| Project Nr.: | 2.1.1.2 |
| Title: | Reliability model for metal detector evaluation. |
| Description: | A proposal to investigate the application of reliability measures developed in NDT (Non Destructive Testing) to the detection of mines with metal detectors. |
| Aim: | To design a standard way of making statistical evaluations of metal detector performance (in particular, detection reliability). |
| Request: | Users of metal detectors |
| Category: | Detection - Metal detector |
| Type: | Methodology |
| Equipment: | Different types of metal detectors |
| Development: | N/A |
| Time Frame: | 2003-01-01 to 2004-05-31 |
| Place: | Germany: BAM (Berlin), Bundeswehr Technical Center for Protective and Special Technologies WTD 52 (Oberjettenberg), European Commission - Joint Research Centre (EC/JRC), Croatia: Center for Testing, Development and Training (HCR - CTRO). |
| Lead Nation: | Germany/BAM |
| Partners: | European Commission - Joint Research Centre (EC/JRC), CROMAC - Center for Testing, Development and Training (HCR - CTRO) |
| Point of contact: | Not available |
| |
| Web site: | Not available |
| Status: | Archived |
| Comments: | The project started in spring 2003 and included three series of metal detector trials. The last trial was carried out in the German Army test facilities of Oberjettenberg (WTD 52) in November 2003.A workshop was held in December 2003 to discuss the results of the trials. Preliminary results were published in the article Proposals for Performance Demonstration and Modular Reliability Assessment for Humanitarian Demining and the final results were summarised in a Journal of Mine Action article and a SPIE 2004 Conference paper. The project was archived in 2004. A continuation of this project (ITEP Project 2.1.1.8) started in April 2005 |
| Results/Conclusions: | The final report includes a series of lessons learned for planning and executing reliability trials of metal detectors. The general conclusions state the following:Some adaptations to the reliability testing section of the CWA 14747:2003 are recommended, most importantly the orthogonal design of the test in combination with the use of Receiver Operator Curve (ROC) diagrams and Probability of Detection (POD) curves.No significant differences were found between the detection performance of either time-domain and frequency domain detectors or between specimens of the same detector model.Experienced, skilled and active deminers should operate the detectors during the testing.Simple approximate formulas for error bars are suitable for statistical representation of the field data.The purpose of the reliability test is a comparison between detectors and not an estimate of the true performance of the detectors in the field.True performance in the field is higher than during the reliability tests, mainly due to a series of human factors. |
