Research And Application Of Detection Algorithms For A Whole Piece Of Leather Surface Defects

With the wide application of leather products in daily life,the consumption demand of leather manufacturing industry is increasing day by day,and the quality and efficiency requirements of production and processing in leather manufacturing industry are also higher and higher.Because the leather raw material takes the auto animal skin(mainly cowhide,sheep hide,pigskin,etc.),there are inevitably various defects in the raw material during the animal growth period and the artificial slaughter,so the leather industry needs to quickly and effectively detect the defect areas in the raw material processing and manufacturing.For a long time,China's leather products industry mainly relies on manual defect detection and location process for the whole leather.However,due to the influence of illumination conditions,different Workers' experience and changes in mood,physical strength,working hours and other factors,it is easy to cause problems such as low efficiency of detection and layout cutting.Therefore,the research of automatic detection method for surface defects of whole leather is of great significance to improve the production efficiency of leather industry and reduce the production cost.At present,researchers in China and abroad have proposed many different methods for the detection of leather surface defects,but most of them are aimed at the detection and classification of certain surface defects of leather,and fail to effectively detect and locate the defects of the whole leather.Therefore,based on digital image processing technology,this paper studies and applies the algorithm of high-definition leather surface defect detection.The main research work and innovation of this paper include the following points.In order to improve the accuracy of leather defect detection,it is necessary to preprocess the leather image.By analyzing the characteristics of natural random and defective areas of leather images,the contrast of leather images was enhanced and image denoising was carried out.When the contrast of the whole leather image is enhanced,the leather region(ROI region)should be extracted first,and the contrast of the partial defect leather image can be enhanced directly.In this paper,according to the characteristics ofleather image,the good methods of image contrast enhancement and denoising are summarized through experimental comparison.On extracting the whole piece of leather leather area(ROI)stage of the image in this paper,based on the saturation of the whole piece of leather effective region extraction method and based on the global significance of the whole piece of leather effective region extraction method,and then to the proposed two methods of comparing with the traditional method of experiment results,through the experiment to the proposed two algorithms in both the extraction efficiency of algorithm and the accuracy of are more advantages than the traditional method,based on the global significance of the whole piece of leather is used to effectively region extraction method in time slightly ahead of the effective area from the whole piece of leather method based on saturation,But the latter can extract the effective area of the whole leather with higher accuracy under the influence of complex conditions.In the stage of defect detection,this paper is divided into two parts: local defect detection of leather and whole leather defect detection.The analysis shows that the leather image is a natural texture image with irregular shape and great difference in defect types.Therefore,a method of leather defect detection based on image saliency is proposed in the stage of local defect detection.In this paper,the saliency model which is most suitable for leather image is obtained through experiments,and the saliency region is detected in the defect image,and the defect part is separated by adaptive threshold segmentation method.In the whole leather defect detection stage,a method based on edge detection was proposed to detect the whole leather surface defect,and the defect grade was classified by the significance of multi-scale extreme value.