Powder metallurgy gears are often accompanied by broken teeth, abrasion, scratches and crack defects. In order to eliminate the defective gears in gear production and improve the yield of gears, this paper presents an improved GA–PSO algorithm, called the SHGA–PSO algorithm. Firstly, the gear images were preprocessed by bilateral filtering, and the images were segmented by the Sobel operator. Then, the geometrical shape, texture feature and color features of the sample were extracted. Next, the BP neural network was reconstructed and SHGA–PSO algorithm was used optimize its structure and weights. Finally, four different gear defect samples were brought into the neural network for calculation, and the performance of the SHGA–PSO algorithm was compared with the GA, PSO and GA–PSO algorithms. Compared with GA–BP algorithm, PSO–BP algorithm, and GA–PSO–BP algorithm, the defect diagnosis of SHGA–PSO–BP algorithm not only enhanced generalization ability, but also improved recognition accuracy.

粉末冶金齿轮常伴有断齿，磨损，擦伤和裂纹缺陷。为了消除齿轮生产中的不良齿轮并提高齿轮的产量，本文提出了一种改进的GA-PSO算法，称为SHGA-PSO算法。首先，通过双边滤波对齿轮图像进行预处理，然后通过Sobel算子对图像进行分割。然后，提取样品的几何形状，纹理特征和颜色特征。接下来，重建BP神经网络，并使用SHGA-PSO算法优化其结构和权重。最后，将四个不同的齿轮缺陷样本引入神经网络进行计算，并将SHGA-PSO算法的性能与GA，PSO和GA-PSO算法进行了比较。与GA–BP算法，PSO–BP算法相比，