|
Click on “Download PDF” for the PDF version or on the title for the HTML version. If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options. Image Acquisition for Site-specific Weed Monitoring Using Panoramic Annular LensPublished by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan www.asabe.org Citation: Paper number 051169, 2005 ASAE Annual Meeting . (doi: 10.13031/2013.18884) @2005Authors: Péter Ákos Mesterházi, Gábor Milics, Stephan Maniak, Károly Kacz, Miklós Neményi Keywords: precision agriculture, weed control, plant protection, PAL The practical application of ground-based site-specific applications using optical devices is set back by some typical limiting factors. In case of optical sensing of weeds, pests and other causative agents or even soil properties, there are some different theories to follow (separate weed species or not, take into account shape parameters or only the spectroscopic differences etc.). The more complex image processing is applied, the more elaborated system build-up and more computing demand are required. Besides, the problem of limited visual field still exists, what can be consequently blamed also for the inadequate operation efficiency. The authors applied a Panoramic Annular Lens (PAL) with a horizontal view angel of 360. allowing the coverage of a field part of 1 ha with a single image. The evaluation of the images is carried out in two parallel ways. After pulling the image on the surface of an inverse sphere identical with the PAL, the 1 ha area can be walked through and the weeds are present can be identified by means of visual evaluation on a given level of probability. Simultaneously, the pixels containing green plant parts are filtered and a so-called plant density value is calculated. Positioning the pixels containing plant parts and taking into account the information of the visual analysis proper site-specific weed control can be planned. The authors review their experiences in connection with the mentioned research tasks and the already reached results. (Download PDF) (Export to EndNotes)
|