采摘機(jī)器人果實夜間自動識別技術(shù)研究.doc
約45頁DOC格式手機(jī)打開展開
采摘機(jī)器人果實夜間自動識別技術(shù)研究,1.85萬字自己的畢業(yè)設(shè)計,原創(chuàng),僅在本站獨家提交,大家放心使用摘要 隨著工業(yè)科技日新月異的發(fā)展,農(nóng)業(yè)也開始走上現(xiàn)代化道路,傳統(tǒng)農(nóng)業(yè)耗費的人力物力都過于巨大,因此農(nóng)業(yè)自動化的實現(xiàn)迫在眉睫。果實目標(biāo)的識別與定位是農(nóng)業(yè)自動化領(lǐng)域中一種非常重要的技術(shù),而今在該領(lǐng)域國內(nèi)外已經(jīng)取得巨大的成就,但...
內(nèi)容介紹
此文檔由會員 jiji888 發(fā)布
采摘機(jī)器人果實夜間自動識別技術(shù)研究
1.85萬字
自己的畢業(yè)設(shè)計,原創(chuàng),僅在本站獨家提交,大家放心使用
摘要 隨著工業(yè)科技日新月異的發(fā)展,農(nóng)業(yè)也開始走上現(xiàn)代化道路,傳統(tǒng)農(nóng)業(yè)耗費的人力物力都過于巨大,因此農(nóng)業(yè)自動化的實現(xiàn)迫在眉睫。果實目標(biāo)的識別與定位是農(nóng)業(yè)自動化領(lǐng)域中一種非常重要的技術(shù),而今在該領(lǐng)域國內(nèi)外已經(jīng)取得巨大的成就,但是人們已經(jīng)不僅僅滿足于白天進(jìn)行果實采摘活動。人們發(fā)現(xiàn)夜晚進(jìn)行機(jī)器采摘活動不僅能避開白天炎熱的天氣,還能縮短采摘周期,減少果實來不及采摘而腐爛等問題,能夠更好的提高機(jī)器采摘的效率。然而如何在夜間條件下準(zhǔn)確的識別仍然是該領(lǐng)域的一個經(jīng)典難題,它涉及面廣,是農(nóng)業(yè)機(jī)器人視覺系統(tǒng)功能的主要分支。
本文主要從以下三個方面對夜間果實目標(biāo)的識別進(jìn)行研究分析。首先,在圖像處理的去噪環(huán)節(jié)中,綜合對比了均值濾波、小波去噪和中值濾波三種濾波方法,并最終選用了較適合本系統(tǒng)的中值濾波;其次,對圖像進(jìn)行增強(qiáng)處理,分別比較研究了直方圖均衡化、直接對比度增強(qiáng)方法與對比度受限自適應(yīng)直方圖均衡化方法,選定效果最好的第三種方法;最后,對已成熟的夜間果實圖像進(jìn)行分割時,通過對閾值分割法和區(qū)域分割法進(jìn)行研究,采用了一種在I1I2I3顏色空間下基于I2顏色因子的OTSU動態(tài)閾值分割法進(jìn)行初分割,并結(jié)合區(qū)域分割法中的閾值面積消去法進(jìn)行再分割,此后,為了進(jìn)一步去除噪聲,對分割后的圖像進(jìn)行數(shù)學(xué)形態(tài)學(xué)處理,得到了較為理想清晰的圖像,最終實現(xiàn)了夜間條件下對成熟果實的準(zhǔn)確識別。
關(guān)鍵詞:夜間 果實目標(biāo) 識別 圖像處理
The fruit picking robot automatic identification technology research at night
Abstract With the rapid development of industrial technology,agriculture also embarked on the road of modernization,traditional agriculture costs too much in human and material,so the realization of agricultural automation is imminent. Fruit and vegetable identification and location of the target is a very important technology in the field of automation of agriculture research in the field,although at home and abroad have achieved certain results,people have not only satisfied with the fruit picking activities during the day. It was found that picking activities in night can not only avoid the hot weather during the day,but also shorten the harvest period,reduce the fruit rot for late picking and other issues. However,how to accurate identification and location is still a classic problem in the field it involves a wide range of the main branches of the agricultural robot vision system functions.
