Research On Vision Detection Of The Aircraft Skin Normal Vector

The application of automatic drilling technology in aircraft assembling haveimproved the assembly effect, reduced costs and solved many problems of manualdrilling such as ineffective, non-uniformity of quality; The automatic drillingequipment needs to detect the normal vector of aircraft skin to make holeperpendicular to the skin surface. The detection of aircraft skin normal vector is animportant part of automatic drilling. On the basis of research at home and abroad, thepaper proposed a method of using machine vision to detect the skin normal vector.At first, the paper established a vision detecting system, which using cross-lightlaser to remark two curves in the skin, and take photos for the curves by monocularshooting ways, then using FPGA as processor to calculate the normal vector form theimages of the light.Secondly, designed the image acquisition hardware circuit and the control logicof FPGA. And remove the noise in the image by spatial filtering; extract samplepoints of the marked lights curve by sample widow method. According to cameraimaging geometry theory, the paper established monocular imaging model whichbased on the positional relationship of laser and camera. Using this model,3Dinformation of sample points can be obtained form it’s2D information. Then optimizethe cubic B-spline fitting algorithm so that it can be applied on FPGA. With thealgorithm, the equation of marked lights curve can be calculated. Then get twotangents of drill point form curve equation. The normal vector could calculate bycross product calculation on these two tangents.At last, designed a detecting platform according to the monocular imaging model.And built this platform with modular fixture. Then tested with flat and curved surfaceon this platform. The test shows that detecting skin normal vector with machine visionwould achieve better results.