1. Optimized light source selection and lighting concept, deeply restoring the end face contour 

2. Powerful logic of pre-welding judgment, welds coordinates is acquired accurately 

3. The algorithm is flexible, compatible for multiple end face status from various products

4. Supports welding processes both at the same or different positions, with a strong capability in adapting to work conditions 

Machine vision technology is one of the necessary technical means to achieve industrial automation, intelligence and digitalization. It is also an important technology in the Intelligent Manufacturing Era and Industrial 4.0 Era. Leveraging the combination of software & hardware to perform analysis & processing on the object under test and converting it into digital signals, the technology can provide guidance and assistance during the execution procedures of the industrial equipment, and it is often referred to as the Eye of Industrial 4.0. 

The pre-welding Visual Guide System newly developed by Haosen is exactly an exemplar application of the machine vision technology in the intelligent manufacturing welding area.

Welding is one of the key processes in the production of Hair-pin motor stators. Under the influence of precedent work process, before the welding process starts, the copper wire at the welds of the stator might be in various conditions such as being not smooth at the end faces, being irregular in the contour, having gaps between copper wires or having deviations, which is prone to incur errors for the visual guide in aspects of welds coordinates, pre-welding logical judgment and etc., and which in turn will affect the OK rate of the welding. This is not only a painful issue for the industry but also a technically difficult issue.

Haosen Visual Guide System adopts a vision algorithm independently developed by Haosen, deeply integrates a Hair-pin motor welding system, sifts the welds out that are qualified for welding by means of judgment on area, clearance and deviation, and it can swiftly & accurately feedback the positions of welds to the welding mechanism. With ＜50ms  checking time for a single point, it can achieve an accuracy of 0.1mm for spot welding and ensure the welding OK rate.

Four major breakthroughs of Haosen Visual Guide System in the aspects of industry technology :

I. On the hardware side, it selects the optimized light source type and lighting concept, and the vision effects deeply restoring the real contour of the copper wire end face at the welds, thus avoiding the recognition error and guidance error that are caused by the instability of the vision contour.

Comparison between the vision effects produced from Haosen Visual Identity System (Right) and other vision equipment 

II. The software algorithm is having powerful logic of pre-welding judgment; it outruns a simple function of recognition to help the customer make a decision. In case the copper wires are not in an ideal condition, the vision equipment will normally tell it as not proper for welding. However, is it truly so in the scenario of an actual application? By ways of multi-dimensional analysis and computing, Haosen Visual Identity System will precisely tell if the copper wires are ready for welding. This is not only saving time for the decision-making of the customer and for the stator repairing, but also providing an escort for the welding process. Meanwhile, the acquisition of welds coordinates is rather different than traditionally risky rim-locating methods and is not disturbed by the changes of wire contour.

III. It is compatible for various conditions such as dents or burrs that exist on the copper wire end faces at the welding positions, or the contours are in various shapes. The vision algorithm is flexibly open for parameter settings, which is adaptable for the welding process requirements in various products status, bringing more choices to the customer and improving the welding OK rate to a maximum extent.

IV. It has a much stronger capability in adapting to work conditions and supports welding processes both at the same or different positions. Considering the relatively complicated work conditions at the stations, and the scenarios where space is limited for the installation of vision system, Haosen Visual Identity System allows different positions for the visual photo-taking and the welding process, i.e. the actuation of vision mechanism and welding mechanism at different positions. During the R&D process, Haosen renders multiple pairing solutions each specific to a unique work condition, and the customer is free to choose from process recipes based on the work condition requirements.

At present time, Haosen Visual Guide System has been successfully applied to the welding of Hair-pin motor stators, and the pre-welding visual guide has achieved a success rate of 99.9%, which is greatly improving the reliability of welding. The advanced visual guide technology can be applied to various scenarios and it can also be of help to the automated and highly precise welding in the field of battery welding.

Apart from visual guide, machine vision technology is applied to all business areas of Haosen, such as workpiece fuzzy gripping, information recognition and effect inspection. The vision technology team of Haosen is performing R&D tasks specifically meeting the various demands from the industry customers, is rich in business experience and technological accumulation, and is playing a leading role in the industry.

With the deepening of industrial automation and intelligence, the customers are bringing forward increasingly high requirements on product safety, performance and quality. The applicable scenarios for machine vision will be much wider. Haosen will continuously make progress, will give full play to the advantages of its technical teams, will strengthen its capability in competition and will lead the way at the frontier of new technologies and new demands.