CN103792235A - Diffuse transmission spectrum and image information fusion method for detecting internal quality of honeydew melons on line and device - Google Patents

Abstract

The invention relates to a method and device for online detection of the internal quality of honeydew melons based on fusion of diffuse transmission spectrum and image information. Signal control unit, diffuse transmission spectrum acquisition device, honeydew melon image acquisition device, honeydew melon internal quality detection software system, the conveying device is provided with tow rings for holding honeydew melons, and each tow ring passes through the honeydew melons in turn The image acquisition room of the image acquisition device and the spectrum acquisition room of the diffuse transmission spectrum acquisition device; the laser sensor and two industrial cameras in the image acquisition room, as well as the laser sensor and the near-infrared spectrometer in the spectrum acquisition room communicate with the honeycomb device through the signal control unit. It is connected with the computer of the internal quality inspection software system of melon. It can effectively detect the internal quality of the honeydew melon, and overcomes the deficiency that the internal quality index of the honeydew melon cannot be accurately detected through image information.

Description

On-line detection method and device for internal quality of honeydew melon based on fusion of diffuse transmission spectrum and image information

the

technical field

The invention relates to an online non-destructive detection method and device for the internal quality of honeydew melons, in particular to an online detection method and device for the internal quality of honeydew melons using diffuse transmission spectrum and image information fusion. the

the

Background technique

Honeydew melon is an important economic crop in Northwest China. For a long time, the maturity and quality classification of honeydew melons in the sales process has only been done by traditional manual methods or destructive sampling inspections. Good and bad are mixed, which reduces the market competitiveness and sales price of the fruit. To solve the above-mentioned problems of honeydew melons, we cannot simply rely on cultivating new varieties, improving planting and storage and transportation conditions, but more attention should be paid to quality testing and grading processing technologies in the process of commercialization. the

Quality grading is an important link in the commercialization of honeydew melons. The main indicators that determine the quality of honeydew melons are its internal quality indicators (sugar content, hardness). It contains its internal quality information, as well as its external feature information, such as shape and size information, and the difference in shape and size of honeydew melon has a significant impact on the diffuse transmission spectrum, which in turn affects the quality detection accuracy. Currently, there is no reliable way to remove these effects from diffuse transmission spectra. Therefore, if we know the shape and size information of honeydew melon when analyzing its diffuse transmission spectrum, and then consider the influence of shape and size on the spectrum when establishing a detection model, the detection accuracy will definitely be improved. The image has a unique advantage in the rapid and non-destructive detection of the appearance characteristics of the fruit. It can not only detect the shape and size of the honeydew melon, but also detect its appearance color characteristics, and the color characteristics have a certain correlation with the internal quality of the honeydew melon. Therefore, if the spectral information and image information are fused, the internal quality of honeydew melons can be detected with the help of multi-sensor information fusion, which has better performance than a single sensor. The feature that the appearance and color characteristics of honeydew melon are correlated with its internal quality can be used to improve the detection accuracy, and at the same time, it also overcomes the deficiency that the image cannot go deep into the inside of honeydew melon for accurate detection of quality indicators. Therefore, it is of great significance to carry out the research on the internal quality detection method and device based on the fusion of honeydew melon diffuse transmission spectrum and image information. the

the

Contents of the invention

In order to overcome the defects of the prior art, the present invention provides an on-line detection method and device for the internal quality of honeydew melons by fusion of diffuse transmission spectrum and image information. The online detection method for the internal quality of melons includes the steps of establishing a non-destructive detection model for the internal quality of honeydew melons and performing online detection of the internal quality of honeydew melon samples, specifically as follows:

(1) Establish a model for non-destructive testing of the internal quality of honeydew melon:

a. Carry out online diffuse transmission spectroscopy and image acquisition of the front and side of honeydew melons for batches of honeydew melon samples;

b. Calibrate and preprocess the collected sample spectra;

c. Extract quality index sugar content and hardness sensitive spectral information, and establish a model of characteristic bands;

d. Preprocessing the collected honeydew melon images, extracting quality index sugar content and hardness-sensitive color feature subsets;

e. Calculate the volume of honeydew melon;

f, calculate the fruit shape index of honeydew melon;

g. The color feature subset, volume, and fruit shape index are fused as image feature variables and spectral information to establish an online detection model for the internal quality of honeydew melons.

