CAJ | 학술논문

精准识别农作物对于及时准确估计农作物种植面积、产量等关键农情信息具有重要意义。合成孔径雷达（synthetic aperture radar，SAR）以其不受云雨天气影响，可全天时、全天候监测等优点已被广泛应用于农情遥感监测领域，为大区域尺度的农作物遥感识别提供了强有力的数据保障和技术支持。该文以雷达技术的发展进程为论述主线，对20余年来国内外农作物SAR识别研究与实践应用的新进展进行了系统总结，具体归纳为4个方面：早期研究（20世纪80年代末-2002年），特征是以单波段、单极化、多时相SAR数据为主；基于多极化、多波段SAR数据进行农作物识别与面积监测研究；利用SAR与光学遥感相结合提高农作物的识别精度与效率研究；农作物SAR分类算法研究。在今后农作物SAR识别研究中，对于复杂种植结构背景下的旱地作物识别，如何优化组合 SAR 系统工作参数（极化方式、频率及入射角等）及与光学遥感融合来提高农作物识别精度与时效性，发展机理性的农作物SAR分类算法将是需要重点解决的3个问题。
정준식별농작물대우급시준학고계농작물충식면적、산량등관건농정신식구유중요의의。합성공경뢰체（synthetic aperture radar，SAR）이기불수운우천기영향，가전천시、전천후감측등우점이피엄범응용우농정요감감측영역，위대구역척도적농작물요감식별제공료강유력적수거보장화기술지지。해문이뢰체기술적발전진정위논술주선，대20여년래국내외농작물SAR식별연구여실천응용적신진전진행료계통총결，구체귀납위4개방면：조기연구（20세기80년대말-2002년），특정시이단파단、단겁화、다시상SAR수거위주；기우다겁화、다파단SAR수거진행농작물식별여면적감측연구；이용SAR여광학요감상결합제고농작물적식별정도여효솔연구；농작물SAR분류산법연구。재금후농작물SAR식별연구중，대우복잡충식결구배경하적한지작물식별，여하우화조합 SAR 계통공작삼수（겁화방식、빈솔급입사각등）급여광학요감융합래제고농작물식별정도여시효성，발전궤이성적농작물SAR분류산법장시수요중점해결적3개문제。
Crop recognition is the initial phase and key link of an agricultural condition monitoring system. The accurate identification of a crop can achieve a good estimation for crop sown acreage, planting structure, and spatial distribution, as well as provide key input parameters for a crop yield estimation model. Due to that crop sown acreage, yield information is the important basis for making national food policy and an economic plan. Therefore, it is very important to conduct the study on crop identification. In view of the advantages of high temporal resolution, wide coverage, and low cost, remote sensing has been used in a wide array of earth observation activities, and thus provides a useful tool for crop recognition and planting acreage monitoring on a large scale. Since the 1980’s, optical remote sensing has been widely used to identify various crops, and consequently, it has made obvious progress, no matter whether in the aspect of theory and technology. However, optical images are not often available in the key growth period of crops, owing to the cloudy and rainy weather. Thus, it has a negative effect on the accuracy and timeliness of crop area monitoring. As a new high-technology with an advantage of all-weather, all-time, high resolution, and wide coverage, synthetic aperture radar (SAR) has been widely applied in the agricultural condition monitoring field and thus provides a strong complement and support for crop identification in the data and technology aspects. As the updating and improvement of function parameters and performance index of radar sensors, it has been an important field of agriculture remote sensing in obtaining the information of crop sown acreage, growing condition, and yield by SAR. In this paper, according to a mainline of the development progress of radar technology in the recent twenty years, the progress of studies and applications on crop discrimination by SAR is systematically summarized, and the conclusion includes four aspects:the first is that early studies (from the late 1980’s to 2002), are characterized by using single band, single polarization, and multi-temporal SAR data for crop identification;The second is crop acreage monitoring based on multi-polarization, multiband SAR data. Furthermore, this section can be divided into two subsections:one is crop recognition by multi-polarization SAR, the other is using multiple SAR sensors for crop classification;The third is studies on improving the accuracy and efficiency of crop identification by combining SAR with optical remote sensing;The last is the studies on crop classification algorithm using SAR data. According to the summary of previous studies, the problems existing in the crop identification by SAR can be analyzed as follows:the first is that crop types identified by SAR are still single;the second is that the accuracies of crop identification are not yet high;the last is that mechanism studies on the classification algorithm are lacking. Furthermore, the development trends are presented in this study. Dryland crop discrimination using SAR images under a complex crop planting structure, improving the accuracy and timeliness of crop identification by optimizing the operational parameters (e.g. polarization, frequency and incidence angle) of SAR system and combining it with optical remote sensing, and developing the mechanism-based algorithm of crop classification will be three area that will urgently be needed to be studied in the future.