| 摘要 |
PROBLEM TO BE SOLVED: To reduce an operational load on operators working in by-group object identifying processing by improving the accuracy of automatically identifying objects by groups. SOLUTION: A sensor 10 acquires time series input signals. A time frequency analysis processing part 20 converts the input signals into gram images. A line extraction part 30 extracts signal lines from the gram images for pre-processing characteristic element extraction. A center frequency extraction part 40, a line width extraction part 50 and a signal strength extraction part 60 perform characteristic element extraction from the line extraction part 30. A catalogue data 70 accumulates three types of characteristic elements for each object and data that the characteristic elements are multiplied by a weighting factor. Certainty factor calculation parts 80-100 calculate certainty factors for each characteristic element. A certainty factor joint part 110 calculates the certainty factor Pk of an object k in accordance with a Dempster joint rule. An object determination part 120 determines an object showing the maximum certainty factor as a target to be detected to obtain a determination result. COPYRIGHT: (C)2004,JPO
|