(2) After establishing the online detection model of the internal quality of honeydew melon, conduct online detection of the internal quality of unknown honeydew melon samples:

a. Place the honeydew melon horizontally on the drag ring of the conveying device, and move forward at a constant speed;

b. When the melon reaches the diffuse transmission spectrum collection room, the built-in laser sensor in the collection room outputs a melon arrival feedback signal to the signal control unit, and the signal control unit sends a command to trigger the near-infrared spectrometer, and the near-infrared spectrometer starts to collect the measured honey For the spectral information of the melon, after the honeydew melon passes, the laser sensor outputs a feedback signal of the honeydew melon leaving to the signal control unit, and the signal control unit issues a command to stop the near-infrared spectrometer from working;

c. When the honeydew melon arrives at the honeydew melon image acquisition room, the laser sensor in the image acquisition room outputs a honeydew melon arrival feedback signal to the signal control unit, and the signal control unit sends a command to trigger two industrial cameras, and the two industrial cameras respectively capture the Measure the front image information and side image information of the honeydew melon. After the honeydew melon passes, the laser sensor in the image acquisition room outputs a feedback signal of the honeydew melon leaving to the signal control unit, and the signal control unit issues an order to stop the industrial camera from working;

d. Input the collected spectrum and image information into the online detection model of the internal quality of honeydew melon established in step (1), obtain the internal quality of the measured honeydew melon, and then judge the grade of honeydew melon;

A device for the online detection method of the internal quality of honeydew melon based on the fusion of diffuse transmission spectrum and image information, including a conveying device, a signal control unit, a diffuse transmission spectrum acquisition device, a honeydew melon image acquisition device and a computer, and the conveyor belt of the conveying device is set There are drag rings holding honeydew melons, and each drag ring passes through the image acquisition room of the honeydew melon image acquisition device and the spectrum acquisition room of the diffuse transmission spectrum acquisition device in turn; the laser sensor in the image acquisition room and two industrial cameras, spectrum The laser sensor and near-infrared spectrometer in the collection room are all connected to the computer installed with the internal quality detection software system of honeydew melon through the signal control unit.

The conveying device (19) includes a support connection frame (26) and a first belt support roller (27), a second belt support roller (28), a first drive sprocket ( 29), the driving chain (30), the second driving sprocket (31), the speed regulating motor (32) and the conveyor belt (5), the second driving sprocket (31) is installed on the speed regulating motor (32), the first The drive sprocket (29) is installed on the first belt support roller (27), and the first drive sprocket (29) and the second drive sprocket (31) are connected by drive chain (30), and the conveyor belt ( 5) Wrap around the periphery of the first belt support roller 27 and the second belt support roller (28) on the supporting link frame, and several drag rings are evenly fixed on the conveyor belt (5). the

The diffuse transmission spectrum acquisition device includes a spectrum acquisition room (18), an arc-shaped light source fixing frame (8), a halogen light source (10), a laser sensor (7) in the spectrum acquisition room, an optical fiber probe (15), a near Infrared spectrometer (21), wherein the halogen light source (10), the laser sensor (7) in the spectrum collection room, and the arc-shaped light source fixing frame (8) are located inside the spectrum collection room (18), and the halogen light source (10) is evenly distributed On the arc-shaped light source fixing frame (8), two arc-shaped light source fixing frames (8) are arranged oppositely, and the drag ring passes through the middle of the two arc-shaped light source fixing frames (8); in the spectrum collection room (18 ) is provided with a vent on the top of the vent, a fan (9) is installed on the top of the vent, the optical fiber probe (15) is located at the bottom of the drag ring, and is connected to the near-infrared spectrometer (21) through an optical fiber. the

