The Application And Research Of Texture Analysis Technology In Tanning Leather Auto Classification System

Our country is the leather industrialized nation, but leather industry production technology in our country still paused in the mechanized stage. It mainly relies on the sense organ appraisal from preparation before leather is tanned to reorganize the stage in leather production process that will be very difficult to guarantee the uniformity and the stability of the product. It is a challenge for the leather industry to change experienced control to numerical control. Uses of the sense organ appraisal in the leather production process are mainly to judge and recognize grain, texture and textile fiber arrangement under the external force of leather. What the method obtains is the visual picture information, with difficult objectivity and accuracy, therefore we can adopt the picture processing technology to transform the leather visual information into the numerical information and realize automation in leather production process.In order to use the texture analysis technology to realize tanning leather classification effectively, the paper presents a series of pre-processing to chrome tanning and vegetable tanning leathers. Pictures in the industrial production are not only affected by noises as a result of illumination and equipment, but it also may be polluted and destructed by a variety of factors which cannot be known in advance. Therefore, the paper abandons color information of tanning leathers. Because the difference of tanning leathers is very small, histogram equalization certainly cannot achieve the most superior enhancement effect, overall gray level transformations is selected to carry on the enhancement and is proved its accuracy through experiments. According to repeated experimental comparison of every kind of filter, the paper selects wavelet packet to filter noises in order to improve SNR of tanning leather pictures.The paper proposes the automatic discrimination system of tanning leather in the industrial production for the first time with four kinds of textures analysis technology used to classify tanning leathers. To begin with, the best parameters of gray level co-occurrence matrix are analyzed and determined in the paper, and the five features which can reflect tanning leathers'characteristics effectively are extracted. Furthermore, the paper improves statistical geometrical features and introduces the concept of average density. Experiments prove that the improved statistical geometry features not only discriminate tanning leathers effectively, it also greatly reduces computation cost compared with the original method. What's more, the paper applies fractal dimension and lacunarity to arrive tanning leathers classification and obtained a satisfactory effect. At last, in term of pore difference, the paper proposes a method of feature extraction based on pore shape of tanning leathers .Even if folded chrome tanning leathers are joined to disturb classification, the tentative data still indicate that results of the method based on pore shape are more stable and have a better performance than others to four picture sets with different number of tanning leather pictures..In summary, the paper researches automatic classification of tanning leathers using texture analysis technology. Because texture of tanning leather belongs to stochastic and natural texture, some texture analysis methods certainly cannot be effective to carry on the classification to tanning leathers. In view of features of tanning leathers, the paper not only carries on optimization and improvement to classical texture analysis technologies, and proposes a new texture analysis method to tanning leathers which can perform a high recognition rate. It has the certain instructive significance to develop automation of leather industry.