Thomas F. Schatzki

DEFECT DETECTION IN APPLES BY MEANS OF X-RAY IMAGING

X-ray radiograms were obtained for 400 to 700 each of four State of Washington apple cultivars (‘Fuji’, ‘Granny Smith’, ‘Red’ and ‘Golden Delicious’) and 79 ‘Braeburns’ carrying assorted defects (bruises, senescence browning, rot, watercore and insect damage). Radiograms of whole apples, most out of long term storage, were obtained with line scanning x-ray, suitable for real-time inspection, and with high resolution film, at two orientations, following which the apples were sliced and photographed. Apples were characterized as defective or not based on the appearance in these photos. Sets of x-ray images for a given cultivar/orientation (good and bad apples mixed randomly) were inspected by human observers and the recognition rates recorded. When still images were viewed on a computer screen, acceptable recognition (= 50% of defective apples recognized, = 5% of good apples classified defective) of images was obtained for senescence browning of ‘Red Delicious’, for watercore and stem rot in ‘Fuji’ (requiring orientation), possibly for watercore in ‘Red Delicious’, and for codling moth damage in the first four cultivars 8 to 19 days after larval entry. However, when images were scrolled across the screen at increasing rates, simulating a three-chain sorting line, recognition fell off to unacceptable levels at rates one half that corresponding to a commercial sorting line. This decrease is not unexpected in light of previous work in more structured situations. The implications for devising an apple sorting system based on x-ray are discussed.

Detection and segmentation of items in X-ray imagery

Processing of real–time X–ray images of randomly oriented and touching pistachio nuts for product inspection is considered. Processing to isolate individual nuts (segmentation) is emphasized. The processing consisted of a blob coloring algorithm, filters, and watershed transforms to segment touching nuts, and morphological processing to produce an image of only the nutmeat. Each operation is detailed and quantitative data for each are presented. These techniques are useful for many different product inspection problems in agriculture and other areas.

New feature extraction method for classification of agricultural products from x-ray images

Classification of real-time x-ray images of randomly oriented touching pistachio nuts is discussed. The ultimate objective is the development of a system for automated non- invasive detection of defective product items on a conveyor belt. We discuss the extraction of new features that allow better discrimination between damaged and clean items. This feature extraction and classification stage is the new aspect of this paper; our new maximum representation and discrimination between damaged and clean items. This feature extraction and classification stage is the new aspect of this paper; our new maximum representation and discriminating feature (MRDF) extraction method computes nonlinear features that are used as inputs to a new modified k nearest neighbor classifier. In this work the MRDF is applied to standard features. The MRDF is robust to various probability distributions of the input class and is shown to provide good classification and new ROC data.

Expanded image database of pistachio x-ray images and classification by conventional methods

In order to develop sorting methods for insect damaged pistachio nuts, a large data set of pistachio x-ray images (6,759 nuts) was created. Both film and linescan sensor images were acquired, nuts dissected and internal conditions coded using the U.S. Grade standards and definitions for pistachios. A subset of 1199 good and 686 insect damaged nuts was used to calculate and test discriminant functions. Statistical parameters of image histograms were evaluated for inclusion by forward stepwise discrimination. Using three variables in the discriminant function, 89% of test set nuts were correctly identified. Comparable data for 6 human subjects ranged from 67 to 92%. If the loss of good nuts is held to 1% by requiring a high probability to discard a nut as insect damaged, approximately half of the insect damage present in clean pistachio nuts may be detected and removed by x-ray inspection.

Modified binary watershed transform for segmentation of agricultural products

Segmentation of agricultural products on a conveyor belt is a first step for product inspection. We consider a new algorithm to achieve segmentation. We use x-ray images which provide useful internal information on the state of the product. The product items considered vary in size, shape, gray-scale properties, and lie at random orientations. Many items touch and thus new techniques are necessary to detect each item at any orientation, estimate the number of touching items and their centers, and segment overlapping touching items to provide separate image filters for touching input items. Test results are presented for x-ray film and linescan images of various agricultural products including one extensive database.

