模式識別導論本一PPT學習教案

1、會計學1模式識別模式識別(m sh sh bi)導論本一導論本一第一頁，共26頁。第1頁/共26頁第二頁，共26頁。中南海故宮(ggng)如何(rh)讓計算機自動分析不同地物類？第2頁/共26頁第三頁，共26頁。未成年人(wi chn nin rn)成年人身高(shn o)體重(tzhng)計算機如何自動來進行判別？第3頁/共26頁第四頁，共26頁。計算機如何(rh)自動分析對話？第4頁/共26頁第五頁，共26頁。第5頁/共26頁第六頁，共26頁。第6頁/共26頁第七頁，共26頁。第7頁/共26頁第八頁，共26頁。第8頁/共26頁第九頁，共26頁。第9頁/共26頁第十頁，共26頁。第10頁/共

2、26頁第十一頁，共26頁。第11頁/共26頁第十二頁，共26頁。第12頁/共26頁第十三頁，共26頁。第13頁/共26頁第十四頁，共26頁。第14頁/共26頁第十五頁，共26頁。第15頁/共26頁第十六頁，共26頁。第16頁/共26頁第十七頁，共26頁。第17頁/共26頁第十八頁，共26頁。第18頁/共26頁第十九頁，共26頁。nkjkikijXXd1|nkjkikijXXd12nkjkikqijXXqqd1|1)(第19頁/共26頁第二十頁，共26頁。|max)(1jkiknkijXXd1)(jijiTijXXXXMd第20頁/共26頁第二十一頁，共26頁。nkjknkiknkjkikijX

3、XXXC12121x1x2x1x2x3XXji,第21頁/共26頁第二十二頁，共26頁。nknkjkjikinkjkjikiXXXXXXXXrij11221jiXX,第22頁/共26頁第二十三頁，共26頁。第23頁/共26頁第二十四頁，共26頁。The use of pattern recognition methods in remote sensingHas a long history with air photo-interpreters perhaps the first to use intuitive methodsThe main distinguishing characte

4、ristics between remote sensing data and other imaging data is the data volumeMethods that can be applied in other applications may not be suited to the analysis of remote sensing dataA further point is the discrepancy between the dimensionality of remote sensing data sets and the training data avail

5、ableHughes effect: increasing error with increasing data dimensionality第24頁/共26頁第二十五頁，共26頁。New trends of pattern recognition methods in remote sensing:New and sophisticated decision rules are continued to be introduced for remote sensing data analysisThe use of combination of data derived from different sensors, or from different time period or even extracted from GIS databasesIn addition of spectral information, derived measures such as values of texture and context are usedNew challenges: computational cost, date scale, data reliabilitySophisticated algorithms can not comp