Abstract
As a typical application of indirect-time-of-flight (ToF) technology, photonic mixer device (PMD) solid-state array Lidar has gained rapid development in recent years. With the advantages of high resolution, frame rate and accuracy, the equipment is widely used in target recognition, simultaneous localization and mapping (SLAM), industrial inspection, etc. The PMD Lidar is vulnerable to several factors such as ambient light, temperature and the target feature. To eliminate the impact of such factors, a proper calibration is needed. However, the conventional calibration methods need to change several distances in large areas, which result in low efficiency and low accuracy. To address the problems, this paper presents an improved calibration method based on electrical analog delay. The method firstly eliminates the lens distortion using a self-adaptive interpolation algorithm, meanwhile it calibrates the grayscale image using an integral time simulating based method. Then, the grayscale image is used to estimate the parameters of ambient light compensation in depth calibration. Finally, by combining four types of compensation, the method effectively improves the performance of depth calibration. Through several experiments, the proposed method is more adaptive to multiscenes with targets of different reflectivities, which significantly improves the ranging accuracy and adaptability of PMD Lidar.
Highlights
Three-dimensional information acquisition has gained extensive attention in the field of computer vision, robot navigation, human–computer interaction, automatic driving, etc. [1]
The grayscale image was used to estimate the parameters of ambient light compensation in depth calibration in this method
The grayscale image was used to estimate the parameters of ambient light compensation in depth calibration, which is the self-adaptive grayscale correlation based depth calibration method (SA-GCDCM)
Summary
Three-dimensional information acquisition has gained extensive attention in the field of computer vision, robot navigation, human–computer interaction, automatic driving, etc. [1]. The methods to obtain three-dimensional information mainly include stereo vision [2,3], structural light [4], single-pixel 3D imaging [5] and time-of-flight (ToF) [6]. Stereo vision needs advanced matching algorithm to obtain accurate depth information, which is vulnerable to ambient light. Compared with the two methods above, the ToF sensor system utilizes active infrared laser to achieve depth information acquisition, which has the advantages of low cost, high frame frequency and high reliability [7,8]. There inevitably exist several errors sources (such as ambient light, integration time, temperature drift and reflectivity), which reduce the performance of the ToF sensor significantly. The equipment needs to be properly calibrated to achieve reliable depth information acquisition [11]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
