基于图像识别的煤仓煤位监控系统的研究与设计-模式识别与智能系统专业论文.docxVIP

Subject : Research and Design of Coal Level Monitoring System of Coal Warehouses Based on Image Recognition Specialty : Pattern Recognition and Intelligent System Name ：Yu Yuanyuan (Signature) Instructor ：Ma Xianmin (Signature) ABSTRACT Depending on the size scale of coal mine, coal companies usually set a number of coal warehouses, which are mainly used to decrease the pressure of coal conveyance, storage and relay of raw coal resources. This research primarily aims to the coal warehouses on the ground, because coal warehouse monitoring is an important process in mine production. Traditional coal warehouse monitoring methods have the disadvantages such as time-consuming, labor-intensive, mechanization, lack of automatic management. This article presents the program based on image processing to realize the ground monitoring of coal warehouses in order to solve these problems. However, the storage environment of coal warehouses is poor, and is often severely affected by dust, so the picture quality of coal warehouses is poor. Therefore, image pretreatment is imperative. The main purpose of image pretreatment is mainly to remove or weaken the background and noise in the coal level monitoring, making valuable information more highlighted in the image, to fulfill the task of coal level monitoring of coal warehouse via images. This research adopts ARM processor as the main controller, which uses S3C2440A as the core chip. It collects the coal-related images of coal warehouse by ZC301-oriented chip camera and transfer them to the PC machine via the cable. After image pretreatment and the image processing that uses auto threshold and region growing algorithm, SIFT combined with k-clustering algorithm is used for image matching, the image matching here refers to images obtained after the processing on the PC and the picture of full coal warehouse stored in the PC. If it reaches a certain amount of similarity, the image is considered as full warehouse. Once this happens