In recent years, data obtained from sensors attached to excavators are being widely used in assisting the operators. The processing unit logic of such a sensor system has been developed to ensure flexibility in the design of the sensor system. As a part of this development, an effort has been made to solve the problem of detecting disturbances, such as rain, snow, and insects, when the depth sensor is used for outdoor measurements. An object placed in front of a depth sensor in a raining environment was moved back and forth while continuously measuring the appearance of the object. In the measured data, points with possible disturbances were chosen. The speed, direction, and acceleration between frames were calculated as feature quantities, which led to the development of a disturbance rejection filter that eliminates the points above set thresholds. This filter has been confirmed to detect the position of objects while ignoring the influence of rain in real applications.
展开▼