Search

The cleaning vehicle is a special equipment for real-time detection and cleaning of the track. To ensure normal operation of the cleaning vehicle, we need to establish a remote monitoring system to collect all the state parameters of the entire cleaning vehicle through sensors to monitor them in real time. According to the parameters collected by sensors, we can evaluate the state of the vehicle's working condition and give corresponding treatment suggestions. The system collects various status parameters of the cleaning vehicle through STM32. After storing the data into the database, it uses VC++ to establish man-machine interface and give corresponding processing recommendations, making the vehicle more reliable and more intelligent.

Vision-based dynamic gesture recognition is an important means of human-computer interaction. This paper designs a dynamic gesture recognition control system based on USB single camera. After obtaining the image through the camera, the original image is binarized, eroded, smoothed and filled with holes, then the circumscribed rectangles and the centroid points of the gesture image are tracked, then the movement direction of the gesture is determined according to the motion track. Finally, through the gesture control movement of the mouse pointer to move and simulate the left-click operation. The test results show that the system recognition rate is high, therefore the research has some practical values.

In the environment of large forest, the factors causing fire are nonlinear and uncertain. If the data collected by the sensor is simply analyzed and compared, the false alarm rate will be higher. How to combine the data of several sensors for effective fire warning is a difficult point. In order to improve the accuracy of prediction, aiming at the shortcomings of traditional forest fire prevention early warning system, we propose a forest fire prevention early warning method based on fuzzy Bayesian network. Firstly, we combine the fuzzy control system and the Bayesian network in series, and pre-process the collected sensor data. The pre-processed data is sent to the previously trained Bayesian network for processing. Then the calculated open fire probability, smoldering fire probability, and no fire probability are used as input data of fuzzy control system, and fuzzy inference is performed. Finally, we de-fuzzify the results of fuzzy reasoning and get the probability of fire. Simulation results show that our method can effectively combine the data collected by multiple sensors, quickly and accurately determine fire occurrence probability, improve the accuracy of forest fire prevention warning, and reduce the false positive rate.