Author Affiliations
School of Electromechanical and Vehicle Engineering, East China Jiaotong University, Institute of Intelligent Electromechanical Equipment Innovation, Nanchang 330013, Chinashow less
Fig. 1. Schematic diagram of on-line detection equipment for fruit internal quality by near infrared diffuse transmission spectroscopy
1: Porximity switch; 2: Coding mask; 3: Chain wheel; 4: Link chain; 5: Optica fiber; 6: Camera bellows; 7: Optical source; 8: Specimen; 9: Terminal PC; 10: PLC control cabinet; 11: Fruit cup; 12: Bionic boot; 13: Grading export; 14: Electromotor; 15: Reducer; 16: Drive sprocket
Fig. 2. Diffuse transmission testing mechanism (a), Gongli placement position (b)
Fig. 3. Establishment of local model and global model and experimental verification scheme
Fig. 4. Average spectra of fruit No.11 in six directions
Fig. 5. Scatter plot of omnidirectional verification
Parameter | Data set | Samples | Mean | S.D | Range |
---|
SSC/ (°Brix) | Calibration | 115 | 12.06 | 1.05 | 9.53~14.70 | Prediction | 35 | 12.09 | 0.93 | 9.60~13.37 |
|
Table 1. Range, standard deviation and average value of SSC content in calibration set and prediction set
Orientation | Pretreatment | Calibration | Prediction |
---|
Rc | RMSEC | Rp | RMSEP |
---|
A1 | Raw | 0.916 | 0.406 | 0.870 | 0.456 | | SGS | 0.870 | 0.497 | 0.873 | 0.453 | | MSC | 0.886 | 0.468 | 0.839 | 0.503 | | GFS | 0.898 | 0.443 | 0.873 | 0.452 | A2 | Raw | 0.987 | 0.151 | 0.898 | 0.327 | | SGS | 0.955 | 0.280 | 0.869 | 0.374 | | MSC | 0.938 | 0.327 | 0.879 | 0.398 | | GFS | 0.982 | 0.175 | 0.897 | 0.331 | A3 | Raw | 0.922 | 0.377 | 0.919 | 0.363 | | SGS | 0.900 | 0.424 | 0.856 | 0.339 | | MSC | 0.913 | 0.397 | 0.885 | 0.315 | | GFS | 0.913 | 0.397 | 0.872 | 0.342 | A4 | Raw | 0.961 | 0.278 | 0.871 | 0.399 | | SGS | 0.922 | 0.387 | 0.883 | 0.375 | | MSC | 0.965 | 0.262 | 0.865 | 0.402 | | GFS | 0.950 | 0.311 | 0.880 | 0.386 | A5 | Raw | 0.945 | 0.301 | 0.863 | 0.384 | | SGS | 0.892 | 0.414 | 0.846 | 0.407 | | MSC | 0.932 | 0.333 | 0.896 | 0.344 | | GFS | 0.930 | 0.337 | 0.859 | 0.389 | A6 | Raw | 0.835 | 0.527 | 0.794 | 0.508 | | SGS | 0.825 | 0.542 | 0.794 | 0.509 | | MSC | 0.910 | 0.402 | 0.762 | 0.542 | | GFS | 0.831 | 0.534 | 0.794 | 0.508 |
|
Table 2. Local model and local prediction effect in six directions
Orientation | Pretreatment | LVs | Calibration | Prediction |
---|
A1 | A2 | A3 |
---|
Rc | RMSEC | Rp | RMSEP | Rp | RMSEP | Rp | RMSEP | Omnidirectional | Raw | 16 | 0.844 | 0.406 | 0.810 | 0.497 | 0.766 | 0.523 | 0.792 | 0.479 | GFS | 16 | 0.828 | 0.424 | 0.818 | 0.446 | 0.765 | 0.525 | 0.799 | 0.478 |
|
Table 3. Global model and prediction effect of A3, A4 and A5 directions
Orientation | Pretreatment | LVs | Calibration | Prediction |
---|
A4 | A5 | A6 |
---|
Rc | RMSEC | Rp | RMSEP | Rp | RMSEP | Rp | RMSEP | Omnidirectional | Raw | 16 | 0.844 | 0.406 | 0.801 | 0.538 | 0.785 | 0.492 | 0.821 | 0.612 | GFS | 16 | 0.828 | 0.424 | 0.821 | 0.538 | 0.794 | 0.486 | 0.824 | 0.619 |
|
Table 4. Global model and prediction effect of A4, A5 and A6 directions
Orientation | LVs | Calibration | Prediction |
---|
A1 | A2 | A3 |
---|
Rc | RMSEC | Rp | RMSEP | Rp | RMSEP | Rp | RMSEP | Omnidirectional | 16 | 0.828 | 0.424 | 0.818 | 0.446 | 0.765 | 0.525 | 0.799 | 0.478 | A1 | 13 | 0.898 | 0.443 | 0.873 | 0.452 | 0.182 | 1.431 | 0.731 | 0.816 | A2 | 14 | 0.982 | 0.175 | NA | 0.336 | 0.897 | 0.331 | 0.765 | 0.914 | A3 | 9 | 0.913 | 0.397 | NA | 13.972 | NA | 3.493 | 0.872 | 0.342 | A4 | 12 | 0.950 | 0.311 | 0.328 | 0.410 | 0.532 | 1.735 | 0.814 | 0.545 | A5 | 14 | 0.930 | 0.337 | NA | 0.134 | 0.552 | 1.058 | 0.797 | 0.633 | A6 | 7 | 0.831 | 0.534 | 0.348 | 0.330 | NA | 1.133 | 0.794 | 0.056 |
|
Table 5. Prediction effects of local model and global model on A1, A2 and A3 direction validation sets
Orientation | LVs | Calibration | Prediction |
---|
A4 | A5 | A6 |
---|
Rc | RMSEC | Rp | RMSEP | Rp | RMSEP | Rp | RMSEP | Omnidirectional | 16 | 0.828 | 0.424 | 0.821 | 0.538 | 0.794 | 0.486 | 0.824 | 0.619 | A1 | 13 | 0.898 | 0.443 | 0.684 | 0.767 | 0.490 | 1.015 | 0.372 | 0.089 | A2 | 14 | 0.982 | 0.175 | NA | 1.196 | NA | 2.024 | NA | 1.618 | A3 | 9 | 0.913 | 0.397 | 0.714 | 0.791 | 0.471 | 1.564 | 0.709 | 0.820 | A4 | 12 | 0.950 | 0.311 | 0.880 | 0.386 | 0.153 | 2.144 | 0.190 | 1.946 | A5 | 14 | 0.930 | 0.337 | 0.752 | 0.650 | 0.859 | 0.389 | 0.722 | 0.659 | A6 | 7 | 0.831 | 0.534 | 0.728 | 0.632 | 0.827 | 0.537 | 0.794 | 0.508 |
|
Table 6. Prediction effects of local model and global model on A4, A5 and A6 direction validation sets