西部矿区煤层赋存条件好，开采强度高，上覆岩层破坏严重，易诱发采空塌陷和地裂缝等灾害，造成地表生态损伤，甚至诱发遗煤自燃，威胁煤矿安全生产。为快速、及时、准确识别地裂缝，提出了基于无人机红外遥感及图像边缘检测技术的地裂缝识别方法。以神东矿区上湾煤矿12401工作面为工程背景，对工作面上方地裂缝发育观测区域进行全天候监测，获取了不同时刻的红外图像，并对不同时刻红外图像中裂缝、沙子、植被的温度信息以及裂缝长度进行了统计和分析。运用多种边缘检测方法及提出的改进边缘检测算法，对采集的典型红外图像进行地裂缝检测，评价了不同边缘检测方法的裂缝检测效果。对比分析了不同时刻地裂缝检测结果，给出了针对本文研究条件下无人机红外遥感技术识别地裂缝的最佳时间窗口。研究结果表明：基于无人机搭载红外相机及边缘检测技术可有效识别采矿导致的地裂缝；夜间相比于白天，地裂缝更易被识别，特别是3:00 am~5:00 am期间，地裂缝的识别效果最佳。提出的改进边缘检测算法的Pratt品质因数（PFoM）值为0.571，检测性能优于文中其他边缘检测方法；1:00 am至5:00 am和7:00 pm至11:00 pm，地裂缝边缘检测结果优于其他时间，检测结果中噪声较少，裂缝边缘清晰突出；其中尤以3:00 am、5:00 am时的地裂缝检测效果最好。

The western mining area in China has good coal deposit conditions,high mining intensity and serious damage to overlying strata,which can induce goaf collapse,ground fissures and ecological destruction on the surface.Moreover,the spontaneous combustion of residual coal also occurs,which can threat the safety of coal mine production.To collect the information of mining induced ground fissures quickly,timely and accurately,a method to identify ground fissures was proposed based on an unmanned aerial vehicle (UAV) equipped with an infrared camera and image edge detection technology.Taking the working face No.12401 of the Shangwan coal mine in Shendong mining area as the engineering background,a round the clock monitoring of ground fissure development area above the working face was conducted,and some infrared images at different times were obtained.The temperature information of fissure,sand,vegetation and the length of fissure in infrared images at different times were statistically analyzed.A variety of edge detection methods and the proposed improved edge detection method were used to detect the fissures in the typical infrared image,and the fissure detection results were evaluated.By comparing and analyzing the results of fissure detection at different times,the optimal time for identifying the ground fissures by UAV infrared remote sensing technology under the conditions studied in this paper was given.The results show that an UAV equipped with an infrared camera and edge detection technology can be used to identify mining induced ground fissures effectively.Compared with the daytime,the ground fissures are easier to be identified at night,especially during 3:00 am and 5:00 am.The Pratt’s figure of merit (PFoM) value of the improved edge detection method is 0.571.Compared with the other edge detection methods,this method has a good detection effect for ground fissures.From 1:00 am to 5:00 am and 7:00 pm to 11:00 pm,the results of fissure detection are better than those of other times.The noise in the detection results is less,and the edges of the fissures are clearly prominent.Especially the fissures detection effect at 3:00 am and 5:00 am are the best.

1 地裂缝监测试验 

　　1.1 工程背景 

　　1.2 监测设备

　　1.3 监测方案

2 地裂缝监测结果及分析

　　2.1 不同时间点地裂缝红外图像特征 

　　2.2 地裂缝辐射温度特征及规律

　　2.3 不同时间点地裂缝尺寸特征

3 地裂缝检测方法改进

　　3.1 不同边缘检测方法下地裂缝检测

　　3.2不同时间点的地裂缝检测

4 结论