2016 Fiscal Year Final Research Report
Study on a novel modeling system for non-stationary objects such as fruit trees in the outdoor environment
| Project/Area Number |
26450366
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Agricultural environmental engineering/Agricultural information engineering
|
|---|
| Research Institution | Wakayama University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2014-04-01 – 2017-03-31
|
| Keywords | スキャンマッチング / CIF特徴量 |
|---|
| Outline of Final Research Achievements |
In order to detect invariant features from multiple three dimensional data acquired under the outdoor environment, we conducted experiments using three kinds of cameras with different characteristics. As a result of investigating various conditions necessary for detecting the invariant feature, it was found that it can be used as a useful invariant feature by adjusting the generation parameters of such invariant feature which was previously proposed.
As a preparation method of robustly stitching mltiple three-dimensional data, we proposed a measure representing its shape of two-dimensional range data. We also proposed a method for robustly stitching multiple two-dimensional range data, which uses the measure to decide whether to include useful shape features of the environment in the two-dimensional range data and has our SVM classifier learn to conduct such judgement.
|
| Free Research Field |
機械学習
|
