Industrial production of edible fungi is the future development direction of China's edible fungi industry. At present, the industrial production of edible fungi in China is dominated by wood-rotting fungi, and its production process includes bottling, sterilization and a series of links until the end of bottle digging. However, after bottle digging, the bottle cleaning still adopts manual cleaning, which leads to low washing efficiency and high labor costs. Based on analyzing the requirements of bottle washing in the industrial production of edible fungi, this paper introduces the working principle of bottle washing machines in the industrial production of edible fungi and designs the control system of bottle washing machines on this basis. Around the hardware and software system, this paper analyzes the functions and main model selection of the drive module, actuator module, sensor module, and designs the pneumatic drive system. According to the process of the bottle washing machine, using Guangyang SH-32R as a programmer, the electrical schematic diagram is drawn, and the control program of the bottle washing machine is written in level language. The trial production and test show that the bottle washing machine runs smoothly, the average pressure of the nozzle is 0.33 MPa, the water consumption of each bottle is 0.1 kg, and the single cleaning time is 10 s, which meets the requirements of industrial bottle washing.

Edible mushrooms are rich in nutrients; however, harvesting mainly relies on manual labor. Coarse localization of each mushroom is necessary to enable a robotic arm to accurately pick edible mushrooms. Previous studies used detection algorithms that did not consider mushroom pixel-level information. When these algorithms are combined with a depth map, the information is lost. Moreover, in instance segmentation algorithms, convolutional neural network (CNN)-based methods are lightweight, and the extracted features are not correlated. To guarantee real-time location detection and improve the accuracy of mushroom segmentation, this study proposed a new spatial-channel transformer network model based on Mask-CNN (SCT-Mask-RCNN). The fusion of Mask-RCNN with the self-attention mechanism extracts the global correlation outcomes of image features from the channel and spatial dimensions. Subsequently, Mask-RCNN was used to maintain a lightweight structure and extract local features using a spatial pooling pyramidal structure to achieve multiscale local feature fusion and improve detection accuracy. The results showed that the SCT-Mask-RCNN method achieved a segmentation accuracy of 0.750 on segm_Precision_mAP and detection accuracy of 0.638 on Bbox_Precision_mAP. Compared to existing methods, the proposed method improved the accuracy of the evaluation metrics Bbox_Precision_mAP and segm_Precision_mAP by over 2% and 5%, respectively.