Research On Leather Visible Defects Detection Based On Texture Analyse

Leather is welcomed to factories and customers as a kind of good material. It has so abroad usage as used in shoes, bags, clothes and gloves, apparels, gym things, musical instrument and industry aspect. Other materials can't be compared to it. It still developed well in the financial crisis. In the process of leather machining,the effect of lookuping disfigurement and classification will impact production efficiency and product quality directly. For the purpose of using raw material effectively and avoid inferior interfusing, detecting the material and product is necessary. But for a long period of time, defecting method is mainly depending on hands, the direct aftermath is reducing production efficiency and increasing production cost. Along with the development of computer vision technical, the handwork of detecting leather can be finished by computer more effectively. Then workers can be liberated from heavy work, and cost can be cut maximally. According to the present situation. I put forward a detecting scheme for leather process.Firstly, this paper classifying materials according their different kinds using analyzing method of texture leather, which can be used in inspecting frontispiece and inverse. Because of different products coming from different kinds od leather, and arranging, incising, detecting are all done on frontispiece, so classification and inspecting technical is necessary. Via calculating the wave crest and troughs statistical features on leather scanning beam, auto-identifying and classifying can be carried out.The main process of leather machining is arranging and incising various product parts. But the surface is not perfect, and exist all kinds of defects, such as bite, scar, wound, etc. For making using of material effectively, detecting the position of defects is the first task before arranging and incising. The paper introduced an image defects detection algorithm based on gray- run length cumulation. This method attempted to utilize the information of both gray level and run- length in an image. In this approach, the gray- run length cumulation of every pixel was calculated through evaluating the differences of gray levels and run- length cumulation, and then used as threshold for segmenting the textures. Experimental results showed that it could effectively segment texture images which had different local defects, and its precision of classification was superior to that of the gray- level co- occurrence matrix and gray level- gradient co- occurrence matrix. Besides, different parts of leather have different roughness, so it is necessary to detect roughness grades to be fit for different products. The paper proposes a new very simple method to determine roughness of a surface of a leather material from its scanning electron microscopy image. The method has been used in assessment roughness of implant materials—fractal method used in nonlinear time series analysis. Last, the classification of defects is also done for the practical work.