微型电脑应用

供热管道往往铺设在地下或建筑物内部，且覆盖较大的地理区域。若巡检覆盖面小，易出现漏检问题。为此，提出基于红外热成像机器人的供热管道安全状态巡检方法。采用红外热成像机器人采集供热管道现场图像，利用简单线性迭代聚类(SLIC)算法，通过快速聚类的方式将图像分割成多个块状区域；灰度化处理分割后的图像，确定不同块状区域的热点温度和不同缺陷类型下的发热面积，全面提取供热管道状态特征，通过扩大巡检覆盖面的方式避免漏检；将提取到的特征输入宽度卷积神经网络，通过多层卷积操作学习特征；通过多层神经元间的复杂连接，更好地适应不同类型的供热管道状态特征，减少误检的概率，实现安全状态巡检。实验表明，所提方法的热管道安全状态巡检效果更优，可以快速、精准检测供热管道安全状态，漏检率和误检率均较低。

Heating pipelines are often laid underground or inside buildings, and cover a large geographical area. If the inspection coverage is small, it is easy to have missed inspections. To this end, a safety status inspection method for heating pipelines based on infrared thermal imaging robot is proposed. An infrared thermal imaging robot is used to capture on-site images of heating pipelines, the simple linear iterative clustering(SLIC) algorithm is used to quickly cluster the images into multi-block regions. Gray-scale processed segmented images are used to determine hot spot temperature in different block areas and the heating area under different defect types. The status features of heating pipelines are comprehensively extracted, and missed detections are avoided by expanding the inspection coverage. The extracted features is input into a width convolutional neural network, and the features are learned through multi-layer convolutional operations. The system can better adapt to different types of heating pipeline status features through complex connections between multi-layer neurons, reduce the probability of false inspection and achieve safety status inspection. The experiment shows that the proposed method has a better inspection effect on the safety status of heating pipelines, and can quickly and accurately detect the safety status of heating pipelines. The missed detection rate and false detection rate are both low.