Based On The Detection Of Hyperspectral Hami Melon Growth Process Quality

Fruits and vegetables are the main components of people's diet in daily life.China has always been one of the major producers of large fruits and vegetables in the world.Hami melon is a branch of melon.The unique natural conditions make the Hami melon produced in Xinjiang superior in quality,and is favored by the fruit market and the people.The traditional quality assessment methods often rely on external indicators such as fruit size,reticulate distribution,epidermis color,and presence or absence of damage to make subjective judgments.This detection method is inefficient and has low accuracy.Therefore,the research on the rapid,accurate and objective Hami melon quality evaluation technology and evaluation system is particularly urgent under the current Hami melon market situation.The research mainly includes five parts: the monitoring of the ecological environment in different planting areas of Hami melon;the dynamic monitoring of the quality change of Hami melon fruits in different growth periods;the comparison of the quality of Hami melon and the influence of environmental factors on the quality of Hami melon in different planting areas;the high internal quality of Hami melon fruits in different growing stages Spectral detection model;Hami melon fruit hyperspectral detection model in different growing areas.(1)The monitoring results of environmental factors in different planting areas show that different eco-environmental indicators are different in different planting areas,but there are obvious differences in individual environmental factor indicators;(2)The monitoring results of internal and external quality indicators of fruit samples of different growth stages of Hami melon showed that the dynamic changes of internal and external quality indicators of Hami melon were different,and the specific manifestations were as follows: The content of sugar in the fruits during the fruit set-up was increased in the sugar content.Slower,rapid increase in mid-growth content,increase in late maturity and then slow down;protein content trends to increase first and then slightly decrease and then increase;the trend of vitamin C content showed a significant decrease-slow increase-rapid increase The law of change is that the change of titratable acid content is gradually increasing.The growth rate in the fruit setting period is slow,the growth rate in the middle development period is slightly accelerated,and the increase trend in the late maturity period is reduced.From the fruit setting to fruit ripening,the speed of single fruit weight growth of Hami melon fruit showed a “slowquick-slow” process.The fruit shape index did not change much,remained around 1.6,and the melon fruit shape remained oblong during four different sampling periods.(3)The results of comparative analysis on the fruit quality of Hami melon during mature period in different planting areas show that due to the different ecological environment conditions in the planting area,there are significant differences among the different internal qualities of the fruit: the sugar content in the new and planted areas is 28.8% lower than that in the Kashgar planting area.,35.9% lower than the Aksu planting area,sugar content in the Kashgar planting area is 11.1% lower than in the Aksu planting area;titratable acid content in the Aksu planting area reaches 0.59%,55% higher than in the new planting area,than in the Kashgar planting area The highest protein content was 13.5% in the Aksu planting area,8.93% higher than the Kashi planting area,53.14% higher than in the Xinhe planting area,and 40.58% higher in the Kashgar planting area than in the Xinhe planting area;the vitamin C content was highest in the Aksu planting area,compared with Kashgar The planting area is 7.35% higher,53.59% higher than the new planting area,and the Kashgar planting area is 43.08% higher than the new planting area.This paper analyzes the correlation between environmental factors and the quality of Hami melon in the new and planted areas.The results show that there is a very significant correlation between the accumulated sugar content and protein content during the growing season,and there are significant correlations between vitamin C content and other environmental factors.The titration acid content has no obvious correlation.(4)The detection models of different quality indicators of Hami melon fruits established in different growth stages using different spectral pretreatment methods can all achieve better prediction results.For the prediction model of sugar content,the prediction model established by the SG-2 spectra in the first and the fourth sampling period had the best effect;the second and the third sampling period had the best prediction effect after the standard normal transformation spectrum.For the protein content model,the first sampling period is best modeled after the SG-2 processed spectrum;the second sampling period model is the best after the standard normal transformed spectrum;third and fourth The best predictor for the sampling period is the spectrum after SG-1 treatment.For the titratable acid content model,the first,second,and fourth sampling periods were best modeled after the SG-2 processed spectrum;the third sampling period the best effect was after the standard normal transformed SNV transform.The spectrum.For the vitamin C content model,the fruit samples in the first sampling period were best modeled after the standard normalized SNVtreated spectra;the second and fourth sampling periods were the best models for the SG-2 model.The spectrum after the transformation;the best effect in the third sampling period is the spectrum after the SG-1 transformation.(5)For the establishment of the quality index model of melon fruits in the ripening period of the new and growing areas and the Alar cultivation area,both could obtain better prediction results.For the prediction model of sugar content,the model constructed by the SG-2 treated spectrum in Xinlao and Alaer planting areas performed best.For the prediction model of titratable acid content,the model constructed by the SG-2 treated spectra in the Xinhe and Alar planting areas had the best results.For the prediction model of protein content,the model constructed by the SG-1 treated spectrum in Xinlao and Alaer planting areas performed best.For the prediction model of vitamin C content,the model constructed by the SG-2 treated spectra in the Xinhe and Alar planting areas had the best results.