研究目的
Investigating the use of morphological closing operation and morphological features restriction for accurate detection of foreign objects on overhead line towers pictured by UAV.
研究成果
The proposed algorithm can recognize common foreign objects with high accuracy (>95%), label the contour of foreign objects, and acquire the centroid location, thus reducing the difficulty of inspection. It is a new, concise, and effective means for identification of foreign objects on power equipment.
研究不足
The accuracy of detection may be influenced by the background and lighting conditions. The method requires adjustment of parameters for different sizes of foreign objects.
1：Experimental Design and Method Selection:
The study employs morphological closing operation and morphological features restriction for foreign object detection.
2：Sample Selection and Data Sources:
Images of overhead line towers with foreign objects such as nests, kites, and balloons captured by UAV.
3：List of Experimental Equipment and Materials:
Software environment: Windows 7; Hardware environment: i5-6200U+4G RAM+500G Hard disk.
4：Experimental Procedures and Operational Workflow:
Pre-treatment of images including gradation processing and denoising, OTSU threshold segmentation, morphological closing operation, canny edge detection, morphological fill, and cycle screening in area and circularity.
5：Data Analysis Methods:
Accuracy of foreign object detection is measured in object level.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
