| Today, it is information time and a leap development in gain, machining, processing and application of information has taken place. Along with the development of computer technology at very fast speed, digital image processing technology has gotten rapid development and comes into being new subject. Since a few years, it has been applied widely to many field including aviation, medicine etc. Today, digital image processing and identifying is the indispensable component of modern agricultural development. With specialization of digital image processing, descending of hardware cost and increasing of CPU speed, this technology is more and more important in the way of agricultural application, and is vital technology of achieving agricultural information and automation as well.Seed purity is an important target of seed quality and primary basis assessing seed grade for seller and user. Recently, along with wide application of new breed, especially hybrid breed, the status of seed is more important in agricultural producing, at the same time, some problems that are caused by seed quality are more and more serious as well, the farrago that is caused by purity of hybrid maize is to 1.5 hundred million tons every year. According to the calculation, seed purity descends 1%, the crop seed output will decrease about 135kg, although the purity of seed presents ascending trend, it is low but unsteady. These existing methods, including morphological approach, physiochemical approach and molecular biology approach, can come true the purpose of determining purity of seed, but all of them are achieved by manual determining, and some of them not only cost a lot of financial and material resources but also waste too much time. Therefore, finding a method that can determine rapidly and precisely is very important to country, agriculture and farmers as well as is necessary to achieving increasing production and harvest and agricultural modernization.We based on the scientific research item of JILIN Province science and technology office —Artificial intelligence image identifying and processing system of rapid determining purity and vigor of maize seed, and make digital image processing system of determining purity of maize seed more perfect by means of increasing variety of maize, aiming at these varieties of YEDAN12, DONGNONG248 and 478 etc., using these implements of computer, pick-up camera, image gathering card, electrophoresis equipment and home-made lighting room, synthetic applying image processing technology, statistics theory of data analysis, and combing digital image processing technology and biology approach of determining purity of seed . The thesis can be divided into three parts:(1) Research of biology approach of determining purity of maize seed Choosing the most adaptive biology approach of determining purity of maize seed by analysis to biology approach of determining seed purity, referring to many books and literatures and contrasting some experiment methods, and improving definition of electrophoresis band images by the method after ameliorating. We can receive reliable artificial determining result by the method, and make these electrophoresis band images become into image sources of computer image processing.(2) Setting up hardware system Establishing hardware system of digital image processing, and getting nicer condition of screening(3) Applying of image processing technology According to character of electrophoresis band images, we applied image processing technology that includes grey scale transformation, image enhancing, image unitary equilibrium, image partition and character pick-up etc. and ameliorated some technologies of image processing. At the time of identifying image, we must find the optimal matching template firstly, and then begin to match all the images by template matching method.(4) Setting up software system Establishing work flow of image processing, applying VC++ to program, and setting up software system of image processing and identifying.(5) Reliability research of software system Contrasting the result between manual determining and image processing determining, and verifying and adjusting image processing system.I. Research of biology approach of determining purity of maize seedWe analyzed carefully current determining methods of the seed purity, and found that protein gelatinous electrophoresis is rapid, accurate, steady and economic, not affected by environment condition. By contrast results between electrophoresis determining and field experimentation, we chose protein gelatinous electrophoresis as determining method of the seed purity, and did chemical experiments by reactive system after ameliorating of our national tentative corn protein gelatinous electrophoresis, and we would obtain electrophoresis band images more clear.II. Setting up hardware systemAccording to above determining method chosen of maize seed, we designed hardwarecomposing of image processing and identifying system that includes IBM-PIII computer, CCD inter line camera, image collection card and lighting room.III. Applying of image processing technologyFirstly, we carried out image pretreatment to electrophoresis band, and adopted weighted means to realize grey scale transformation. Secondly, in the process of gathering image, it must produce noise, otherwise, we focus on the image character that we are interested in when we do it, so we should use the technology of removing noise by smoothness and filtering, histogram equalization of image, windows conversion to enhance the image definition. Thirdly, we adopted the minimum error partition method to pick up character that we are interested in. Finally, we established normal template library, and adopted global template matching method to identify image that we need treat.IV. Setting up software systemWe made the best of these virtues of WINDOWS interface and VC++ software to program of the purity of maize seed digital image processing and identifying system, and perfected this system.V. Reliability research of software systemThrough contrasting the results between manual determining and image processing determining, we have verified and adjusted image-processing system. In the period of the purity determining, accuracy percent of the result of manual determining and image processing determining is above 95%, In the test, we found that optics distortion pick-up camera and digital camera and quality problem of seeds themselves affect these results of image processing. Primary conclusion of study is: to 18 maize seeds, utilizing image processing and identifying method to determine purity of maize seed is feasible in theory and practice.The innovations of this research are as below:(1) According to actual determining methods of seed purity, we chose protein gelatinous electrophoresis as biology approach of determining purity of maize seed, and utilized national tentative salting proteinic PAGE gelatinous electrophoresis reactive system to do chemical experiment, its result testified that we would obtain electrophoresis band images more clear by reactive system after ameliorating and provide good image data forcomputer image processing. We established hardware system of digital image processing, and got nicer condition of screening.(2) We ameliorated some technologies of image processing, and combined expressing biology character (electrophoresis band) and image processing and identifying technology to determine purity of 18 maize seeds. Furthermore, we utilized 50 pieces of electrophoresis band as image data source for computer, and obtained high accuracy by contrasting the result between computer determining and manual determining .The average accuracy is to 98.37%, individual accuracy is 94.05%.(3) We established work flow of image processing, applied VC++ to program, and set up software system of image processing and identifying. We came true rapid, accurate determining by the method of digital image processing technology, and the technology takes on advanced character and nicer popularization value. |
PDF Full Text Request