无人机巡检
一、无人机基础 | Drone (UAV) Fundamentals
二、无人机巡检应用 | Drone Inspection Applications
三、巡检数据处理 | Inspection Data Processing
四、行业专用无人机 | Industry-Specific Drones
五、合规与安全 | Compliance & Safety
Characters: O&M Engineer (运维工程师 OE) / Drone Pilot (无人机飞手 DP) · Characters: Project Controls Manager (控制经理 PC) / Survey Manager (测量经理 SM)
We have a scheduled thermal inspection of the 50 MW solar array today. The weather is clear with light wind — perfect conditions. What's the flight plan?
今天计划对50MW光伏阵列做热成像巡检。天气晴微风——完美条件。飞行计划?
I've programmed an autonomous grid flight at 30 meters altitude, covering all 156,000 panels in 12 battery swaps. The thermal camera will capture at 640×512 resolution with a 1.5 cm/pixel ground sampling distance. We're looking for three things: hot spots indicating cell-level defects, bypass diode failures, and string-level disconnections. The total mission time is about 4 hours.
已编好自动网格飞行:高度30米,覆盖15.6万块光伏板,换12次电池。热成像相机 640×512分辨率,地面采样距离1.5厘米/像素。找三样东西:电池片级热斑缺陷、 旁路二极管故障、组串级断连。总任务约4小时。
Good. After the flight, how long until we have actionable data? Last time the processing took two days.
好。飞行后多久能有可操作数据?上次处理花了两天。
We've upgraded our processing pipeline. The onboard edge processor does preliminary hotspot detection during the flight — by the time I land after the last swap, 90% of anomalies are already flagged. The full orthomosaic with AI defect classification will be ready within 2 hours after landing. You'll get a report showing a heat map of the entire array with GPS-tagged defect locations ranked by severity. I expect we'll find about 15-20 actionable defects. We've already reduced the inspection cycle from 5 days with manual methods to 4 hours with the drone.
处理管线升级了。机载边缘处理器在飞行中就做初步热斑检测——最后一次换电降落时, 90%异常已标出。完整正射影像加AI缺陷分类落地后2小时内就绪。你会收到一份报告: 全场热力图加GPS定位的缺陷点按严重度排序。预计找出15-20个可操作缺陷。巡检 周期已从人工法的5天缩短到无人机的4小时。
We're running the weekly drone survey of the site today. I need the orthomosaic overlaid with the latest design to verify earthworks progress on the access road and the substation platform. The client is questioning our progress claim of 78% on bulk earthworks.
今天做每周无人机现场测绘。需要正射影像叠加最新设计图,核实进场道路和变电 站平台的土方进度。业主在质疑我们78%路基土方的进度申报。
The drone survey was completed this morning. I've processed the data into a digital surface model and compared it against the design surface using cut-and-fill analysis. The numbers: access road earthworks are at 82%, not 78% — we're actually ahead. The substation platform is at 71%, a bit behind due to the rock excavation we encountered. Total bulk earthworks volume: 94,000 m³ completed out of 127,000 m³ planned — that's 74%, close to our 78% claim.
今早完成了无人机测绘。数据处理成数字表面模型,与设计面对比做了挖填分析。 数据:进场道路土方82%,不是78%——实际超前。变电站平台71%,因遇到石方略滞后。 总土方量:已完9.4万方/计划12.7万方——74%,接近我们申报的78%。
So our 78% claim is slightly optimistic on the overall, but defensible with the road progress compensating for the platform. Let's present the drone data at the progress meeting — the orthomosaic with the colored cut-and-fill map is powerful evidence. Much better than arguing over manual survey figures. How often can we run these drone surveys?
所以总体78%略乐观,但道路超前进度弥补了平台滞后,数据站得住。在进度会上用 无人机数据——正射影像加彩色挖填图是有力证据。远比争论手工测量数据强。这种 无人机测绘多久能做一次?
We're doing weekly now, but we can go twice a week during critical earthworks phases. The cost per survey is about USD 800 for a site this size — that's minimal compared to the value of accurate progress data for claims and client reporting.
现在每周一次,关键土方阶段可以每周两次。这个规模的现场每次测绘约800美元—— 跟准确的进度数据对索赔和业主报告的价值比微不足道。