根長測定への画像処理の応用
書誌事項
- タイトル別名
-
- An application of computer graphics for root length measurement.
- コンチョウ ソクテイ エ ノ ガゾウ ショリ ノ オウヨウ
この論文をさがす
説明
There has been much descriptive work carried on plant root systems in the past, but quantitative studies have lagged behind equivalent work on plant tops, probably because of the practical difficulties of measuring root lengths. Root length has normally been measured by excavating soil and washing out roots, and then using a map measuring wheel, a ruler or the Newman method on the cleaned roots or photograph of them. The measurement methods of root length are very simple, but they require a great deal of labor and a lot of time.<br>The technique of image analysis has come to be more widely used in recent years. The measurement of the root length is made by using an image analyzing computer. The basic principle employed here is after the Newman method. It relates root length to intersections between roots and randomly oriented straight lines according to:<br>R=πNA/2H<br>where R is an estimated root length in an area A, N is the number of intersections between roots and randomly oriented straight lines, and H is the total length of the lines.<br>A root photographic image which is put into a computer, is separated into the “black” roots and the “white” background. The randomly oriented straight lines are replaced by scanning lines of image analyzing computer.<br>The results obtained by the present experiments can be summarized as follows:<br>The estimated value of a sample differed with the direction of the scanning lines. This variation depends on that the Newman method holds in a strict sense only when the scanning is done in a randomly oriented manner. The inaccuracy which comes from using a computer can be improved by rotating a sample image which is taken in the computer.<br>The accuracy of this system is related to the distance between the scanning lines. The narrower the distance is, the closer the measured value approaches the true value.<br>When we adopt this measuring method using a computer or hand-operated, it is necessary that the following conditions be satisfied.<br>1) In the process of measuring a sample, the number of rotation times of the sample must change more than twice. However two rotation times of the sample are adequate for the accuracy of the measurement and in practical application. Therefore the length of the sample can be estimated at the average of 2 times measurements taken after rotating the sample 90°.<br>2) The distance between scanning lines must be less than 10mm.
収録刊行物
-
- 農業気象
-
農業気象 44 (1), 1-6, 1988
日本農業気象学会
- Tweet
詳細情報 詳細情報について
-
- CRID
- 1390282679646959744
-
- NII論文ID
- 130003665490
-
- NII書誌ID
- AA11530034
-
- ISSN
- 18810136
- 00218588
-
- NDL書誌ID
- 3186341
-
- 本文言語コード
- ja
-
- 資料種別
- journal article
-
- データソース種別
-
- JaLC
- IRDB
- NDLサーチ
- Crossref
- CiNii Articles
- OpenAIRE
-
- 抄録ライセンスフラグ
- 使用不可