We were studied and analyzed for the nighttime identification and location of the above three aspects of the fruit and vegetable target in this paper. At the first,in the image processing to remove noise in the link,we compared the average filtering,median filtering and wavelet denoising. Eventually,we choosed median filtering which is more relatively suitable for the system. Secondly,the method of Histogram equalization was attempted to used in the nighttime image,and we also compared the Directly contrast enhancement method with the Contrast limited adaptive histogram equalization method,we finally choosed the third one. At last,in the Image segmentation of ripe fruit at night, we adopted a I1I2I3 color space based on I2 color factor under the dynamic OTSU threshold segmentation method for initial segmentation,and connected the regional segmentation method of threshold value area expunction method for segmentation based on the research of threshold segmentation and region segmentation. After that,in order to remove the noise further,we did the mathematical morphology processing at the segmentation image,got a more clear ideal image,ultimately realized the accurate identification of the ripe fruit of nighttime conditions.
Key words:Nighttime fruit target recognition image processing
目 錄
第一章 緒 論 1
1.1課題研究意義 1
1.2果實采摘機(jī)器人概述及特點 1
1.3國內(nèi)外研究現(xiàn)狀 2
1.4 研究的內(nèi)容和目標(biāo) 4
1.5 論文的體系結(jié)構(gòu) 4
第二章 夜間果實圖像去噪處理 5
2.1 噪聲來源及去噪的意義 5
2.2 噪聲的分類及其特點 5
2.3 圖像中的噪聲 6
2.4 濾波去噪法 6
2.4.1 均值濾波 7
2.4.2 小波變換 8
2.4.3 中值濾波 8
2.5 本章小結(jié) 10
第三章 夜間圖像增強(qiáng)處理 12
3.1 圖像增強(qiáng)的含義與意義 12
3.2 圖像增強(qiáng)算法及分類 13
3.2.1 基于空域的算法 13
3.2.2 基于頻域的算法 13
3.3 常用的圖像增強(qiáng)算法 13
3.3.1 直方圖均衡化 13
3.3.2 直接對比度增強(qiáng)算法 15
3.3.3 對比度受限自適應(yīng)直方圖均衡化(CLAHE) 16
3.4 本章小結(jié) 17
第四章 果實圖像分割及數(shù)學(xué)形態(tài)學(xué)處理 18
4.1 圖像分割基本概念 18
4.1.1 圖像分割的定義 18
4.1.2 圖像分割法分類 19
4.2 閾值分割法 19
4.2.1 灰度閾值分割法 20
4.2.2 直方圖閾值分割法 20
4.3 區(qū)域分割法 27
4.3.1 閾值面積消去法 27
4.4 數(shù)學(xué)形態(tài)學(xué)處理 28
4.4.1 腐蝕 28
4.4.2 膨脹 28
4.4.3 開運算 29
4.4.4 閉運算 29
4.4.5 已成熟的芒果圖像形態(tài)處理 30
4.5 本章小結(jié) 31
第五章 總結(jié)與展望 32
5.1 論文工作總結(jié) 32
5.2 展 望 32
致 謝 34
參考文獻(xiàn)35
附 錄38
1.85萬字
自己的畢業(yè)設(shè)計,原創(chuàng),僅在本站獨家提交,大家放心使用
摘要 隨著工業(yè)科技日新月異的發(fā)展,農(nóng)業(yè)也開始走上現(xiàn)代化道路,傳統(tǒng)農(nóng)業(yè)耗費的人力物力都過于巨大,因此農(nóng)業(yè)自動化的實現(xiàn)迫在眉睫。果實目標(biāo)的識別與定位是農(nóng)業(yè)自動化領(lǐng)域中一種非常重要的技術(shù),而今在該領(lǐng)域國內(nèi)外已經(jīng)取得巨大的成就,但是人們已經(jīng)不僅僅滿足于白天進(jìn)行果實采摘活動。人們發(fā)現(xiàn)夜晚進(jìn)行機(jī)器采摘活動不僅能避開白天炎熱的天氣,還能縮短采摘周期,減少果實來不及采摘而腐爛等問題,能夠更好的提高機(jī)器采摘的效率。然而如何在夜間條件下準(zhǔn)確的識別仍然是該領(lǐng)域的一個經(jīng)典難題,它涉及面廣,是農(nóng)業(yè)機(jī)器人視覺系統(tǒng)功能的主要分支。
本文主要從以下三個方面對夜間果實目標(biāo)的識別進(jìn)行研究分析。首先,在圖像處理的去噪環(huán)節(jié)中,綜合對比了均值濾波、小波去噪和中值濾波三種濾波方法,并最終選用了較適合本系統(tǒng)的中值濾波;其次,對圖像進(jìn)行增強(qiáng)處理,分別比較研究了直方圖均衡化、直接對比度增強(qiáng)方法與對比度受限自適應(yīng)直方圖均衡化方法,選定效果最好的第三種方法;最后,對已成熟的夜間果實圖像進(jìn)行分割時,通過對閾值分割法和區(qū)域分割法進(jìn)行研究,采用了一種在I1I2I3顏色空間下基于I2顏色因子的OTSU動態(tài)閾值分割法進(jìn)行初分割,并結(jié)合區(qū)域分割法中的閾值面積消去法進(jìn)行再分割,此后,為了進(jìn)一步去除噪聲,對分割后的圖像進(jìn)行數(shù)學(xué)形態(tài)學(xué)處理,得到了較為理想清晰的圖像,最終實現(xiàn)了夜間條件下對成熟果實的準(zhǔn)確識別。
關(guān)鍵詞:夜間 果實目標(biāo) 識別 圖像處理