The honeydew image acquisition device includes a honeydew image acquisition room (11), a first industrial camera (1), a second industrial camera (4), an LED light source (3), a laser sensor (2) in the image acquisition room, Among them, the first industrial camera (1), the second industrial camera (4), the laser sensor (2) and the LED light source (3) in the image acquisition room are located inside the honeydew image acquisition room (11), and the laser sensors in the two image acquisition rooms The sensors (2) are arranged oppositely, and the LED light sources (3) are respectively arranged at the four corners of the honeydew melon image collection room; the first industrial camera (1) and the second industrial camera (4) collect the frontal image information and the side image information of the measured honeydew melon respectively. image information. the

There are two conveyor belts (5), which are arranged parallel to each other, forming a gap in the middle; the two conveyor belts are installed on the support connecting frame through the first belt support roller and the second belt support roller, and the two ends of the drag ring are respectively Fixed installation on two conveyor belts. the

the

The advantages of the present invention are: the above scheme provides a method and device for on-line detection and grading of the internal quality of honeydew melons based on the fusion of diffuse transmission spectrum and image information. The internal quality of the honeydew melon can be improved, and at the same time, the deficiency that the internal quality index of the honeydew melon cannot be accurately detected through the image information is overcome.

the

Description of drawings

Fig. 1 is the schematic flow chart of the method for modeling of the present invention;

Fig. 2 is a schematic flow chart of the fruit online quality detection method after the model of the present invention is established;

Fig. 3 is the schematic diagram of device of the present invention;

Fig. 4 is the schematic diagram of delivery device in Fig. 3;

Fig. 5 is the top view of honeydew melon image acquisition device among the present invention;

Fig. 6 is a side view of Fig. 5;

Fig. 7 is a top view of the diffuse transmission spectrum acquisition device in the present invention;

Fig. 8 is a side view of Fig. 7;

Figure 9 is a schematic diagram of the connection between PLC and external equipment in the signal control unit of the present invention;

Figure 10 is a schematic diagram of the method for calculating the volume of honeydew melon;

Figure 11 is a schematic diagram when Δh in Figure 10 becomes very small;

Fig. 12 is a schematic diagram of a method for calculating the fruit shape index of honeydew melon. the

the

Detailed ways

The present invention will be further described below in conjunction with specific embodiments, and the advantages and characteristics of the present invention will become clearer along with the description. However, these embodiments are only exemplary and do not constitute any limitation to the scope of the present invention. Those skilled in the art should understand that the details and forms of the technical solutions of the present invention can be modified or replaced without departing from the spirit and scope of the present invention, but these modifications and replacements all fall within the protection scope of the present invention. the

Referring to Figures 1 to 3, the present invention relates to an online detection method for the internal quality of honeydew melon based on the fusion of diffuse transmission spectrum and image information, including the steps of establishing a non-destructive detection model for the internal quality of honeydew melon and online detection of the internal quality of honeydew melon samples ,details as follows:

(1) Establish a model for non-destructive testing of the internal quality of honeydew melon:

a. Carry out online diffuse transmission spectroscopy and image acquisition of the front and side of honeydew melons for batches of honeydew melon samples;

b. Calibrate and preprocess the collected sample spectrum, using smoothing, differential, multivariate scattering correction, Norris filtering and baseline for correction preprocessing;

c. Extract quality index sugar content and hardness sensitive spectral information, and establish a model of characteristic band; this step adopts the simulated genetic algorithm combining simulated annealing method and genetic algorithm;

