【正文】 s essay affects me very deeply. Not as an urging to bee self reliant – I39。s Liaoning province, decades ago. The solider gave the old man a handmade saddle when they bid farewell. The story inspired Nasun to write Carved Saddle, a song that later became one of his most popular numbers. Now, every year, Nasun recruits young singers and dancers for the troupe. The troupe has also designed a new repertoire, which is mostly based on the daily lives of Mongolian people, especially the lives of nomadic families, and has bined contemporary musical elements with folk songs of the region. Haimu, a 25yearold khoomei (a local variant of overtone singing) singer, joined the troupe three years ago. Along with a sixmember band, he performs fast songs and soft ones that he writes－ all while playing the horsehead fiddle.Although I learned the piano since childhood and grew up listening to various kinds of music, to me, the folk music of Inner Mongolia is the root, he says. Performing in remote villages is pleasant. I feel at home on the boundless grasslands, and the warm people there make me feel fulfilled. The first round of spring auction season in Beijing ended last week, but it failed to create much spring in the art market. Although two pieces of Chinese painting fetched more than 100 million yuan, the decline in trading volume and sa le rate showed a downturn this year. In the ―Grand View: Chinese Painting Highlight‖ session at China Guardian 2022 spring auctions, Pan Tianshou’s representative work Eagle, Rock and Flora hit a record auction price of 279 million yuan, while Li Keran’s masterwork Jinggang Mountain fetched million yuan, an unexpected high in recent years. However, the trading volume fells to billion yuan from billion yuan in the same period the year before. The Huangchen 2022 Spring Auctions, which recorded million yuan in total sales, experienced the same. The section number went down to 5 from 12 pared to last year. According to expert Shao Jianwu, the art market did not attract much excitement this year due to the booming stock market and the persistent problems of fery and fake deals. The two pieces of Chinese painting notched up high price this spring due to their own value not due to a revival in the market. The successful transaction of two works with a hammer price of more than 100 million yuan once again confirms that valuable and rare works of high quality always earn the longlasting high price. Chinese painting and calligraphy is still the pillar of the market. However, the overall sales rate has dropped slightly, which reflects the demand for the ordinary auction goods has weakened,‖ said the president of China Guardian’s Hu Yanyan. Besides, the entry of new buyers also makes certain changes to the direction of purchasing. In a word, the overall market trend is very unpredictable. Although China Guardian 2022 Spring Auctions has achieved remarkable results, it does not mean that the market has walked out of the predicament. We still have to wait for the big economic boom to invigorate the art market.‖ Even so, the hot pursuit of some special items appeared this spring. A special session of China Guardian Spring Auctions titled ―Fine Gilt—Bronze Buddhist Images‖ achieved a great deal with total sales volume of million yuan and 92 percent sale rate. The 5th Shamar Rinpoche Statue from 1617th century of Tibet was sold for million yuan. The Beijing Chengxuan Auctions featured almost 2,800 items of coins and stamps in three sessions with good sale rate. The Huachen Auctions also set a special session of photographs under the condition of largescale decline of auction sessions. There was a palpable dull thud of disappointment that acpanied the return of the imperial entourage of Zhen Huan to her homeland. It followed a couple years of hushed excitement as Chinese fans were fed tidbits about their proud concubine who was supposed to conquer the high ground of the North American market. Zhen Huan is, of course, the title character of The Legend of Zhen Huan, a 2022 television series that swept China off its feet and later took other Asian countries by storm. Two years ago, it was reported that HBO, a premium cable service headquartered in the United States, was going to air it in North America after some modification. Now, a condensed version that provides English subtitles but no dubbing has finally been made available on Netflix for online streaming. This version, highly anticipated as a milestone in China39。80s. We sat on the back of pickup trucks for hours. The sky was blue, and we couldn39。本文之所以能夠順利完成，和呂老師的幫助是分不開的，值此我表示對導(dǎo)師深深的謝意。 其次是現(xiàn)在神經(jīng)網(wǎng)絡(luò)模型的確定還沒有成熟的理論依據(jù) ,在理論上如何實(shí)現(xiàn)神經(jīng)網(wǎng)絡(luò)模型最優(yōu)化是一個難點(diǎn) ,也是神經(jīng)網(wǎng)絡(luò)廣泛應(yīng)用的瓶頸。但由于所提取的樹葉形態(tài)特征存在差異，也出現(xiàn)了識別錯誤的現(xiàn)象。第三部分是樹葉圖像特征提取，它的作用是把圖像預(yù)處理后的樣本進(jìn)行選擇某些特征參數(shù) ,并進(jìn)行分析。通過不斷調(diào)整隱元，觀測實(shí)驗(yàn)結(jié)果獲取最佳的隱含層單元數(shù)為 8。常用的人工神經(jīng)網(wǎng)絡(luò)包括 RBF 網(wǎng)絡(luò)、 BP 網(wǎng)絡(luò)、Hopfield 網(wǎng)絡(luò)等。其中長軸是樹葉輪廓上任意過質(zhì)心的兩點(diǎn)直線的最長距離，短軸是長軸兩側(cè)與其距離最長的左右兩點(diǎn)的距離和。 figure,imshow(BW_laplace),xlabel(39。sobel 檢測 39。)。gaussian39。該算子既能檢測邊緣點(diǎn)，還能抑制噪聲的影響。 %顯示網(wǎng)絡(luò)線 axis on。原圖 39。 axis([0,600,0,600])。 8 圖像分割 從圖像采集到圖像分析必然要進(jìn)行 圖像分割，圖像分割 是把 圖像 分成若干個特定區(qū)域。5*5 中值濾波 39。)。,5),K)/255。 K=imnoise(J,39。 均值濾波是用模板內(nèi)的全體像素點(diǎn)的均值或加權(quán)均值來代替原始值；其數(shù)學(xué)公式如21： 設(shè) ? ?,f i j 是含有噪聲的圖像， ? ?,gi j 是經(jīng)過理后的圖像， ? ? ? ? ? ?, , / , ,g i j f i j N i j M??? (21) 其中 M 是鄰域內(nèi)各鄰近像素的坐標(biāo)值， N 是鄰域中臨近像素的個數(shù)。 2 樹葉圖像預(yù)處理 首先闡述了基于神經(jīng)網(wǎng)絡(luò)的樹葉識別流程圖，然后詳細(xì)介紹了有關(guān)圖像采集、裁剪、去噪、邊緣提取等操作的理論、程序及效果圖。 4 神經(jīng)網(wǎng)絡(luò)的研究進(jìn)展 人工神經(jīng)網(wǎng)絡(luò)，簡稱 ANN，它 是由很多神經(jīng)元相互連接而組成的一個信息處理系統(tǒng)。利用數(shù)字圖像處理技術(shù)對植物葉片分類是完全可行的。propose method based on BP neural work recognition of the leaves ,besides ,constructed a model based on neural work ensemble classifier .Finally carried on the experiment test to this system ,hen get to the good result . Key Words： image processing； neura