Unmanned aerial vehicle visual inspection and recognition method for crane complex steel structure surface defects

The invention discloses an unmanned aerial vehicle visual inspection and recognition method for crane complex steel structure surface defects, which comprises the following steps: carrying a high-resolution visible light camera by using an inverted unmanned aerial vehicle platform, and collecting images through a cattle-tilling type full-coverage inspection path; constructing a classification algorithm fusing a support vector machine, a deep convolutional network and a generative adversarial network, subjecting multi-scale and multi-type defects on the surface of the crane structure under the complex background to classification detection, and marking the positions of the defects with a minimum bounding rectangular frame of the defects; segmenting the detected defect target frame area to extract a connected domain, establishing a pixel equivalent accurate calibration model based on multi-point laser ranging, and counting and collecting parameters such as length, width and area of the defect through a pixel po