and welding parts of the weld;

((((((((((((((((2) Due to the high accuracy requirements of material: in the face of some high-strength or characteristic alloy materials, their requirements for welding heat input and weld quality are high, and thermal deformation and stress concentration during welding are required to be precisely controlled;

(((((((((((((((3) Manual intervention causes low efficiency: When encountering some larger sizes, their welding paths and welding state quality cannot be understood in real time, and manual participation is often required, resulting in problems such as difficult connection, low efficiency, and time-consuming;

In response to the above problems, Yunji Technology takes advantage of its own visual application development advantages to dig deep into customer needs. The core of the weld visual tracking technology developed is the real-time identification of weld features and its feature extraction. Only by accurately identifying different types of weld features and extracting them into data information recognized by the welding robot can we achieve no manual intervention during the welding process, so that the welding robot can adjust the real-time trajectory according to the weld form by itself, so as to adapt to different welding occasions and achieve rapid welding production in small batches, multiple occasions, and multiple tasks.

Practical application of visual tracking of welds

Yunkey Technology uses a gigabit network industrial camera as the weld welding visual guidance, and combines a customized visual tracking system and visual algorithm to effectively solve this problem.This machine vision tracking system is a tracking system that takes full advantage of the high flexibility of welding robots and uses laser displacement sensors and cloud key technology industrial cameras to automatically identify gaps and guide welding.It has the following advantages: (1) High visual recognition accuracy, and can obtain the size and position changes of the weld in real time; (2) Timely adjustments in the face of changes in the weld to ensure the quality of the welding process; (3) No manual intervention is required to adapt to various complex environments; (4) High recognition rate, low error rate, and improve production efficiency; (5) Full visual monitoring to understand the welding situation in real time;