）。读取河套蜜瓜图像转化成二值图，针对蜜瓜图像特点通过对图像分别进行腐蚀、膨胀、开运算进行形态学去噪，采用比较成熟的Otsu方法计算图像的全局阈值，并且以此阈值将图像转换成二值图，然后对二值图进行填充，这样做的目的是按照灰度级对像素集合进行划分，得倒的每个子集形成一个与现实景物相对应的区域，各区域内部具有一致的属性，而相邻区域布局有这种一致属性，通过这种方法实现阈值分割，对蜜瓜图片进行完开运算后，并且找到蜜瓜的形心，以蜜瓜形心为基准在蜜瓜图片上标记出六个方形框，这六个框就是我们采集糖度的区域，对已标记的图像，通过MATLAB语句实现颜色平均值的提取，分别提取了R、G、B、H、S、I 、L*、a*、b*等九种颜色特征，对这九种颜色特征采用BP神经网络进行颜色最优特征值得选取，找到与蜜瓜内部品质相关性最高的特征子集作为输入量，为后续信息融合提供数据支持。 d. Preprocess the collected honeydew melon images, extract quality index sugar content, hardness-sensitive color feature subsets; use morphological denoising, threshold segmentation, and image labeling to preprocess the collected honeydew melon images, and use neural networks The algorithm extracts quality index sugar content, hardness-sensitive color feature subsets (H, L*, a*, b*,

). Read the Hetao honeydew melon image and convert it into a binary image. According to the characteristics of the honeydew melon image, the image is corroded, expanded, and opened to perform morphological denoising. The more mature Otsu method is used to calculate the global threshold of the image, and the threshold Convert the image into a binary image, and then fill the binary image. The purpose of this is to divide the pixel set according to the gray level, and each subset of the inverted image forms an area corresponding to the real scene. The interior of each area It has a consistent attribute, and the adjacent area layout has this consistent attribute. Through this method, the threshold segmentation is realized. After the honeydew picture is opened, and the centroid of the honeydew is found, the centroid of the honeydew is used as the benchmark in the Six square boxes are marked on the honeydew melon picture. These six boxes are the areas where we collect sugar content. For the marked image, the color average value is extracted through MATLAB statements, and R, G, B, H, and S are extracted respectively. , I , L*, a*, b* and other nine color features, use BP neural network to select the optimal color feature value for these nine color features, and find the feature subset with the highest correlation with the internal quality of honeydew melon as input Quantity, to provide data support for subsequent information fusion.

、

=

，如图10所示；当Δh逐渐变得很小时，由EF、CD所围部分可近似看成圆台，如图11所示，由此可计算圆台的体积为：  e. Calculate the volume of the honeydew melon; since the honeydew melon is approximately elliptical, it can be assumed that the two-dimensional image of the ideal honeydew melon is symmetrical about the longitudinal diameter, and the idea of layered integration is adopted (the pixel volume of the ideal honeydew melon can be regarded as the volume of a series of conical frustums in the longitudinal direction) Integral) to calculate the volume of the honeydew melon; assuming that the ideal honeydew melon is a symmetric body of rotation about the longitudinal diameter, the following figure is the outline of the ideal honeydew melon, in the figure L is the longitudinal diameter, and is the vertical direction, A is the vertex, and the straight lines CD, EF It is perpendicular to the longitudinal diameter, the left intersection with the contour is C, the right intersection with E is D, F, the distance between CD and EF is Δh, =

,

=

, as shown in Figure 10; when Δh gradually becomes very small, the part surrounded by EF and CD can be approximately regarded as a circular frustum, as shown in Figure 11, from which the volume of the circular frustum can be calculated as:

Therefore, from the perspective of calculus, the volume of an ideal honeydew melon can be regarded as the integral of a series of circular frustums in the longitudinal direction L

横径，进而计算蜜瓜果形指数为：

。 f. Calculate the fruit shape index of the honeydew melon; in view of the approximate elliptical characteristics of the honeydew melon image, the shape of the honeydew melon is obtained through the G channel interception method, the edge information of the honeydew melon is extracted, the centroid of the honeydew melon is found according to the edge information, and then the software caliper is used method to calculate the fruit shape index of honeydew melon; use the minimum circumscribed rectangle method to detect the vertical and horizontal of honeydew melon, preprocess the saved pictures, select the RGB color space, extract the G component image, and convert it into a binary image, using cnny The operator performs edge detection, extracts the edge of the honeydew melon and calculates the smallest circumscribed rectangle containing the edge to obtain the longitudinal diameter of the honeydew melon

Diameter , and then calculate the fruit shape index of honeydew melon as:

.

the

g. The color feature subset, volume, and fruit shape index are fused as image feature variables and spectral information to establish an online detection model for the internal quality of honeydew melons.

