Fruits quality evaluation

Abstract

In developing country, the grading of fruits are almost handled by human. That is, neverthelessm, time-consuming process and easily inconsistent. Therefore, researchers in this field has developed various algorithms for evaluating quality using computer vision and machine learning. In this research project, I have also proposed a method using image processing technique to inspect quality of fruits involves image segmentation and feature extraction step.

Acknowledgement

I would like to express my special thanks of gratitude to my tutor (Tran Tien Duc), who is eaching at HCMC University of Technology and Education. He helped me a lot in doing this research and also is one of my loved teacher I've ever met and worked together.

Programming languages

• C / C++
• Python
• MATLAB

Library / Packages

• OpenCV (4.5.0)
• Scikit-learn
• OpenCV-CUDA (Optional)

Datasets

Dataset of fruits were obtained from COFILAB team.

Mango	Apple golden

Segmentation and contour detection processes

Process	Mango	Apple golden
Gray
Blur
Inrange
Canny
Dilate
FloodFill
Remove noise
Contour
Area
Dimension
Isolation

Feature extraction

• Mango

File name	Area measured by image analysis (number of pixels)	Width measured by image analysis	Height measured by image analysis	Mass measured by weighting (g)	Status
Mango_01_B.JPG	267919	637	556	436.78	Maturity
Mango_02_B.JPG	314896	694	586	511.79	Maturity
Mango_03_B.JPG	294155	646	561	444.76	Maturity
Mango_04_B.JPG	311824	694	567	459.47	Maturity
Mango_05_B.JPG	299544	676	591	457.67	Maturity
Mango_06_B.JPG	296080	699	533	481.72	Maturity
Mango_07_B.JPG	313918	712	576	474.54	Maturity
Mango_08_B.JPG	301669	669	572	473.52	Maturity
Mango_09_B.JPG	315649	697	600	479.97	Maturity
Mango_10_B.JPG	302939	679	599	478.47	Maturity
Mango_11_B.JPG	321777	706	586	477.33	Maturity
Mango_12_B.JPG	285883	655	562	455.42	Maturity
Mango_13_B.JPG	315587	682	595	487.60	Maturity
Mango_14_B.JPG	295082	653	599	472.67	Maturity
Mango_15_B.JPG	293621	662	576	480.83	Maturity
Mango_16_B.JPG	291285	672	543	440.62	Maturity
Mango_17_B.JPG	314757	675	607	524.13	Maturity
Mango_18_B.JPG	300686	684	578	472.50	Maturity
Mango_19_B.JPG	303050	673	589	508.37	Maturity
Mango_20_B.JPG	333246	720	622	516.80	Maturity
Mango_21_B.JPG	323454	701	617	458.57	Maturity
Mango_22_B.JPG	310483	648	597	590.16	Maturity
Mango_23_B.JPG	294239	658	597	545.44	Maturity
Mango_24_B.JPG	289883	680	569	458.00	Maturity
Mango_25_B.JPG	296397	657	591	463.54	Immaturity
Mango_26_B.JPG	277712	665	545	401.21	Immaturity
Mango_27_B.JPG	287156	643	586	437.34	Immaturity
Mango_28_B.JPG	325324	703	613	532.63	Immaturity
Mango_29_B.JPG	285720	659	566	437.36	Immaturity
Mango_30_B.JPG	285922	674	548	417.73	Immaturity
Mango_31_B.JPG	303157	655	612	494.06	Immaturity
Mango_32_B.JPG	314016	678	608	513.60	Immaturity
Mango_33_B.JPG	304348	707	575	477.24	Immaturity
Mango_34_B.JPG	321335	709	607	508.97	Immaturity
Mango_35_B.JPG	301064	704	580	464.89	Immaturity
Mango_36_B.JPG	293339	710	545	427.72	Immaturity
Mango_37_B.JPG	381422	775	634	640.44	Immaturity
Mango_38_B.JPG	309734	709	576	492.18	Immaturity
Mango_39_B.JPG	411628	834	654	739.40	Immaturity
Mango_40_B.JPG	420399	845	661	757.72	Immaturity
Mango_41_B.JPG	375253	773	629	658.54	Immaturity
Mango_42_B.JPG	360528	748	603	620.49	Immaturity
Mango_43_B.JPG	345198	728	645	592.82	Immaturity
Mango_44_B.JPG	335209	718	612	557.35	Immaturity
Mango_45_B.JPG	312361	752	544	484.43	Immaturity
Mango_46_B.JPG	401422	800	649	663.17	Immaturity
Mango_47_B.JPG	352250	734	614	600.40	Immaturity
Mango_48_B.JPG	273928	664	553	505.75	Immaturity
Mango_49_B.JPG	348721	746	615	458.05	Immaturity
Mango_50_B.JPG	310015	677	585	541.46	Immaturity

• Apple golden

File name	Area measured by image analysis (number of pixels)	Width measured by image analysis	Height measured by image analysis	Mass measured by weighting (g)
Golden_01_3.JPG	325456	713	620	223.61
Golden_02_3.JPG	292818	605	587	212.49
Golden_03_3.JPG	308248	691	573	216.70
Golden_04_3.JPG	330242	639	644	237.87
Golden_05_3.JPG	258536	606	551	170.47
Golden_06_3.JPG	324243	654	610	224.15
Golden_07_3.JPG	281514	616	581	197.17
Golden_08_3.JPG	292763	651	608	206.68
Golden_09_3.JPG	310381	659	624	215.81
Golden_10_3.JPG	262452	590	543	180.21
Golden_11_3.JPG	310246	652	594	228.82
Golden_12_3.JPG	295091	612	617	208.29
Golden_13_3.JPG	294691	687	599	208.82
Golden_14_3.JPG	318046	647	617	212.37
Golden_15_3.JPG	340940	693	633	236.41
Golden_16_3.JPG	317143	663	584	219.46
Golden_17_3.JPG	275108	590	587	189.83
Golden_18_3.JPG	328042	700	600	226.72
Golden_19_3.JPG	329161	672	623	236.05
Golden_20_3.JPG	333618	688	589	229.25
Golden_21_3.JPG	315626	645	609	219.08
Golden_22_3.JPG	279182	616	563	193.42
Golden_23_3.JPG	319838	644	607	227.74
Golden_24_3.JPG	336852	712	638	241.11
Golden_25_3.JPG	318909	621	636	230.26

Regression analysis for estimation of the weight

• Mango

Parameters	Value
Coefficient	0.002
Intercept	215.755

Score	Value
Training	0.97
Test	0.96

• Apple golden

Parameters	Value
Coefficient	0.0008
Intercept	16.0667

Score	Value
Training	0.91
Test	0.92

Contact

Cao Le Cong Minh - caolecongminh1997@gmail.com