A Comparison of Feature Extractors for Panorama Stitching in an Autonomous Car Architecture

Abstract

Panorama stitching consists on frames being merged to create a 360° view. This technique is proposed for its implementation in autonomous vehicles instead of the use of an external 360-degree camera, mostly due to its reduced cost and improved aerodynamics. This strategy requires a fast and robust set of features to be extracted from the images obtained by the cameras located around the inside of the car, in order to effectively compute the panoramic view in real time and avoid hazards on the road. This paper compares and creates discussion of three feature extraction methods (i.e. SIFT, BRISK and SURF) for image feature extraction, in order to decide which one is more suitable for a panorama stitching application in an autonomous car architecture. Experimental validation shows that SURF exhibits an improved performance under a variety of image transformations, and thus appears to be the most suitable of these three methods, given its accuracy when comparing features between both images, while maintaining a low time consumption. Furthermore, a comparison of the results obtained with respect to similar work allows us to increase the reliability of our methodology and the reach of our conclusions.

Panoramic image stitching technology and its application in the field of autonomous driving

“…Hence, the circulation in the market is relatively small. A small number of existing full range of panoramic camera equipment that can be bought in the market is also relatively high price(table 3) [10]. In addition, because one of the characteristics of the use of panoramic camera equipment is to install this equipment must be installed in the outer body of the vehicle, generally placed in the roof position, so there is a great risk of theft, coupled with the high cost of the equipment, if stolen will lead to higher cost losses [10].…”

“…On the other hand, the vehicle in the process of driving will produce huge airflow hedge, and the airflow in the vehicle to cause obstruction will often become a major factor affecting the development of the speed of the vehicle [10]. The feature that the panoramic camera needs to be installed outside the body will lead to the inefficiency of these cameras in terms of aerodynamics, which greatly limits the development of vehicle performance in terms of speed [10].…”

“…On the other hand, the vehicle in the process of driving will produce huge airflow hedge, and the airflow in the vehicle to cause obstruction will often become a major factor affecting the development of the speed of the vehicle [10]. The feature that the panoramic camera needs to be installed outside the body will lead to the inefficiency of these cameras in terms of aerodynamics, which greatly limits the development of vehicle performance in terms of speed [10]. Jingxiong Zhang et al has also proposed that ordinary panoramic equipment is prone to the problem of serious distortion of the stitched-out images, which also reflects that the research and development technology of panoramic equipment is not yet perfect [11].…”

“…Cortes-Gallardo, Edgar et al discusses and compares three feature extraction methods, SIFT, BRISK, and SURF, for panoramic stitching applications in Autonomous Vehicle Architecture [10]. After experimental validation, the results show that SURF is the most suitable of the three methods.…”

Research on semi-trailer panoramic surround view system based on feature point matching

“…In recent years, there have been many studies on vehicle-mounted panoramic surround view systems, mainly focusing on fisheye camera distortion correction [2] , feature point matching [3] and image fusion technologies [4] . For example, Gallardo et al [5] compared and discussed the advantages and disadvantages of various feature extraction methods in the vehiclemounted panoramic surround view imaging system, and determined the optimal feature extraction algorithm through experiments; Cheng et al [6] studied the generation method of the panoramic surround view system of large trucks, and proposed an adaptive brightness equalization method based on γ transformation, which improved the quality of the final surround view image.…”