Research On Online Detection System For Defects In Tire Belt Joints

Vehicle safety is becoming a major concern due to the increasing number of cars owned every year.Many components in the composition of the car,the tire as an essential component of the car,its quality has a vital impact on the safety of driving.Therefore,the tire defect detection has become the focus of the most attention for domestic and foreign tire manufacturers.However,the common defect detection systems are only for finished tires,and the defective products can only be reworked or downgraded,which increases the production cost and wastes resources.Defects in final tires often exist due to poor quality control in the production line.The most typical defective case is the tire belt joint defects.When the tire belt is laminated online,the joint docking is a critical factor in the safety of the tire,and its quality is the key aspect of corporate control.At present,the detection method for the quality of the tire belt joint is mainly manual and a few traditional visual detection methods.Due to the characteristics of small target and low contrast of joint defects in the tire belt,the result of manual detection is influenced by human subjective consciousness,while traditional visual detection has high error detection rate and low efficiency due to factors such as tread roughness and characteristics of joint defects.This paper proposes a computer vision and line laser-assisted online detection technology for tire belt joint defects,which can be applied to the detection of tire belt defects on the assembly line with high accuracy,high efficiency and low cost,and has an engineering significance.The main tasks are as follows.（1）Based on the actual production,we analyzed and summarized the common types of defects at the joints of the belt bundle layer and proposed a defect detection program for each category.In the centerline extraction stage,combined with the real conditions,the traditional grayscale center of gravity method is improved,which makes the extracted centerline outline more realistic to reflect the characteristics of the tire belt joint.（2）To effectively prevent the interference of belt surface roughness on the feature point extraction accuracy,a feature point extraction method combining integration and mean slope was proposed in the feature point extraction stage,this method has the advantages of traditional methods.（3）To further ensure the accuracy of the detection results,a targeted accuracy correction model was developed to weaken the influence of the transition region generated by the line laser stripes at the tire belt joints.（4）Combined with the special characteristics of the tire belt joint structure,a defect diagnosis method based on secondary sampling is proposed.Set a detection model and design experiments to choose a reasonable sampling spacing,which can ensure accuracy of detection,the efficiency of the system is also improved.（5）The experimental platform was built to verify the reliability and interference resistance of our designed defect detection scheme and the precision of the model.Experimental results show that this method is more efficient and accurate than traditional detection methods.It can meet the requirement of on-line detection of belt joint defects in tire production line.