空地数据链系统中图像自动剪切方法研究

An Image Automatic Clipping Method for Air-ground Data Link System

在空地数据链路通信中,鉴于存在需要传输的各种机载图像传感器的数据量越来越大与已有的数据链带宽相对较小的矛盾,为了提高数据链路在战场环境中的图像传输效率,对于用CCD航空相机以中心投影方式获取的静态图像数据,要求远程快速高保真地进行信息传递,提出了一种机载自动数据缩减方案。该方案的关键步骤是图像的自动匹配与剪切,它的实现算法利用了航拍相邻图像间具有强逻辑相关性的特点,基于二维仿射变换模型,使用通过傅里叶-梅林变换计算得到的图像匹配参数,在没有检测到突现目标的情况下,对相邻图像中的场景重叠部分进行自动剪切操作。自动剪切及镶嵌的试验结果表明,本算法复杂度较低,鲁棒性和可操作性较强,能够有效地缩减下行链路中待传输的数据量,具有进一步实用化的潜力。

In communications using air-ground data link, the data that comes from various kinds of airborne image sensors is becoming huger and huger, but the bandwidth of the data link is relatively small. In order to improve image transmission efficiency in battle field condition, there is a requirement for CCD aerial camera using central projection acquired images to be transmitted with high fidelity and speed over long distance. This paper proposed an airborne automatic data reduction scheme to solve the aforementioned problem. The key of this scheme is automatic image matching and clipping. The algorithm utilizes the strong logic relativity between adjacent aero-photograph and it is based on the model of 2-D affine transform. Under the conditions of detecting no highlight moving target, the algorithm performs automatic clipping on overlapping part of two adjacent images by the use of Fourier-Mellin transform. The experimental results of automatic clipping and mosaic show that the proposed algorithm is computationally efficient and very easy to realize. It is a robust algorithm and can effectively reduce data quantity of the downside data link. All these prove that the proposed algorithm has great potential of being applied to practical use.