Plant Roots Under Complex Background Image Feature Extraction Methods In The Research And Implementation

Currently, precision agriculture, precision agriculture ideas put forward for the development of agriculture opened up a new space. High-tech applied to agricultural crop production and reducing production costs and increase crop yields and improve the quality of agricultural products and reduce environmental pollution in the production of great significance.Plant root extract image data and provide technical support for these studies is an important basis for the content of the above research. The complexity and diversity of plant root root itself surroundings while making root extract has become a complex image, challenging work. Therefore, the study of plant root extract complex background image has important theoretical significance and practical value.This topic using computer vision technology and image processing methods to achieve the image feature extraction plant root complex background to complete the images captured by plant roots separated from natural scenes, as well as the feature parameter extraction plant roots. Its implementation will be further research to provide accurate data to support and efficient means for observing the development of agriculture and forestry and related research department, technology workers.The main contents of the research are: research existing image segmentation algorithms and the application of new technologies and methods of plant root complex background image segmentation, observation and comparative analysis of the experimental results obtained best practices; in the existing basis for plant roots can be simple background characteristic parameters extraction prototype system, based on a complex background processing extensions, support systems for portable use extensions,system encryption extensions, so that it can be able to be applied in practice.In this thesis, a comprehensive comparative analysis of the threshold segmentation method, edge detection and region segmentation method three traditional image segmentation method is applied to the case of complex background plant roots image segmentation. Spiking Neural Network presents a method with the traditional image segmentation method combining image segmentation roots.Image segmentation method for the study of natural scenes under multiple original images obtained and the application of traditional methods of combining SNN were extracted, and the success rate of extraction, the extraction rate and the amount ofhuman interaction and the statistical analysis. Concluded: single use traditional image segmentation method is relatively simple algorithm, extraction speed is relatively fast,but the extraction of the low success rate, can not achieve the desired results; combining SNN image segmentation method is relatively complex, relatively slow extraction, and a small amount of human interaction, but the success rate of extraction, can be applied to reality. Finally, based on the research results of this project to design and implement a practical analysis of the root system software.