Turbine blades are mainly distinguished by the positive and negative of the fan blades. A large number of repetitive turbine detection can easily interfere with the traditional manual detection and confuse it, so it can not accurately classify the turbines. Through the most direct judgment of the appearance of the turbine, it can be judged that we can get the software algorithm to distinguish the positive and negative sides of the turbine through simple lighting. In this analysis, we do not need to use 3D detection, only need to use 2D detection under good lighting conditions to solve the problem.
1. Detect turbine blade blades.
2. The detected object is placed in a rectangular container.