To solve the problems of poor sawing surface quality, severe blade wear and high power consumption caused by unreasonable working parameters in the process of Caragana korshinskii (C.K.) stumping, this study explored the effects of branch diameter (D), sawing speed $\left(v_c\right)$, feeding speed $\left(v_f\right)$, cutting inclination (α), number of circular saw teeth (T) and moisture content (M) on sawing power consumption (P) and sawing surface quality (A) through a single-factor test using a homemade branch sawing bench. Based on the Box–Behnken design principle, a multi-factor test was carried out based on a single-factor test with $v_c$,$v_f$, α and T as influencing factors and with P and A as targets, establishing a regression model. The test results show that the sawing power consumption (P) increases with increasing D, decreases with increasing M, and decreases first and then increases with increasing $v_c$, $v_f$, α and T; the sawing surface quality (A) increases first and then decreases with increasing D, increases with increasing M, and first increases and then decreases with increasing $v_c$, $v_f$, α and T. The optimum combination of parameters for the regression model was obtained with $v_c$ of 45.24 m/s, $v_f$ of 0.34 m/s, α of 10° and T of 100, which resulted in the P of 177.46 J and A of 85.87%. The errors between the predicted and actual values of P and A are 3.1% and 6%, respectively. The study can provide information to support the development of subsequent C.K. stubble equipment.

In order to reduce the impact of pests and diseases on the yield and quality of Ningxia wine grapes and to improve the efficiency and intelligence of detection, this paper designs an intelligent detection platform for pests and diseases. The optimal underlying network is selected by comparing the recognition accuracy of both MobileNet V2 and YOLOX_s networks trained on the Public Dataset. Based on this network, the effect of adding attention mechanism and replacing loss function on recognition effect is investigated by permutation in the Custom Dataset, resulting in the improved network YOLOX_s + CBAM. The improved network was trained on the Overall Dataset, and finally a recognition model capable of identifying nine types of pests was obtained, with a recognition accuracy of 93.35% in the validation set, an improvement of 1.35% over the original network. The recognition model is deployed on the Web side and Raspberry Pi to achieve independent detection functions; the channel between the two platforms is built through Ngrok, and remote interconnection is achieved through VNC desktop. Users can choose to upload local images on the Web side for detection, handheld Raspberry Pi for field detection, or Raspberry Pi and Web interconnection for remote detection.