摘要
针对大视场经典图像拼接算法流程速度慢,且嵌入式平台算力较低的限制,采用了离线参数计算的实时图像拼接方法。在实际应用场景中,由于多台摄像头的位置是固定的,因此离线计算出的矩阵参数,在实时拼接中可以直接加载进行后续计算,且可重复使用,从而省去了大量的特征提取和配准的时间,在低算力的条件下,在保证对图像拼接质量影响不大的情况下实现了摄像头画面的实时拼接效果。
Aiming at the limitation of the slow process speed of the classical image stitching algorithm with large field of view and the low computing power of the embedded platform,the real-time image stitching method of offline parameter calculation is adopted.In practical application scenarios,since the positions of multiple cameras are fixed,the matrix parameters calculated offline can be directly loaded for subsequent calculation in real-time stitching,and can be reused.In this case,a lot of feature extraction and configuration time can be saved.Under the condition of low computing power,the real-time stitching of the cameras is achieved while ensuring that the quality of the image stitching is not obviously affected.
出处
《工业控制计算机》
2023年第1期119-120,145,共3页
Industrial Control Computer