(2) After establishing the online detection model of the internal quality of honeydew melon, conduct online detection of the internal quality of unknown honeydew melon samples:

a. Place the honeydew melon horizontally on the drag ring of the conveying device, and move forward at a constant speed;

b. When the honeydew melon reaches the diffuse transmission spectrum collection room, the built-in laser sensor in the collection room outputs a honeydew melon arrival feedback signal to the signal control unit, and the signal control unit sends a command to trigger the spectrometer, and the spectrometer starts to collect the spectral information of the measured honeydew melon , after the melon passes, the laser sensor outputs a melon leaving feedback signal to the signal control unit, and the signal control unit issues a command to stop the spectrometer;

c. When the honeydew melon arrives at the honeydew melon image collection room, the built-in laser sensor in the collection room outputs a honeydew melon arrival feedback signal to the signal control unit, and the signal control unit sends a command to trigger two industrial cameras. Measure the front image information and side image information of the honeydew melon. After the honeydew melon passes, the laser sensor outputs a feedback signal of the honeydew melon leaving to the signal control unit, and the signal control unit issues an order to stop the industrial camera from working;

d. Input the collected spectrum and image information into the online detection model of the internal quality of honeydew melon established in step (1), obtain the internal quality of the measured honeydew melon, and then judge and classify the honeydew melon;

Referring to Figures 3 to 9, a method and device for online detection of the internal quality of honeydew melon based on fusion of diffuse transmission spectrum and image information, including a conveying device, a signal control unit 14, a diffuse transmission spectrum acquisition device, a honeydew melon image acquisition device and computer, the conveyor belt of the conveying device is provided with tow rings for holding honeydew melons, and each tow ring passes through the image acquisition chamber 11 of the honeydew melon image acquisition device and the spectrum acquisition chamber 18 of the diffuse transmission spectrum acquisition device in turn; The laser sensor 2 in the image collection room and the two industrial cameras (the first industrial camera 1, the second industrial camera 4, the laser sensor 7 in the spectrum collection room, and the near-infrared spectrometer 21 all communicate with the internal quality of honeydew melon through the signal control unit 14). The computer 24 of the detection software system is connected.

The conveying device 19 includes a supporting connecting frame 26 and a first belt supporting roller 27, a second belt supporting roller 28, a first driving sprocket 29, a driving chain 30, a second driving belt mounted on the supporting connecting frame (26). Sprocket 31, speed-regulating motor 32, conveyor belt 5 are made up, and the second driving sprocket 31 is installed on the speed-regulating motor 32, and the first driving sprocket 29 is installed on the first belt support roller 27, and described first driving chain Wheel 29 and the second drive sprocket 31 are connected by drive chain 30, and the conveyor belt 5 is wound around the periphery of the first belt support roller 27 and the second belt support roller 28 on the support connection frame, and several drag rings 6 are evenly fixed on on the conveyor belt. the

There are two conveyor belts, which are arranged parallel to each other, and a gap is formed in the middle; the two conveyor belts are installed on the support connecting frame through the first belt support roller and the second belt support roller; the two ends of the drag ring are respectively fixed on the On the two conveyor belts, when the honeydew melons are placed on the support ring, the bottom of the honeydew melon is exposed. The near-infrared spectrometer is located outside the spectrum collection room, at the bottom of the gap between the two conveyor belts, and is used to check the honeydew melons in the drag ring. For spectral collection, its front (ie top) is photographed by an industrial camera located on the front. the

