Real-time obstacle detection in a darkroom using a monocular camera and a line laser

Abstract

When a disaster strikes, one of the highest priorities is to save the lives of the victims. The proportion of victims that rescuers can save is strongly related to how quickly rescue efforts can begin. Therefore, early detection of disaster victims is very important. However, significant risks are involved in rescue operations immediately after a disaster. In fact, in the Great East Japan Earthquake, approximately 250 firefighters died while rescuing victims. Under such circumstances, rapid and safe rescue operations are needed at disaster sites. For this purpose, it is important to improve the technology of disaster relief robots. In this paper, we propose an algorithm to measure the linear distance to an obstacle in real time using only a line laser and a monocular camera. This approach allows the use of a camera to obtain more information than the one-dimensional information such as that obtained by ultrasonic sensors. Moreover, this method of obstacle detection for disaster rescue robots is smaller and more durable than large measurement systems such as LiDAR that have been used in the past. In addition, since only one camera is used, the processing cost is low and the processing equipment is expected to be small. The proposed method’s effectiveness is indicated by comparing the distance measured from the image processing results in a dynamic environment with the actual distance between the obstacle and a crawler robot by having the robot move straight toward to an obstacle.

A task level fusion autonomous switching mechanism

“…In indoor scenarios where firefighting robots perform tasks, GPS signals are often unavailable [5]. The current indoor positioning methods have many methods, including ultra wide band (UWB), inertial measurement unit (IMU), infrared depth sensor (IDS), camera [4,[6][7][8][9][10][11]. UWB has high temporal resolution but is susceptible to indoor environments [4,6].…”

“…IDS is easily detected at close range but has limited detection distance, high noise, and low accuracy [8,9]. Camera accuracy is great, but the real-time performance is poor and vulnerable to light conditions [10,11]. These sensors can be used to their advantage in different usage scenarios and are difficult to replace.…”