| 摘要: |
| 为达到组培苗分割移植机器人的嵌入式控制系统对图像处理器小型化和快速化的要求 ,采用基于TMS32 0C6 711的嵌入式图像处理平台 ,开发了运行于该平台的软件系统 ,实现了组培苗识别算法在该平台上的移植。针对DSP软件开发难度大、涉及因素多 ,以及在本研究中需要大量数据存取的情况 ,在保证功能的前提下尽量精简软件结构 ,编写了轻量简洁的Bootloader和适应大量数据存取的连接命令文件。对软件进行了优化 ,将算法中占用时间较长的环节用线性汇编语言编写 ,使总处理时间降低了近 80ms。该图像处理器实验室测试结果为 ,苗节点识别准确率 96 % ,每棵苗的处理时间为 0 7～ 0 8s ,10× 2 4h连续运转稳定 ,性能达到设计要求。该图像处理器的体积只有PC机的 1/ 2 0 ,达到了小型化的要求。 |
| 关键词: 识别算法 DSP TMS320C6711 软件开发 移植 |
| DOI：10.11841/j.issn.1007-4333.2004.06.160 |
| 修订日期：2004-06-14 |
| 基金项目:国家自然科学基金资助项目 (6 0 375 0 36 ) |
|
| Realization of plantlet recognition algorithm on a TMS320C6711 based embedded image processing system |
|
|
|
| Abstract: |
| An embedded image processing system based on TMS320C6711, as well as its software system, was adopted to realize the transplantation of the tissue culture plantlet recognition algorithm on the system, which satisfied the needs of an image processor with high speed and small size used for a tissue culture plantlet robot. The crucial problems, such as programming bootloader, interrupt vector, command file, and assigning memory space, optimizing code and achieving mix-language programming, were discussed aiming at the characteristics of the embedded system during the software development. The system has a good stability, strong real-time capability, and high accuracy. |
| Key words: recognition algorithm,DSP,TMS320C6711,software exploitation,transplantation |
