Study On The Husking Rate Of Corn Combine Harvester Husking Device Based On Color Image Segmentation

Peeling rate is an important parameter of the performance of corn combine, which canintuitively and effectively reflect the working condition of peeling device. So rapidly andaccurately detecting the peeling rate of the corn combine peeling device plays an important role.With the rapid development of information technology, image processing techniques areincreasingly applied to agriculture. When using gray level image processing, the original imageinformation effectively often loss too much; Along with the rapid increase of the computerinformation processing ability and the precision of image acquisition equipment, color imageprocessing technology obtained the unprecedented attention by more longer retaining theoriginal image information effectively. So applying color image processing technology to thestudy of peeling rate of corn combine harvester peeling device have been a new method to studythe peeling rate.This paper studies the peeling rate of the corn combine harvest machine peeling devicebased on color image segmentation technique. First, the hardware device of image acquisitionhas been build. Manual industrial is installed on the digital industrial camera lens, and digitalcamera collection transmission card is plug-in PC motherboard slot. Then link the digitalindustrial camera to the PC through a reticle. After building the hardware acquisition device,operate the image acquisition software, and then adjust the focus ring and aperture of the cameralens. The pictures of ear of corn and residual bract after peeling are deposited in PC with JEPGformat.Then preprocess the images and enhance the images with the method of color imagehistogram equalization. In RGB color space, conversion the artwork master to R、G、Bcomponents, and proceed histogram equalization to the three-component respectively. By thehistogram average after the three-component to reconstruct an RGB image. Vector medianfiltering method is used on the original image denoising.Use improved k-means clustering segmentation method to proceed color imagesegmentation to the images after preprocessing. Use HSV model transformation to the images,and determine H component histogram. According to the number of the histogram peaksestimate the clustering number. Then L*a*b*model transformation is proceed to the images.According to the estimated number of clustering, an automatic clustering center is obtained.Clustering segmentation is taken to the images, and color image segmentation of the ear of cornand residual corn husks is obtained. Finally, gray process has proceeded respectively to the ear of corn and residual corn husksafter segmentation, then is the binarization processing respectively. To get the image binarizationimage area connected and tags, wipe off the target object, statistic the corn grain numberstripped clean and after peeling. The ratio of the two is the traditional corn peeling rate.Meanwhile, calculate the pixel percent of the residual corn bract and the pixel percent of thecorn ear. Then, calculate the radio of the pixel percent of the corn ear and the two above toillustrate the extent of peeling of the corn. Devised a system interface for the calculation of netrate, the interface can be intuitive and accurately display the main image processing and thecalculation results.The experimental results proved that the method proposed by this paper can effectivelyobtained the peeling rate of corn united harvest machine peeling device, but because of theinfluence of the segmentation, stripping net rate calculation results appeared deviation.