Font Size: a A A

Research On Strawberry Garden Environmental Regulate And Control Algorithm Based On ARM

Posted on:2014-11-10Degree:MasterType:Thesis
Country:ChinaCandidate:F M YangFull Text:PDF
GTID:2268330422453237Subject:Control engineering
Abstract/Summary:PDF Full Text Request
The growth of crops, environmental and biological issues will be severelyaffected by herbicides.To reduce the pollution of herbicides on crops and theenvironment.this paper mainly researched strawberry leaves and weeds identificationand the control of the greenhouse environment of strawberry Park,based on the homeand abroad on weeds and crop characteristics extraction and recognition,combined theautomatic control technology with ARM core.The main contents of the study could bebriefly summarized as follows:1. image segmentation.This paper focuses on how to identify plant from soilbackground in plant identification, the image obtained by the usual plants-background segmentation methods (such as ExG, Modified Hue) have the poor qualityof and time consuming,according to the segmentation results and the expended timeof those algorithms. A new background segmentation method is proposed which takespixel wise and to judge segmentation affect. In this approach, take advantage of theinterrelationship among the R-G-B channels of each pixel belonging to the plant andnon-plant region, The aim is to find possible complementary classification criteria todelineate plants from the soil and residual background. The experimental resultsshowed that the method can improve the accuracy and efficiency of segmentation.2. The various features combined with color, shape and texture was used toidentify weeds.This method transformed the RGB images into HIS image andextracted the H and S components of order3color characteristics, extracted thegeometry of features such as perimeter, area, circularity, dispersion and seven Huinvariant moment features,then extracted the eight texture features compose ofsecond-order, correlation, contrast, the same nature, mean, standard deviation,consistency, entropy. The effective integration of these characteristics, selectingSupport Vector Machine (SVM) as classifier, using SVM for classification andidentification of multiple images,can increase the rate of recognition up to87.5%.3. The transplantation of program and the control of temperature and humidity.Transplanting the recognition algorithm developed on PC to embedded system basedon ARM processor,designing temperature and humidity acquisition circuit, alarmcircuit, LCD circuit and programming,etc,achieved the regulation of temperature andhumidity on Strawberry Par、k.
Keywords/Search Tags:image segmentation, feature extraction, svm, weed identification, temperat-ure and humidity control
PDF Full Text Request
Related items