| Harmonic gears are widely used in military and civilian fields due to their large reduction ratio,small size,and high accuracy.Harmonic gears are a key component especially for robots.However,the performance of the harmonic reducer(such as transmission smoothness,transmission accuracy,transmission torque,etc.)are directly related to the actual tooth thickness of the flexible wheel and rigid wheel.Therefore,if there is a measurement method that can quickly and accurately detect the tooth thickness of the flexible and rigid wheels,and effectively control the clearance between the flexible and rigid wheels,then it will have a great impact on improving the performance of the harmonic reducer.Based on these problems,this thesis measures the tooth thickness and backlash of the flexible and rigid wheels in the harmonic drive system,and conducts flaw detection on the outer surface of the rigid wheels to improve the quality of the products.During these years,the following works have been completed:(1)In this thesis,a hardware experimental platform based on visual detection technology is built.Firstly,the selection of equipment required for visual measurement is carried out.Secondly,the positioning calibration experiment of industrial cameras.Then,internal and external parameters of the camera are obtained.Then,the re-projection error rate of the calibration experiment is obtained.At last,the hardware platform design and software flow design are completed.(2)In this thesis,the adaptive tooth profile detection algorithm based on the low rank matrix in the pixel-level category is proposed for the characteristics of harmonic gears with small modulus,large number of teeth and the defects of existing tooth profile detection algorithms.Firstly,the algorithm uses low rank recovery theory to remove tooth burr interference.Then,an adaptive threshold segmentation method to solve the problem of edge breakpoints present in tooth profile detection is adopted.The experimental results show that the algorithm achieves the effect of reducing dentate burr interference with enhanced dentate extraction integrity.(3)In this thesis,the influence of harmonic flex wheel and rigid wheel tooth structure on the performance of harmonic reducer is investigated.Then,the proposed pixel-level detection method combined with the interpolated sub-pixel positioning method is used.At last,the chord tooth thickness measurement strategy is adopted to complete the tooth thickness measurement experiment.Firstly,the experiment extracts the tooth profile edge information.Secondly,the gear center point is determined.Then,the tooth groove line and the tooth thickness line are located.Finally,the experiment completes the visual measurement of the tooth thickness based on the intersection of the tooth thickness line and each tooth to determine the measurement point of the string tooth thickness.(4)The results are compared with the actual tooth thicknesses obtained from the gear center meter using multiple sets of measurements obtained from the experiment in this thesis.The absolute error of tooth thickness is around 0.02 mm,which proves that the measurement method in this thesis is effective and the detection accuracy is high.The above works show that the adaptive tooth profile detection algorithm based on the low rank matrix theory is well integrated with the chord tooth thickness visual measurement strategy on the designed visual measurement platform.Finally,the tooth thickness of harmonic gear measurement is completed.The experimental results show that the absolute error between the measurement results in this thesis and the actual measurement results is in accordance with the processing error range and meets the requirements of the topic. |