The fruit picking robot automatic identification technology research at night
Abstract With the rapid development of industrial technology,agriculture also embarked on the road of modernization,traditional agriculture costs too much in human and material,so the realization of agricultural automation is imminent. Fruit and vegetable identification and location of the target is a very important technology in the field of automation of agriculture research in the field,although at home and abroad have achieved certain results,people have not only satisfied with the fruit picking activities during the day. It was found that picking activities in night can not only avoid the hot weather during the day,but also shorten the harvest period,reduce the fruit rot for late picking and other issues. However,how to accurate identification and location is still a classic problem in the field it involves a wide range of the main branches of the agricultural robot vision system functions.
We were studied and analyzed for the nighttime identification and location of the above three aspects of the fruit and vegetable target in this paper. At the first,in the image processing to remove noise in the link,we compared the average filtering,median filtering and wavelet denoising. Eventually,we choosed median filtering which is more relatively suitable for the system. Secondly,the method of Histogram equalization was attempted to used in the nighttime image,and we also compared the Directly contrast enhancement method with the Contrast limited adaptive histogram equalization method,we finally choosed the third one. At last,in the Image segmentation of ripe fruit at night, we adopted a I1I2I3 color space based on I2 color factor under the dynamic OTSU threshold segmentation method for initial segmentation,and connected the regional segmentation method of threshold value area expunction method for segmentation based on the research of threshold segmentation and region segmentation. After that,in order to remove the noise further,we did the mathematical morphology processing at the segmentation image,got a more clear ideal image,ultimately realized the accurate identification of the ripe fruit of nighttime conditions.
Key words:Nighttime fruit target recognition image processing
目 錄
第一章 緒 論 1
1.1課題研究意義 1
1.2果實采摘機(jī)器人概述及特點 1
1.3國內(nèi)外研究現(xiàn)狀 2
1.4 研究的內(nèi)容和目標(biāo) 4
1.5 論文的體系結(jié)構(gòu) 4
第二章 夜間果實圖像去噪處理 5
2.1 噪聲來源及去噪的意義 5
2.2 噪聲的分類及其特點 5
2.3 圖像中的噪聲 6
2.4 濾波去噪法 6
2.4.1 均值濾波 7
2.4.2 小波變換 8
2.4.3 中值濾波 8
2.5 本章小結(jié) 10
第三章 夜間圖像增強(qiáng)處理 12
3.1 圖像增強(qiáng)的含義與意義 12
3.2 圖像增強(qiáng)算法及分類 13
3.2.1 基于空域的算法 13
3.2.2 基于頻域的算法 13
3.3 常用的圖像增強(qiáng)算法 13
3.3.1 直方圖均衡化 13
3.3.2 直接對比度增強(qiáng)算法 15
3.3.3 對比度受限自適應(yīng)直方圖均衡化(CLAHE) 16
3.4 本章小結(jié) 17
第四章 果實圖像分割及數(shù)學(xué)形態(tài)學(xué)處理 18
4.1 圖像分割基本概念 18
4.1.1 圖像分割的定義 18
4.1.2 圖像分割法分類 19
4.2 閾值分割法 19
4.2.1 灰度閾值分割法 20
4.2.2 直方圖閾值分割法 20
4.3 區(qū)域分割法 27
4.3.1 閾值面積消去法 27
4.4 數(shù)學(xué)形態(tài)學(xué)處理 28
4.4.1 腐蝕 28
4.4.2 膨脹 28
4.4.3 開運算 29
4.4.4 閉運算 29
4.4.5 已成熟的芒果圖像形態(tài)處理 30
4.5 本章小結(jié) 31
第五章 總結(jié)與展望 32
5.1 論文工作總結(jié) 32
5.2 展 望 32
致 謝 34
參考文獻(xiàn)35
附 錄38