Research On The Similarity Detection Algorithm Of Tire Pattern

The tire pattern similarity detection technology belongs to the field of image detection technology.The update and improvement of this technology will promote the standard development of the tire industry.With the rapid development of the tire industry,various companies pay more and more attention to the maintenance of intellectual property rights.Patent protection for tire patterns has received significant attention.The application of tire pattern similarity detection technology can not only prevent the pattern design result from being similar to the existing pattern,causing unnecessary disputes,but also prevent malicious infringement by competitors.Although image detection technology is widely used in the Internet,industry,agriculture and other fields,there are still many gaps in tire pattern similarity detection.There are fewer similarity detection methods designed for the characteristics of tire images.The algorithm designed in this paper mainly solves the problems such as tire pattern rotation,tire pattern repetition,and uneven illumination,etc.The specific requirement is that under the premise of accurately detecting the similarity of these tire patterns,the algorithm’s adaptability and stability to the problems of the tire pattern image itself are improved,and the tire pattern similarity detection algorithm has high practical significance.In response to this requirement,this paper first introduces the tire pattern image extraction technology,and specifically analyzes the practical application effects of several typical template matching algorithms.What’s more,the original template matching algorithm is improved so that it has better effect whenapplied totire pattern image extraction.Then,in order to reduce the interference of tire patterns such as uneven illumination and anti-skid patterns,an improved single-scale Retinex algorithm is proposed.This article introduces the Retinex algorithm and analyzes the problems of applying the single-scale Retinex algorithm to the tire pattern image.Combining bilateral filtering with the single-scale Retinex algorithm,the single-scale Retinex algorithm is improved.The improved algorithm is applied to the tire pattern image,and compared with the original algorithm.Thirdly,local binary patterns and perceptual hash algorithm are introduced for the features of high repetition,easy rotation,and similar color distribution of the tire pattern image.This paper proposes a method for deduplication of tire pattern images based on hash function,which can ensure the unity of the tire patterns for similarity detection.The traditional LBP algorithm and the hash algorithm are combined to obtain an improved LBPalgorithm that obtains the texture features of the image by rotating the invariant pattern LBP.Then it extracts the fingerprint information of the image,and obtains the similarity of these tire patterns by calculating the distance between the fingerprint information of these tirepatterns.Finally,these improved algorithms are actually applied to the tire pattern image,and the processing effect of each stage is demonstrated.The experiment proves that the tire pattern similarity detection algorithm proposed in this paper can better adapt to the characteristics of the tire image It can more accurately and quickly extract a single tire pattern image,and overcome the problems of the tire image itself.