Visual detection of particulates in x-ray images of processed meat products

A study was conducted to test the efficacy of detecting particulate contaminants in processed meat samples by visual observation of line-scanned x-ray images. Six hundred field-collected processedproduct samples were scanned at 230 cm2/s using 0.5 X 0.5-mm resolution and 50 kV, 13 mA excitation. The x-ray images were image corrected, digitally stored, and inspected off-line, using interactive image enhancement. Forty percent of the samples were spiked with added contaminants to establish the visual recognition of contaminants as a function of sample thickness (1 to 10 cm), texture of the x-ray image (smooth/ textured), spike composition (wood/bone/glass), size (0.1 to 0.4 cm), and shape (splinter/round). The results were analyzed using a maximum likelihood logistic regression method. In packages less than 6 cm thick, 0.2-cm-thick bone chips were easily recognized, 0.1-cm glass splinters were recognized with some difficulty, while 0.4-cm-thick wood was generally missed. Operational feasibility in a time-constrained setting was confirmed. One half percent of the samples arriving from the field contained bone slivers >1 cm long, 1/2% contained metallic material, while 4% contained particulates exceeding 0.3 cm in size. All of the latter appeared to be bone fragments.

Defect detection in apples by means of x-ray imaging

The possibility of using x-ray radiographic imaging for detecting intenal defects in apples has been investigated. Four hundred to seven hundred each of five Washington State cultivars [Red and Golden Delicious (RD< GD), Fuji (FJ), Granny Smith (GS) and Braeburn (BR)], both defect free and with assorted internal defects (bruises, senescence browning, rot, insect damage and watercore), were imaged using film and on-line line-scanning x-ray equipment. Both axial (stem-to-calyx) and radial images were obtained. The resulting images were presented to human operators, both as still shots and by scrolling them across the screen of a PC at rates approximating that of a commercial packing line. Good recognition [greater than 50% recognized, less than 10% false positives] could generally be obtained on selected cultivars when still shots were inspected. Apple orientation was required for recognition of watercore and rot. However, recognition rapidly fell off as scrolling speeds across the screens approached commercial rates. It is concluded that such inspection may be possible using machine recognition, but probably cannot be achieved using human operators.

Neural net classification of x-ray pistachio nut data

Classification results for agricultural products are presented using a new neural network. This neural network inherently produces higher-order decision surfaces. It achieves this with fewer hidden layer neurons than other classifiers require. This gives better generalization. It uses new techniques to select the number of hidden layer neurons and adaptive algorithms that avoid other such ad hoc parameter selection problems; it allows selection of the best classifier parameters without the need to analyze the test set results. The agriculture case study considered is the inspection and classification of pistachio nuts using x- ray imagery. Present inspection techniques cannot provide good rejection of worm damaged nuts without rejecting too many good nuts. X-ray imagery has the potential to provide 100% inspection of such agricultural products in real time. Only preliminary results are presented, but these indicate the potential to reduce major defects to 2% of the crop with 1% of good nuts rejected. Future image processing techniques that should provide better features to improve performance and allow inspection of a larger variety of nuts are noted. These techniques and variations of them have uses in a number of other agricultural product inspection problems.

Machine Recognition of Navel Orange Worm Damage in X-ray Images of Pistachio Nuts

Insect infestation increases the probability of aflatoxin contamination in pistachio nuts. A non- destructive test is currently not available to determine the insect content of pistachio nuts. This paper uses film X-ray images of various types of pistachio nuts to assess the possibility of machine recognition of insect infested nuts. Histogram parameters of four derived images are used in discriminant functions to select insect infested nuts from specific processing streams.

Detection of insect damage in almonds

Pinhole insect damage in natural almonds is very difficult to detect on-line. Further, evidence exists relating insect damage to aflatoxin contamination. Hence, for quality and health reasons, methods to detect and remove such damaged nuts are of great importance in this study, we explored the possibility of using x-ray imaging to detect pinhole damage in almonds by insects. X-ray film images of about 2000 almonds and x-ray linescan images of only 522 pinhole damaged almonds were obtained. The pinhole damaged region appeared slightly darker than non-damaged region in x-ray negative images. A machine recognition algorithm was developed to detect these darker regions. The algorithm used the first order and the second order information to identify the damaged region. To reduce the possibility of false positive results due to germ region in high resolution images, germ detection and removal routines were also included. With film images, the algorithm showed approximately an 81 percent correct recognition ratio with only 1 percent false positives whereas line scan images correctly recognized 65 percent of pinholes with about 9 percent false positives. The algorithms was very fast and efficient requiring only minimal computation time. If implemented on line, theoretical throughput of this recognition system would be 66 nuts/second.