| 基于天空分割的去雾算法在目标检测中的应用 |
| |
| 引用本文： | 朱晓霞,施智平,邵振洲,关永,任长娥. 基于天空分割的去雾算法在目标检测中的应用[J]. 计算机应用与软件, 2020, 37(2): 152-160,225 |
| |
| 作者姓名： | 朱晓霞 施智平 邵振洲 关永 任长娥 |
| |
| 作者单位： | 首都师范大学信息工程学院电子系统可靠性技术北京市重点实验室 北京 100048;首都师范大学成像技术北京市高精尖创新中心 北京 100048;首都师范大学信息工程学院轻型工业机器人与安全验证北京市重点实验室 北京 100048;首都师范大学成像技术北京市高精尖创新中心 北京 100048;首都师范大学电子系统可靠性与数理交叉国际科技合作基地 北京 100048;首都师范大学电子系统可靠性与数理交叉国际科技合作基地 北京 100048 |
| |
| 基金项目： | 基本科研业务费项目;国家自然科学基金;国家重点研发计划;北京市科委项目 |
| |
| 摘 要： | 针对道路目标检测的准确率受雾霾环境影响的问题,提出一种基于雾检测与天空分割的自适应去雾算法。提取道路环境中的雾检测区域并对其进行通道相关性以及亮度和饱和度比值的计算,以判别场景类别;对图像进行天空非天空区域的分割并对不同区域建立去雾模型,用于自适应去雾;将恢复后的图像应用到道路目标检测算法中,进行去雾效果验证和道路目标检测。实验结果表明,该算法提高了雾霾环境下目标检测的准确率。
|
| 关 键 词： | 消失点检测 天空分割 暗通道 自适应去雾 道路目标检测 Yolo |
APPLICATION OF DEFOGGING ALGORITHM BASED ON SKY SEGMENTATION IN OBJECT DETECTION |
| |
| Affiliation: | (Beijing Key Laboratory of Electronic System Reliability and Prognostics,College of Information Engineering,Capital Normal University,Beijing 100048,China;Beijing Advanced Innovation Center for Imaging Technology,Capital Normal University,Beijing 100048,China;Beijing Key Laboratory of Light Industrial Robot and Safety Verification,College of Information Engineering,Capital Normal University,Beijing 100048,China;International Science and Technology Cooperation Demonstration Base of Interdisciplinary of Electronic System Reliability and Mathematics,Capital Normal University,Beijing 100048,China) |
| |
| Abstract: | Aiming at the problem that the accuracy of road object detection is affected by fog environment,we propose an adaptive defogging algorithm based on fog detection and sky segmentation.It extracted the fog detection area in road environment and calculated the channel correlation and the ratio of brightness and saturation so as to identify the scene category.The image was segmented into sky and non-sky regions,and corresponding models were separately established for adaptive defogging.We applied the restored image to the object detection task to verify the defogging effect and detect the road object.The experimental results show that our algorithm improves the accuracy of object detection in fog envi-ronments. |
| |
| Keywords: | Vanishing point detection Sky segmentation Dark channel Adaptive defogging Road object detection Yolo |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