Described diffuse transmission spectrum collecting device comprises spectrum collection room 18, arc-shaped light source fixture 8, halogen light source 10, laser sensor 7 in the spectrum collection room, optical fiber probe 15, near-infrared spectrometer 21, wherein halogen light source 10, The laser sensor 7 and the arc-shaped light source fixing frame 8 in the spectrum collection room are located inside the spectrum collection room 18, the halogen light source 10 is evenly distributed on the arc-shaped light source fixing frame 8, and the two arc-shaped light source fixing frames 8 are arranged opposite to each other. The drag ring passes through the middle of the two arc-shaped light source fixing frames; a vent is arranged on the upper part of the spectrum collection room, and a fan 9 is installed on the upper part of the vent, and the optical fiber probe 15 is located at the bottom of the supporting ring, through the optical fiber (16) Connect with near-infrared spectrometer (21). the

Described honeydew image acquisition device comprises honeydew image acquisition room 11, the first industrial camera 1, the second industrial camera 4, LED light source 3, the laser sensor 2 in the image acquisition room, wherein the first industrial camera (1), the second The laser sensor 2 and the LED light source 3 in the second industrial camera 4, the image acquisition room are located in the honeydew image acquisition room 11, the laser sensors 2 in the two image acquisition rooms are relatively arranged, and the LED light sources 3 are respectively arranged at the four corners of the honeydew image acquisition room . The first industrial camera 1 is positioned at the inner side of honeydew melon image collection chamber 11, is used for collecting the positive image information of measured honeydew melon, i.e. one of the two ends of honeydew melon; The second industrial camera 4 is positioned at the honeydew melon image The bottom of the collection chamber 11 collects image information of the side of the honeydew melon (since the honeydew melon is similar to an ellipse, the side is the side where the long axis of the ellipse is located). the

There are two conveyor belts (5), which are arranged parallel to each other, forming a gap in the middle; the two conveyor belts are installed on the support connecting frame through the first belt support roller and the second belt support roller, and the two ends of the drag ring are respectively Fixed installation on two conveyor belts. the

Claims (6)

1. a honeydew melon inside quality online test method for diffuse transmission spectrum and image information fusion, comprises that setting up honeydew melon inside quality Nondestructive Testing Model and honeydew melon sample interior quality carries out the online step detecting, specific as follows:

(1) set up the model of honeydew melon inside quality Non-Destructive Testing:

A, batch honeydew melon sample is carried out to the positive image acquisition with side of online diffuse transmission spectrum and honeydew melon;

B, to gather sample spectra proofread and correct pre-service;

C, extract index of quality pol, hardness sensitivity spectrum information, set up the model of characteristic wave bands;

D, the honeydew melon image collecting is carried out to pre-service, extract index of quality pol, hardness sensitive colors character subset;

E, calculating honeydew melon volume;

The fruit shape index of f, calculating honeydew melon;

G, color characteristic subset, volume, fruit shape index are merged as characteristics of image variable and spectral information, set up the online detection model of honeydew melon inside quality.

(2) establish after the online detection model of honeydew melon inside quality, unknown honeydew melon sample interior quality detected online:

A, honeydew melon is lain in a horizontal plane on the shackle of conveying device, at the uniform velocity steadily travel forward;

B, arrive when diffuse transmission spectra collection chamber when honeydew melon, gather honeydew melon arrival of indoor laser sensor output of putting and feed back signal to signaling control unit, by the signaling control unit triggering near infrared spectrometer of giving an order, near infrared spectrometer starts to gather the spectral information of tested honeydew melon, after honeydew melon passes through, a honeydew melon of laser sensor output leaves and feeds back signal to signaling control unit, is given an order and is stopped near infrared spectrometer work by signaling control unit;

