| 引用本文: | 袁朝晖,付文龙,李佰霖,等.基于多策略分割融合与形态特征辨识的变电站
保护压板状态识别[J].电力系统保护与控制,2022,50(1):98-106.[点击复制] |
| YUAN Zhaohui,FU Wenlong,LI Bailin,et al.Protection platen status recognition for a smart substation based on multi-strategy segmentation and fusion and morphological feature identification[J].Power System Protection and Control,2022,50(1):98-106[点击复制] |
|
| 摘要: |
| 目前变电站保护压板巡检仍主要采用人工进行，耗时费力且容易出错，制约着变电站二次设备智能化的发展。为此，提出一种基于多策略分割融合与形态特征辨识的变电站保护压板状态识别方法。通过移动端设备采集屏柜压板图像后，首先对压板区域进行透视变换，消除拍摄角度产生的畸变影响。然后采用多策略分割融合方法获取有效压板区域，即在HSV空间利用多阈值分割。同时在Lab空间采用K均值聚类分割，再对两种分割结果进行融合，获得有效压板区域。最后计算有效压板方向角和宽长比形态特征，并分别判别两种特征对应的状态，进一步融合两种状态结果，辨识压板最终运行状态。通过对不同场景下的复杂背景压板图像进行实例研究，结果表明该方法具有良好的准确率和适用性。 |
| 关键词: 保护压板 多阈值 K均值聚类 形态特征 运行状态 |
| DOI：DOI: 10.19783/j.cnki.pspc.210355 |
| 投稿时间：2021-04-01修订日期：2021-06-25 |
| 基金项目:国家自然科学基金项目资助(51741907) |
|
| Protection platen status recognition for a smart substation based on multi-strategy segmentation and fusion and morphological feature identification |
| YUAN Zhaohui,FU Wenlong,LI Bailin,WEN Bin |
| (1. College of Electrical Engineering & New Energy, China Three Gorges University, Yichang 443002, China; 2. Hubei Key
Laboratory of Cascaded Hydropower Stations Operation & Control, China Three Gorges University, Yichang 443002, China) |
| Abstract: |
| The inspection of protection platens in substations at present still relies on manual operation, which is time-consuming and error-prone. It also restricts the development of more intelligent secondary equipment in substations. Given this, a novel status recognition method for a substation protection platen based on multi-strategy segmentation and fusion and morphological feature identification is proposed. With a cabinet image collected by the mobile terminal, the platen region is first dealt with perspective transformation to eliminate the distortion caused by the image acquisition angle. Then, a multi-strategy segmentation and fusion method is proposed to extract the valid platen region, within which multi-threshold segmentation is applied in the HSV space while K-means clustering is adopted for segmentation in Lab space. Thus, the valid platen region is obtained by fusing the two kinds of segmentation results. Subsequently, morphological features of the valid platens including direction angle and aspect ratio are calculated, based on which the state corresponding to different features is distinguished. Then the two state results are merged into the final operating state of platens. Results in application for different cases for cabinet images with complex background show that the proposed method achieves good accuracy and applicability.
This work is supported by the National Natural Science Foundation of China (No. 51741907). |
| Key words: protection platen multi-threshold K-means clustering morphological feature operating state |
