The Design Of Infrared And Visible Image Matching Scheme Based On Position Relation Constraint

According to the survey,China's per capita arable land area is only1.64 mu,only 50.52% of the world level.The small area of cultivated land,coupled with abnormal crop growth,will reduce yield and affect farmers' income and crop futures market.In order to be informed of crop anomalies in a timely manner,it is necessary to collect crop growth information in an efficient manner.In this paper,using the characteristics of abnormal temperature change spree surface change situated by drones,using infrared thermal imaging technology to detect the temperature of plant canopy and predict crop abnormalities.Although the drone's infrared camera efficiently obtains infrared images of crop canopy temperature,it does not know the specific geographic information of the abnormal temperature area and therefore does not provide an accurate target location for the prevention and control operation.In view of this problem,this paper proposes to match the infrared image with the map visible light image,and to get the correspondence between the two pixels,so as to combine infrared and visible light images,the main contributions of this paper are as follows:(1)Based on the study of the existing infrared and visible light image matching scheme,combined with the characteristics of infrared and visible light image used in this paper and the final matching purpose,the matching scheme of infrared and visible light image from coarse to fine is proposed.First of all,the matching image is pre-processed,making full use of the significance information in the image,the second is to match the infrared map with the pre-selected template in the visible light map,the approximate area of the infrared image is determined according to the confidence of the matching related values,and then the effective template in the combined area is matched point pair,and finally the matching point pair is used to combine the infrared with the visible light image.(2)In this paper,three matching methods are selected for validation experiments.By analyzing the experimental effect,the region-based matching algorithm is selected as the matching algorithm designed by the scheme.In view of the local optimal and time-consuming problems of this method,the location relationship constraints of the significant areas and the algorithms related to the covariance are simplified respectively.(3)In this paper,using the designed infrared and visible image matching scheme,the 4 infrared images and visible light images successfully completed from coarse to fine match.By comparing with the SIFT matching method based on feature points,it is shown that the infrared and visible light images proposed in this paper have achieved good results from coarse matching to region matching.In summary,the infrared and visible image matching scheme and the matching algorithm based on location relationship constraints designed in this paper realize the match from coarse match to the region under this application,and effectively combine the infrared image with the map data with the geographic information.Provide the necessary data support for later crop scientific growth management,which promotes automation and intelligence of agricultural management.