C, arrive when honeydew melon image acquisition chamber when honeydew melon, the indoor honeydew melon of laser sensor output of image acquisition arrives and feeds back signal to signaling control unit, give an order and trigger two industrial cameras by signaling control unit, two industrial cameras gather respectively direct picture information and the side image information of tested honeydew melon, after honeydew melon passes through, the indoor honeydew melon of laser sensor output of image acquisition leaves and feeds back signal to signaling control unit, is given an order and is stopped industrial camera work by signaling control unit;

D, the spectrum collecting and image information are input in the online detection model of honeydew melon inside quality of setting up in step (1), obtain the inside quality of tested honeydew melon, and then judge honeydew melon grade.

2. one kind implements the claims the device of the honeydew melon inside quality online test method of 1 diffuse transmission spectrum and image information fusion, it is characterized in that, comprise conveying device, signaling control unit, diffuse transmission spectra collection device, honeydew melon image collecting device and computing machine, the travelling belt of described conveying device is provided with the shackle that holds honeydew melon, and each shackle is all successively through the image acquisition chamber of honeydew melon image collecting device and the spectra collection chamber of diffuse transmission spectra collection device; The indoor laser sensor of image acquisition is all connected with the computing machine that honeydew melon inside quality detection software systems are installed by signaling control unit with two industrial cameras, indoor laser sensor, the near infrared spectrometer of spectra collection.

3. device according to claim 2, it is characterized in that, described conveying device (19) comprises support connecting frame (26) and is arranged on the first belt backing roll (27) on support connecting frame (26), the second belt backing roll (28), the first drive sprocket (29), chain drive-belt (30), the second drive sprocket (31), buncher (32) and travelling belt (5), the second drive sprocket (31) is arranged on buncher (32), the first drive sprocket (29) is arranged on the first belt backing roll (27), described the first drive sprocket (29) and the second drive sprocket (31) are in transmission connection by chain drive-belt (30), travelling belt (5) is around the first belt backing roll 27 on support connecting frame and the periphery of the second belt backing roll (28), several shackles are fixed on travelling belt (5) uniformly.

4. device according to claim 2, it is characterized in that, described diffuse transmission spectrum device for picking comprises spectra collection chamber (18), circular-arc light source fixing frame (8), halogen light source (10), the laser sensor (7) that spectra collection is indoor, fibre-optical probe (15), near infrared spectrometer (21), wherein halogen light source (10), the laser sensor (7) that spectra collection is indoor, circular-arc light source fixing frame (8) is positioned at inside, spectra collection chamber (18), halogen light source (10) is evenly distributed on circular-arc light source fixing frame (8), two circular-arc light source fixing frames (8) are oppositely arranged, shackle passes in the middle of two circular-arc light source fixing frames (8), top in spectra collection chamber (18) is provided with ventilating opening, and fan (9) is installed on the top of ventilating opening, and fibre-optical probe (15) is positioned at the bottom of shackle, is connected with near infrared spectrometer (21) by optical fiber.

5. device according to claim 2, it is characterized in that, described honeydew melon image collecting device comprises honeydew melon image acquisition chamber (11), the first industrial camera (1), the second industrial camera (4), LED light source (3), the laser sensor (2) that image acquisition is indoor, wherein the first industrial camera (1), the second industrial camera (4), the laser sensor (2) that image acquisition is indoor and LED light source (3) are positioned at inside, honeydew melon image acquisition chamber (11), two indoor laser sensors (2) of image acquisition are oppositely arranged, LED light source (3) is separately positioned on four jiaos of honeydew melon image acquisition chamber, the first industrial camera (1) and the second industrial camera (4) gather respectively direct picture information and the side image information of tested honeydew melon.

6. device according to claim 2, is characterized in that, described travelling belt (5) is two, be arranged in parallel each other, the middle space that forms; Article two, travelling belt is all arranged on support connecting frame by the first belt backing roll and the second belt backing roll, and the two ends of shackle are fixedly mounted on respectively on two travelling belts.

Images