| Along with the continuously development and maturity of computer software and hardware, image processing technology is not only widely used in daily life but also plays an important role in the crop growth condition and quality's inspection in the greenhouse. Image processing technology is applied to the image acquired to transform or analysis it for getting useful information contained in it. These can satisfy people's need in agricultural produce and other sides. It is inevitable trend of coming agricultural developments.Vegetables contain plenty of necessary microelements, which have significant effects on human healthy. In China, vegetable consumption quantity is big, so vegetable quality is paid more attention by people. Seedling is basic for the production of plants, and it is beneficial for improving quality and output of vegetable. Seedling under protected condition can not only enhance the quality greatly, but also realize the commercialization and industrialization of vegetable seedling production.Since the advanced technologies have been introduced into protected agriculture, vegetable seedlings, whose planting purpose is transplanting, especially for melon, their productions have achieved rapidly and promising developments. It's a huge challenge for agricultural researchers to recognize the seedling healthy condition accurately, furthermore, carry out unhealthy seedling filtration and separation and the healthy seedling extraction.This reserch focuses on these problems as follows:1. Design a bracket, which is applied to place the Image Collection Equipment (ICE). And the hardware system of Melon's Healthy Seedling Identify System (MHSIS) is established.2. The combination of MV-VS078FC industrial digital camera and M0814-MP industrial lens are chosen to be as the system ICE, and the MHSIS is completed on the basis of SDK.3. The basic principles of image processing and identifying are introduced and the preprocessing of melon seedling image is realized. Several kinds of basic arithmetic of image enhancement and image division are also achieved. The arithmetic of mathematical morphology into image processing is made use, and scolor image histogram is shown correctly. Color and shape character of seedling leaves are extracted successfully finally.4. Cucumber and muskmelon seedling are both planted in greenhouse; and MHSIS is used to collect images'characteristic parameters, the images are collected at two stages, whicn include a few days after seedling is planted and a few days before seedling is transplanted. Microsoft's database software of Access is adopted to establish a Character Database (CD), and the CD can be used as MHSIS data source. At the same time connection methods between VC++ and CD is studied and the method of ADO is chosen to realize connection between foreground application of MHSIS and background application of CD. As a result of this designing, the system can realize the basic function of data citing, insert, modification and so on.I use the image acquiring equipment to get eight color digital images, and analysed the laminae's characters, and finally established the following rules to distinguish the seedling's healthy condition.(1) No matter cucumber's seedling and melon's seedling, it is effective to distinguish seedling's healthy condition through color characters and figure characters. We can establish recognition rules which use figure characters as primary ways and color characters are used as assistant.(2) For color characters, the distributing character of green ponderance is evidence. It is basically to deem healthy if laminae's green ponderance is distributed between 90-200. Howerver, the characters of red ponderance and blue ponderance are not evidence.(3) For figure characters, we can use area character and perimeter character to distinguish seedling's healthy condition.Laminae's erea which between 180-300 (pixels) is basically as healthy, and Laminae's perimeter which between 80-200 (pixels) is basically as healthy. |
PDF Full Text Request