Research on the Error Detection Technology of Complex Parts based on Improved Bee Colony Algorithm

For Parts of complex structures with multiple characteristics, traditional methods are often difficult and inefficient, so the three-dimensional detection techniques have been developed. But the detection used for the complex structural components still get some problems, such as : it's difficult to collect data of complex structural components in the data collection process, because of size is always limited; features of parts retained difficultly in data processing; there are many iterations and calculations in the optimization algorithm of geometric error evaluation, and the search direction is blind. In this paper, the improved colony algorithm is used to get the cylindrical error by establishing the mathematical model with the point cloud data that has been processed. First, collect data acquisition of more structural parts with hand-held laser scanners; Secondly, denoise and streamline the raw point cloud collected before, and then make registration and alignment with the actual CAD model of part; Finally, according to the resultant point cloud data above, in accordance with the principle of minimum condition assessment, establish evaluation model of cylinder error based on improved bee colony algorithm. Experimental results show that for various structural parts, the precision of this detection method is high enough, and detecting speed is also fast.