Research On Surface Defects Matching Algorithm Of Navel Orange Based On Compressed Sensing

Fruit surface defect is one of the most significant factors affecting its quality and price.At present,most of the fruit quality classification relies on the traditional artificial sorting methods are not only labor intensity,low efficiency,and the classification of the results are not reliable,which can not meet the needs of modern logistics of fruit production automatic processing.In recent years,the Compressed Sensing theory is used in the field of image processing and so on widely.Image matching algorithm based on Compressed Sensing and Speeded Up Robust Features(SURF)has gradually become a research hotspot in the field of image processing,and it focus on how to improve image matching accuracy and efficiency to maintain the real-time display.At present,it has not yet achieved satisfactory results.This paper combine machine vision technology and Compressed Sensing(CS)technology,study a rapid and accurate method on navel orange surface defect detection based on SURF to realize the image feature point matching more accurate and faster,and the matching effect is better,therefore the classification of navel orange is more quickly and screen out all unqualified navel orange.The main research work of this paper is as follows:This paper first introduces the basic theory of compressed sensing and its framework in detail,including the basic contents of signal's sparse representation,observation matrix and reconstruction algorithm,then analyzes and illustrates its application in feature extraction and matching.Second,The theory of image preprocessing and image matching are elaborated.After that,the commonly used image preprocessing techniques and several classical image matching algorithms are illustrated in detail.Third,the theoretical knowledge and process of SURF algorithm are introduced,then the author summarizes several commonly used local feature matching algorithms,at the same time,compares their advantages and disadvantages.A fast SURF image matching algorithm of navel orange surface defect which combines Wavelet transform and CS is proposed,The simulation results of Matlab show that the number of feature points detected and matched by the improved algorithm is less than that of the traditional algorithm,which improves the problem of error matching and low efficiency greatly.Improves the speed of navel orange defect image detection,provides a new means for the real-time online classification of navel orange.