In the current standard for flue-cured tobacco, it is divided into four classes and 42 levels.
The vision technology is used to auto recognize and count. Therefore, when select the tobacco leaf grade classification standard, the classification standard should based on appearance and characteristics as the main identification. One of the classification standards relies on the color of different parts of the tobacco leaf as.Divide all possible tobacco leaf colors into eight categories.Divide the maturity of tobacco leaves into two categories.In the leaf structure, it is divided into three categories according to the degree of looseness of the leaf structure.According to the different growing environment of tobacco leaf, the oily contained in the tobacco leaft are also inconsistent.Therefore, three standards have been defined for tobacco leaf based on the oil content in tobacco leaves. Finally, three categories was divided according to the different chromatic of tobacco leaves.
The density of the arrangement of tobacco leaf cells, divided into. Loose, more loose, slightly tight, tight.
In the image taking stage, in order to get high-quality images,the linear array high-resolution imaging equipment is installed to get images.Then through image preprocessing activities, the sharpness of the image can be enhanced, the possible noise in the image can be reduced, the efficiency and accuracy of machine recognition can be improved in the later stage.Remove the image information irrelevant to the recognition in the image, such as the tobacco leaf placement table, and increase the effectiveness of the information.Also, it can simplify the image and reduce the time and cost of later machine learning.
When identifying the appearance of tobacco leaves, multiple pieces of information related to classification and grading can be extracted from one reference image.Different grade standards are delineated from different dimensions. Therefore, it is necessary to get the information in the reference image that meets the classification standard. This process is also an important process of machine learning. Through this process, the machine will match the image with the classification standard and form a database.
In the process of identifying and grading tobacco leaves, the fuzzy recognition of tobacco leaves is the first required.In the process of fuzzy recognition, the method of cluster fuzzy analysis is used,through this method,can make the analysis of the uncertain leaves more reasonable and make several images in the same group belong to the same category,so as to give a more accurate recognition. After the fuzzy recognition, can to to the recognition and grading process.
The collected images are simplified and the effective information related to the classification activities is extracted after automatic image collection.Than machine will enter the classification processing.The system will compare the image information that extracted with the grading standards,than divide the recognition target into different grading,according to the classification standards.and transmit the classification result to the user computer finally.
Product Tags:
|
|
KEYE Tobacco Leaf Automatic Vision Sorting Machine For Quality Inspection Images
|
