TAIWAIN, November 1, 2012 — Researchers from the Department of Bio-Industrial Mechatronics Engineering at the National Taiwan University (Taiwan, ROC) have developed a machine vision system for monitoring the growth of Boston lettuce.
Traditionally, such measurements have been performed manually, but although manual techniques are simple and reliable, they are also time consuming and laborious. In contrast, vision-based methods are nondestructive and highly efficient.
To automate the measurement procedure, the researchers employed a stereo vision system, which was composed of two off-the-shelf cameras with parallel optical axes. Images were continuously recorded by the system and from them panoramic images were constructed.
A GrabCut algorithm from the OpenCV library was used to cut the plant area out from the background and find the center of the image. Then, both 2-D and 3-D plant features — including plant volume, plant height, width, length and diameter — were extracted from the panoramic images.
The researchers believe that such a system could be used to provide useful information in many areas of horticultural research as well as in plant production processes. More information describing the system they developed can be found here.
Recent articles on agricultural imaging that you might also be interested in reading.
1. Linescan-based system counts seeds. Coleman Technologies Inc. (Newtown Square, Pa.) has developed a linescan-based imaging system that can count seeds at rates of up to 1,000 seeds/sec with an accuracy greater than 99 percent.
2. Vision system sorts tomato seedlings. A vision system installed at Westland Plantenkwekerij (Rotterdam, The Netherlands) is capable of sorting tomato seedlings at a rate of 18,000/hr.
3. Smartphone app tracks invasive plants. Bristol University (Bristol, United Kingdom) researchers have developed a new mobile phone app to help the public track the spread of invasive plants that pose a threat to biodiversity.
Source: vision-systems.com, © 2012: PennWell Corporation
If you have any questions or comments, please contact SME Public Relations at 313.425.3000, email firstname.lastname@example.org or fax: 313.425.3403.