欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
相对传统的火焰识别方法，为提高火焰识别精度，提出了一种基于红外热像仪和彩色摄像机的火焰识别算法。对采集得到的红外图像进行温度阂值分割，得到高温疑似火焰区域，通过三维重建技术匹配彩色摄像机中对应的高温疑似火焰区域，对其RGB空间模型进行色彩分析，排除干扰得到疑似火焰区域。在此基础上分析疑似火焰区域的区域增长特性及闪动特性，提取疑似火焰区域的面积及其轮廓，对其轮廓进行离散余弦变换，计算前后两帧图像轮廓的变化，从而进一步确定真实火灾火焰的发生，及时报警，减少误判率和危害性。实验证明了该算法的准确性和有效性。
英文摘要：
Compared with the traditional flame recognition method, a flame recognition algorithm based on thermal infra- red imager and color camera is introduced to improve the flame recognition accuracy. At first the collected infrared image was segmented by temperature threshold segmentation, and the suspected high-temperature flame area was found. Secondly, the corresponding suspected high-temperature flame area in the color camera was matched through the three-dimensional recon- struction technology, and its color component in RGB space model was analyzed to exclude interference and gain suspected fire area. Thirdly, the regional growth characteristic and flashing characteristic of the suspected flame areas were discussed, the contour of a suspected flame area was extracted and its area was computed, and then the discrete cosine transform of the con- tour line was performed, the contour changes of two consecutive images were calculated, which further determined the true in- cidence of the fire and gave an alarm in a timely manner, reduced false positives and dangerous. The accuracy and validity of the algorithm are proved.
同期刊论文项目
同项目期刊论文
期刊信息
位置：