Utilizing Landsat TM Imagery to Map Greenhouses in Qingzhou, Shandong Province, China

To build a rapid and accurate method for greenhouse vegetable landinformation extraction using an index model derived from TM digital data of Qingzhou City, ShandongProvince, based on a systematic analysis of the spectral characteristics of different land use typesin the study area, a subset of the image was first made to eliminate the mountainous region notassociated with vegetable distribution, and then water body pixels were masked. With this the V_Iindex model for greenhouse vegetable land extraction was developed. The index model indicatedgreenhouse vegetable land for Qingzhou in April 2002 was concentrated in the southeast and aroundrural residential areas. Field data used for an accuracy evaluation showed that greenhouse hectaresdetermined with remote sensing were 95.9% accurate, and accuracy for the spatial distribution ofgreenhouse vegetable land cross checked with a random sample was 96.3%. Therefore, this approachprovided an effective method for greenhouse vegetable land information extraction and has potentialsignificance for management of greenhouse vegetable production in the study area, as well as NorthChina.

To build a rapid and accurate method for greenhouse vegetable land information extraction using an index model derived from TM digital data of Qingzhou City, Shandong Province, based on a systematic analysis of the spectral characteristics of different land use types in the study area, a subset of the image was first made to eliminate the mountainous region not associated with vegetable distribution, and then water body pixels were masked. With this the V1 index model for greenhouse vegetable land extraction was developed. The index model indicated greenhouse vegetable land for Qingzhou in April 2002 was concentrated in the southeast and around rural residential areas. Field data used for an accuracy evaluation showed that greenhouse hectares determined with remote sensing were 95.9% accurate, and accuracy for the spatial distribution of greenhouse vegetable land cross checked with a random sample was 96.3%. Therefore, this approach provided an effective method for greenhouse vegetable land information extraction and has potential significance for management of greenhouse vegetable production in the study area, as well as North China.