Figure - available from: Journal of Spectroscopy
This content is subject to copyright. Terms and conditions apply.
Source publication
Bruise may cause spoilage, reduce commodity economic value, and give rise to food quality and safety concerns. Therefore, it is crucial to detect whether a loquat is bruised and when it is bruised to save storage and transportation costs. At present, the bruise of loquats is mainly discriminated by the operator’s naked eye, which is affected by per...
Citations
... The above findings indicate that the method offers potential for non-destructive detection of changes in TSS, AA, total chlorophyll and carotenoid content during fruit maturation. In addition, hyperspectral imaging has been used for peach [11], pear [12,13], jujube [14,15], loquat [16,17] and other fruits for mechanical damage detection. However, for the current mainstream hyperspectral imaging systems, acousto-optic tunable filters (AOTF) and liquid crystal tunable filters (LCTF) are usually used to separate the light source spectrum for hyperspectral imaging [18,19], which are relatively expensive and difficult to be widely used. ...
Hyperspectral imaging, as a non-destructive testing technique with the ability to acquire rich spatial and spectral information, is a potential tool for fruit bruise detection. However, the cost issue is one of the main reasons limiting the popularity of this technique. In addition, it is difficult to detect the entire surface of samples such as fruits due to the uncertainty of the damaged area. Furthermore, directional scattering on the target sample can lead to bright spots and shadows on the acquired hyperspectral image, which can affect the image quality and add additional preprocessing steps. Therefore, a monochromatic illumination model based on the combination of Xenon lamp and reflective grating is proposed. An experimental prototype of the system has realized 101 spectral channels in the 400–700 nm range. The sample is placed in an optical integrating sphere with rollers for indirect illumination to avoid spots and shadows during image acquisition, enabling multi-surface imaging. This prototype is then used to prepare hyperspectral datasets of sound apples and bruised apples. Models built using classic classification algorithms SVM (87.5%), k-NN (82.5%) AlexNet (95%), VGG16 (95%), ResNet (100%), and achieve effective results in tests. The results demonstrate that the HSI system we designed has exceptional performance in apple mechanical damage detection and classification, showing the advantages of good spectral resolution, low cost, and low thermal effect.
... First, we used the competitive adaptive weighted sampling method (CARS) to screen the characteristic bands. CARS is a sampling method for feature variable selection based on regression coefficients, which is widely used for the selection of characteristic wavelengths of spectral data [31,32]. ...
Egg freshness is essential for evaluating the internal quality of eggs. Here, we propose a method based on feature fusion to improve the accuracy of freshness classification. First, hyperspectral reflectance images of 264 egg samples of three freshness grades were acquired by a hyperspectral image acquisition system. Spectral features were extracted from the region of interest of the reflectance hyperspectral images in the range of 440.51-950.24 nm. Second, after pre-processing and characteristic bands selection, the characteristic images were extracted from the images corresponding to the characteristic bands using principal component analysis; then the texture features of each egg sample were obtained from the characteristic images by gray-level co-occurrence matrices. Third, a variety of freshness discrimination models based on spectral and texture features were established, and the models based on the single feature had a maximum accuracy of 89.39%. Finally, well-performing models based on a single feature were merged into a robust model by the stacking ensemble learning method to realize decision-level fusion of the two features, and the highest accuracy of the prediction set was increased to 92.42%. Thus, the feature fusion method based on decision level is feasible for egg freshness classification.
... Siedliska et al. [30] constructed a study to determine bruising in five apple varieties using short-wave infrared hyperspectral imaging technology, and attained a successful classification ratebetter than 93%. Many researchers investigating food quality using the SWIR spectral range have achieved good results in other fruits and vegetables, such as peaches [31], loquats [32], strawberries [33], and potatoes [34]. Therefore, it is also worthwhile to explore early bruise detection in the short-wave infrared spectral range, especially for events that occur within half an hour. ...
Bruising is a common occurrence in apples that can lead to gradual fruit decay and substantial economic losses. Due to the lack of visible external features, the detection of early-stage bruising (occurring within 0.5 h) is difficult. Moreover, the identification of stems and calyxes is also important. Here, we studied the use of the short-wave infrared (SWIR) camera and the Faster RCNN model to enable the identification of bruises on apples. To evaluate the effectiveness of early bruise detection by SWIR bands compared to the visible/near-infrared (Vis/NIR) bands, a hybrid dataset with images from two cameras with different bands was used for validation. To improve the accuracy of the model in detecting apple bruises, calyxes, and stems, several improvements are implemented. Firstly, the Feature Pyramid Network (FPN) structure was integrated into the ResNet50 feature extraction network. Additionally, the Normalization-based Attention Module (NAM) was incorporated into the residual network, serving to bolster the attention of model towards detection targets while effectively mitigating the impact of irrelevant features. To reduce false positives and negatives, the Intersection over Union (IoU) metric was replaced with the Complete-IoU (CIoU). Comparison of the detection performance of the Faster RCNN model, YOLOv4P model, YOLOv5s model, and the improved Faster RCNN model, showed that the improved model had the best evaluation indicators. It achieved a mean Average Precision (mAP) of 97.4% and F1 score of 0.87. The results of research indicate that it is possible to accurately and effectively identify early bruises, calyxes, and stems on apples using SWIR cameras and deep learning models. This provides new ideas for real-time online sorting of apples for the presence of bruises.
... Loquat is an economical fruit for both medicine and food, which is native to China (Han et al., 2022). Skin defects directly affect the price of fresh loquats, due to the most consumers associate skin defects with poor quality. ...
... The band ratio algorithm can enhance the contrast of defective areas, which is widely used in image processing Li et al. (2016). Li et al. (2022) presented a multispectral analysis method, the PCA combined with band ratio algorithm and gray level cooccurrence matrix, with a classification is 91.3 %. Liu et al. (2016) used PCA to select six characteristics wavelengths in the NIR region, the two-band ratio (R 1160 /R 1464 ) combining with image subtraction algorithm (R 1160 -R 1464 ) was used to detect external insect damage in jujubes, and the classification rate of it was 93.1 %. ...
... The average spectra of the region of interest (ROIs) within 10*10 pixels is extracted as the characteristic spectra can reduce the error (Li et al., 2022). Fig.4 shows the average reflectance of loquats with different skin defects in full-band (397-1100 nm). ...
Background and objectives
Skin defects are one of the primary problems that occur in post-harvest grading and processing of loquats. Skin defects lead to the loquat being easily destroyed during transportation and storage, which causes the risk of other loquats being infected, affecting the selling price.
Materials and Methods
In this paper, a method combining band radio image with an improved three-phase level set segmentation algorithm (ITPLSSM) is proposed to achieve high accuracy, rapid, and non-destructive detection of skin defects of loquats. Principal component analysis (PCA) was used to find the characteristic wavelength and PC images to distinguish four types of skin defects. The best band ratio image based on characteristic wavelength was determined.
Results
The band ratio image (Q782/944) based on PC2 image is the best segmented image. Based on pseudo-color image enhancement, morphological processing, and local clustering criteria, the band ratio image (Q782/944) has better contrast between defective and normal areas in loquat. Finally, the ITPLSSM was used to segment the processing band ratio image (Q782/944), with an accuracy of 95.28%.
Conclusions
The proposed ITPLSSM method is effective in distinguishing four types of skin defects. Meanwhile, it also effectively segments images with intensity inhomogeneities.
