The Tissue Culture Seedling Recognition Technology Research For Transplanting Process

With the continuous development of plant tissue culture technology,the production scale of plant tissue culture seedlings expands unceasingly.it is necessary to transplant the individual plantlets into new culture flasks at various phases of their development in order to provide them with a larger growing space and a more suitable nutrient medium to support their further development.However,the transplantation task is still performed manually at present in Taiwan.This not only increases the cost of the production process,but also means that the quality of the transplantation operation is almost entirely dependent upon the skill of the operator performing the transplantation process.Consequently,a requirement exists for an automatic transplantation system to lower the production cost and to improve the quality of the transplantation process.In the process of cultivating bottle of seedling transplant,seedling is unordered,cross and conglutinate.It is necessary that the automatic transplanting equipment has the function of visual identity to accurately capture strains for transplant.Therefore in the process of this topic in view of the culture bottle seedling transplanting seedlings strains identification technology were studied.By analysising of the morphological characteristics of the single tissue culture seedling and combining with the traditional manual process of transplanting the tissue culture seedling,designing the whole process of the tissue culture seedling automatic transplant scheme.In order to correctly identify the target images,the images need be segmented.Through in-depth study related knowledge of image segmentation algorithm,the method based on color features for image gray processing,combined with the automatic threshold segmentation method realizes the target and background segmentation.In this paper,through analysis of many strains of disordered tissue culture seedling cross adhesion type,and analyzes the mild adhesion tissue culture seedling separation method.And through the image enhancement method to improve image quality.Preparing the way for further identification.Analysising the importance of the tissue culture seedling location information in the process of transplanting.In order to obtain the position information of plant seedlings,the first determine the axis of the tissue culture seedling as the characteristics of the stem.By adopting the method which based on edge detection of the seedlings in plant stem as the diameter information,and determine whether to plant seedling,and to determine the location information.The method based on skeleton algorithm for seedling plant location information was analyzed.Through a lot of refinement algorithm to get the skeleton of a large number of redundant frame of analysis,put forward a simple and effective skeleton and pruning method,which can remove unnecessary interference information and according to the length of the skeleton information acquisition seedling plant location information.Finally,this article has carried on the seedling plant location identification experiment.The result was dealed by the statistical analysis,and by comparing the effect of two kinds of feature extraction algorithm,the error questions was summaried.