Leather Cutting Cam System For Image Segmentation Problem

With the development of modem industry and the increase of people's living standard, there are more needs to leather products. The traditional leather producing craftwork which is inefficient and has a high cost has been the major obstacle to the increase of leather product output. To promote the efficiency and decrease the cost, Leather Incising CAM has been developed which is an inosculation of computer technology and the traditional leather producing craftwork. To extract the object curve in the leather image is the key to the system and it is a typical image segmentation problem.Leather surface has a specialty. With the specialty, there will be glisten on leather whatever lamp-house is used. The brightness of the glisten can be very strong and can be stronger than that of object curve. It makes it impossible to solve the leather image segmentation problem by the algorithms based on brightness information. Therefore, color information is to be used to solve this problem.The article analyzed nine characters of several typical leather images. They are Y, H, S, I, I1, I2, I3,I2 'and I3'. Y is usually used in gray image segmentation. H, S, I come from the color model shortening as HSI. I1, I2, I3 are the characters of I1I2I3 color model. I2',I3' is the transmogrification of I2,I3. The output of the experiment shows that three characters- S, I2,I2' have better effect when used to segment leather image.Based on the conclusion and known algorithms, four algorithms are put forward .To compare the validities of the four algorithms, an algorithm evaluating experiment is designed and performed. The experiment result is that the multi-dimensional thresholding algorithm using I1 I2' is most applicable to the CAM system.The selected algorithm is implemented in the software of the CAM system. After the system applied to practice, it is proved that the algorithm has solved the leather image segmentation problem.